Li Et Al-2021-Cochrane Database of Systematic Reviews PDF

Cochrane
Library
Cochrane Database of Systematic Reviews

Glucocorticoid with cyclophosphamide for oral paraquat poisoning
(Review)
Li LR, Chaudhary B, You C, Dennis JA, Wakeford H
Li LR, Chaudhary B, You C, Dennis JA, Wakeford H.

Glucocorticoid with cyclophosphamide for oral paraquat poisoning.
Cochrane Database of Systematic Reviews 2021, Issue 6. Art. No.: CD008084.
DOI: 10.1002/14651858.CD008084.pub5.
www.cochranelibrary.com

Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review)
Copyright © 2021 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
Cochrane Trusted evidence.
Informed decisions.

Library Better health. Cochrane Database of Systematic Reviews
TABLE OF CONTENTS
ABSTRACT..................................................................................................................................................................................................... 1
PLAIN LANGUAGE SUMMARY....................................................................................................................................................................... 2
SUMMARY OF FINDINGS.............................................................................................................................................................................. 4
BACKGROUND.............................................................................................................................................................................................. 6
OBJECTIVES.................................................................................................................................................................................................. 7
METHODS..................................................................................................................................................................................................... 7
Figure 1.................................................................................................................................................................................................. 8
Figure 2.................................................................................................................................................................................................. 9
Figure 3.................................................................................................................................................................................................. 10
RESULTS........................................................................................................................................................................................................ 11
Figure 4.................................................................................................................................................................................................. 15
Figure 5.................................................................................................................................................................................................. 15
Figure 6.................................................................................................................................................................................................. 16
DISCUSSION.................................................................................................................................................................................................. 16
AUTHORS' CONCLUSIONS........................................................................................................................................................................... 18
ACKNOWLEDGEMENTS................................................................................................................................................................................ 18
REFERENCES................................................................................................................................................................................................ 19
CHARACTERISTICS OF STUDIES.................................................................................................................................................................. 21
DATA AND ANALYSES.................................................................................................................................................................................... 29
Analysis 1.1. Comparison 1: All-cause mortality, Outcome 1: All-cause mortality............................................................................ 29
Analysis 1.2. Comparison 1: All-cause mortality, Outcome 2: Sensitivity analysis: all-cause mortality re-including fulminant 30
participants............................................................................................................................................................................................
Analysis 1.3. Comparison 1: All-cause mortality, Outcome 3: Sensitivity analysis: all-cause mortality: excluding studies without 30
plasma paraquat assessment at baseline...........................................................................................................................................
APPENDICES................................................................................................................................................................................................. 30
WHAT'S NEW................................................................................................................................................................................................. 34
HISTORY........................................................................................................................................................................................................ 35
CONTRIBUTIONS OF AUTHORS................................................................................................................................................................... 35
DECLARATIONS OF INTEREST..................................................................................................................................................................... 36
SOURCES OF SUPPORT............................................................................................................................................................................... 36
DIFFERENCES BETWEEN PROTOCOL AND REVIEW.................................................................................................................................... 36
INDEX TERMS............................................................................................................................................................................................... 36
Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) i

Informed decisions.

[Intervention Review]
Glucocorticoid with cyclophosphamide for oral paraquat poisoning
Luying Ryan Li1, Bhuwan Chaudhary2, Chao You1, Jane A Dennis3, Helen Wakeford3
1Department of Neurosurgery, West China Hospital, Sichuan University, Chengdu, China. 2West China Medical School, Sichuan
University, Chengdu, China. 3Cochrane Injuries Group, London School of Hygiene & Tropical Medicine, London, UK
Contact: Luying Ryan Li, lirang58@hotmail.com.
Editorial group: Cochrane Injuries Group.

Publication status and date: New search for studies and content updated (conclusions changed), published in Issue 6, 2021.
Citation: Li LR, Chaudhary B, You C, Dennis JA, Wakeford H. Glucocorticoid with cyclophosphamide for oral paraquat poisoning.
Cochrane Database of Systematic Reviews 2021, Issue 6. Art. No.: CD008084. DOI: 10.1002/14651858.CD008084.pub5.
ABSTRACT
Background
This an update of a Cochrane Review.
Paraquat is a widely used herbicide, but is also a lethal poison. In some low- and middle-income countries (LMICs) paraquat is commonly
available and inexpensive, making poisoning prevention difficult. Most of the people poisoned by paraquat have taken it as a means of
self-poisoning.
Standard treatment for paraquat poisoning prevents further absorption and reduces the load of paraquat in the blood through
haemoperfusion or haemodialysis. The effectiveness of standard treatments is extremely limited.
The immune system plays an important role in exacerbating paraquat-induced lung fibrosis. Immunosuppressive treatment using
glucocorticoid and cyclophosphamide in combination has been developed and studied as an intervention for paraquat poisoning.
Objectives
To assess the effects of glucocorticoid with cyclophosphamide for moderate to severe oral paraquat poisoning.
Search methods
The most recent searches were run in September 2020. We searched the Cochrane Central Register of Controlled Trials (CENTRAL) (which
contains the Cochrane Injuries Trials Register), Ovid MEDLINE(R), Ovid MEDLINE In-Process & Other Non-Indexed Citations, Ovid MEDLINE
Daily and Ovid OLDMEDLINE, Embase Classic + Embase (Ovid), ISI WOS (SCI-EXPANDED, SSCI, CPCI-S, and CPSI-SSH), and trials registries.
We also searched the following three resources: China National Knowledge Infrastructure database (CNKI 数据库); Wanfang Data (万⽅
数据库); and VIP (维普数据库) on 12 November 2020. We examined the reference lists of included studies and review papers.
Selection criteria
We included randomised controlled trials (RCTs). For this update, in accordance with Cochrane Injuries' Group policy (2015), we included
only prospectively registered RCTs for trials published after 2010. We included trials which assessed the effects of glucocorticoid with
cyclophosphamide delivered in combination. Eligible comparators were standard care (with or without a placebo), or any other therapy
in addition to standard care. Outcomes of interest included mortality and infections.
Data collection and analysis

We calculated the mortality risk ratio (RR) and 95% confidence interval (CI). Where possible, we summarised data for all-cause mortality at
relevant time periods (from hospital discharge to three months after discharge) in meta-analysis, using a fixed-effect model. We conducted
sensitivity analyses based on factors including whether participants were assessed at baseline for plasma paraquat levels. We also reported
data on infections within one week after initiation of treatment.
Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) 1
Informed decisions.

Main results
We included four trials with a total of 463 participants. The included studies were conducted in Taiwan (Republic of China), Iran, and Sri
Lanka. Most participants were male. The mean age of participants was 28 years.
We judged two of the four included studies, including the largest and most recently conducted study (n = 299), to be at low risk of bias for
key domains including sequence generation. We assessed one study to be at high risk of selection bias and another at unclear risk, since
allocation concealment was either not mentioned in the trial report or explicitly not undertaken. We assessed three of the four studies to
be at unclear risk of selective reporting, as no protocols could be identified. An important source of heterogeneity amongst the included
studies was the method of assessment of participants' baseline severity using analysis of plasma levels (two studies employed this method,
whilst the other two did not).
No studies assessed the outcome of mortality at 30 days following ingestion of paraquat.
Low-certainty evidence from two studies indicates that glucocorticoids with cyclophosphamide in addition to standard care may slightly
reduce the risk of death in hospital compared to standard care alone ((RR 0.82, 95% CI 0.68 to 0.99; participants = 322); results come
from sensitivity analysis excluding studies not assessing plasma at baseline). However, we have limited confidence in this finding as
heterogeneity was high (I2 = 77%) and studies varied in terms of size and comparators. A single large study provided data showing that
there may be little or no effect of treatment at three months post discharge from hospital (RR 0.98, 95% CI 0.85 to 1.13; 1 study, 293
participants; low-certainty evidence); however, analysis of long-term results amongst participants whose injuries arose from self-poisoning
must be interpreted with caution.
We remain uncertain of the effect of glucocorticoids with cyclophosphamide on infection within one week after initiation of the treatment;
this outcome was assessed by two small studies only (31 participants, very low-certainty evidence) that considered leukopenia as a proxy
or risk factor for infection. Neither study reported infections in any participants.
Authors' conclusions
Low-certainly evidence suggests that glucocorticoids with cyclophosphamide in addition to standard care may slightly reduce mortality in
hospitalised people with oral paraquat poisoning. However, we have limited confidence in this finding because of substantial heterogeneity
and concerns about imprecision. Glucocorticoids with cyclophosphamide in addition to standard care may have little or no effect on
mortality at three months after hospital discharge. We are uncertain whether glucocorticoid with cyclophosphamide puts patients at an
increased risk of infection due to the limited evidence available for this outcome. Future research should be prospectively registered
and CONSORT-compliant. Investigators should attempt to ensure an adequate sample size, screen participants for inclusion rigorously,
and seek long-term follow-up of participants. Investigators may wish to research the effects of glucocorticoid in combination with other
treatments.
PLAIN LANGUAGE SUMMARY
What are the benefits and risks of treating paraquat poisoning with a combination of steroids and cyclophosphamide (an anti-
cancer medicine)?
Key messages
- Steroids given with cyclophosphamide (an anti-cancer medicine) are unlikely to reduce the risk of death after paraquat poisoning in the
short term, or at three months after hospital discharge.
- We are uncertain whether these medicines increase the risk of infection.
- Future studies need to be larger, measure the level of paraquat poisoning of patients accurately, and monitor patients in the long term.
Research into steroids combined with other treatments could be useful.
What happens in people with paraquat poisoning?
Paraquat is used as a herbicide, but is also a deadly poison. Most people who are poisoned by paraquat have taken it as a means of self-
poisoning.
Treatment for paraquat poisoning focuses on the physical removal (via stomach pumping and other methods) of as much paraquat as
possible from the person's digestive system (stomach) and blood. Any paraquat that remains inside the body causes inflammation that
can damage the lungs severely and lead to death.
Steroids and cyclophosphamide (a medicine normally used in cancer treatment) are medicines that fight inflammation and so are also
used to treat paraquat poisoning.
What did we want to find out?

Informed decisions.

We wanted to find out if a combination of steroids and cyclophosphamide (plus usual care) works better than usual care alone to reduce
the number of people who die from paraquat poisoning.
We also wanted to find out if treatment with steroids plus cyclophosphamide causes an increased number of infections in patients.
What did we do?
We searched for studies that investigated the use of steroids and cyclophosphamide (plus usual care) compared with usual care alone in
people poisoned with paraquat.
We compared and summarised the results of the studies and rated our confidence in the evidence, based on factors such as study methods
and sizes.
What did we find?
We found four studies that involved 463 people with confirmed paraquat poisoning. Two studies were conducted in Taiwan (Republic of
China), one in Iran, and one in Sri Lanka.
All participants were given either:
- usual care only, or
- steroids (methylprednisolone alone, or with dexamethasone) plus cyclophosphamide, as well as usual care. Cyclophosphamide was given
before the steroid(s) or at the same time as them.
Two of the studies measured the severity of poisoning by testing patients’ plasma (a component of blood) at the start of the study. Plasma
tests provide the best assessment of how seriously a person is affected by paraquat poisoning.
One study used a placebo (sham) treatment in addition to usual care. Two studies gave patients a steroid (dexamethasone) as part of the
usual care.
Death while in hospital
The combined results of two studies showed that steroids plus cyclophosphamide (plus usual care) may slightly reduce the risk of death
compared to usual care alone (with, or without, placebo) in people with paraquat poisoning.
Death 3 months after hospital discharge
One large study showed that at 3 months after discharge from hospital there may be no difference in the number of deaths between the
people treated with steroids plus cyclophosphamide (plus usual care) and those treated with usual care alone.
Infection
Two small studies checked levels of white blood cells in patients (low levels can increase risk of infection). Neither study reported any
infections in the week following treatment with steroids and cyclophosphamide. Due to the small size of the studies, we are very uncertain
about whether the treatment affects the risk of infection within one week of treatment.
What are the limitations of the evidence?
The four studies differed in terms of the number of people in them, assessment of level of paraquat poisoning, and types of treatment.
This limited our ability to draw firm conclusions from the evidence.
Overall, the studies we found were too small to provide answers to our questions.
How up to date is this evidence?
This review updates our previous review on this subject. The evidence is up to date to September 2020.

SUMMARY OF FINDINGS

Summary of findings 1. Glucocorticoid with cyclophosphamide plus usual care compared to usual care (with or without placebo) for oral paraquat
Library
Cochrane
poisoning
Glucocorticoid with cyclophosphamide plus usual care compared to usual care, with or without placebo, for oral paraquat poisoning
Patient or population: people with moderate to severe oral paraquat poisoning

Setting: hospital
Better health.
Informed decisions.
Trusted evidence.
Intervention: glucocorticoid(s) with cyclophosphamide plus usual care
Comparison: usual care (with or without placebo)
Outcomes № of participants Certainty of the Relative effect Anticipated absolute effects* (95% CI)
(studies) evidence (95% CI)
(GRADE) Risk with standard care Risk with glucocorticoid plus
cyclophosphamide
Mortality at 30 days fol- No included studies reported this outcome.

lowing the ingestion of
paraquat
All-cause mortality at final 322 ⊕⊕⊝⊝ (RR 0.82, 95% CI 0.68 to 0.99) 63 per 100 52 per 100 (43 to 62)
follow-up: at hospital dis- (2 RCTs) LOW 1
charge (results of a sensitivity analysis)
All-cause mortality at final 293 ⊕⊕⊝⊝ RR 0.98 (0.85 to 1.13) 72 per 100 71 per 100 (61 to 81)
follow-up: 3 months (1 RCT) LOW 2,3
New infection within 1 0 ⊕⊝⊝⊝ No clinical infections were diagnosed within 1 week after initiation of methylprednisolone and
week after initiation of (0 RCTs) VERY LOW 4 cyclophosphamide in the included studies. Lin 1999 reported the related outcome of leukopenia
treatment in 8 of 22 participants (36.4%) in the intervention arm. These participants spontaneously recov-

ered 1 week later, with no mortality. Lin 2006 reported leukopenia in 6 of 16 participants (37.4%)
Assessed through clinical in the intervention arm, all of whom recovered within 1 to 2 weeks. Neither Gawarammana 2017,
diagnosis the largest included study, nor Afzali 2008, measured infection or leukopenia.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and
its 95% CI).
CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence

High certainty: we are very confident that the true effect lies close to that of the estimate of the effect.
4

Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is
substantially different.
Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.
Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
Library
Cochrane
1The sample size fell short of the required information size (439 participants required in each arm of the study), therefore we downgraded one level for imprecision, and a further
level because of the high heterogeneity (I2 = 77%).
2We downgraded one level for imprecision, as the sample fell short of the required information size (as above).
3We downgraded once for indirectness: we consider that the reasons for participant mortality whilst being treated after self-poisoning in hospital may differ substantially from
Better health.
Informed decisions.
Trusted evidence.
the reasons for participant mortality at longer-term follow-up, in the absence of any other information on cause of death.
4The largest RCT, Gawarammana 2017, did not assess infection as an outcome, therefore this outcome was assessed by only two small trials. We therefore downgraded twice for
serious imprecision, and once for indirectness (i.e. the outcome of leukopenia is a precursor of, but not identical to, infection).


5

Informed decisions.

BACKGROUND Although the lung is the primary target organ, multiple organs are
affected by paraquat poisoning. Renal function is impaired as the
Paraquat is one of the most widely used herbicides worldwide, body attempts to excrete paraquat, and the hepatobiliary system,
and is highly toxic to humans when ingested. It is commercially nervous system, and heart are also affected, often resulting in
produced and has been sold in around 130 countries since 1961 multiple organ failure (Gao 2020).
(Tomlin 1994). Because it is inexpensive and widely available, it
is a common cause of poisoning, through accidental or voluntary Diagnosis of paraquat poisoning has evolved over time. Currently
ingestion. Paraquat poisoning is most prevalent in lower- and the only reliable biomarker for the condition is time-adjusted
middle-income countries (LMICs), where its use is less stringently paraquat concentration.
regulated.
The prognosis in paraquat poisoning is associated with the amount
Based on prevalence studies, Karunarathne 2020 estimates the of toxin ingested.
global death toll from pesticide poisoning between 1960 and 2018
to be between 9,859,667 and 17,303,333, and notes that this is likely • In low-dose poisoning (< 20 mg of paraquat ion per kilogram of
to be an underestimate. Paraquat is a principal cause of this fatal body weight), patients are often asymptomatic, or may develop
poisoning in many part of Asia and South America (Dawson 2010; vomiting or diarrhoea, but have a good chance of recovery.
Mew 2017). • In moderate-dose poisoning (20 mg to 40 mg of paraquat
ion per kilogram of body weight), initial renal and hepatic
The incidence of self-poisoning increases with age, and is higher in dysfunction is common. Mucosal damage may become apparent
males, although in some LMICs the incidence of self-poisoning for with sloughing of the mucous membranes in the mouth.
young adult females exceeds that of males (Gunnell 2003). Difficulty in breathing may develop after a few days in more
severe cases. After about 10 days, although renal function
The number of deaths due to poisoning, including those by often returns to normal, radiological signs of lung damage
paraquat, is now falling worldwide. This is likely to be a usually develop. Lung damage is usually followed by irreversible
combination of industrialisation and the consequent movement massive pulmonary fibrosis manifested by the progressive loss
of people out of rural areas, and the fact that many of the most of the lungs' ability to breathe, and deterioration continues until
dangerous pesticides, including paraquat, are being banned in an the patient eventually dies, between two and four weeks after
increasing number of countries (Karunarathne 2020). For example, ingestion.
paraquat was phased out and eventually banned in 2010 in Sri
• In high-dose poisoning (> 40 mg paraquat ion per kilogram
Lanka, where it had been the leading cause of death from pesticide
of body weight), toxicity is much more severe, and death
poisoning (Eddleston 2003; Knipe 2014). In addition, the overall
occurs early (within 24 h to 48 h) from multiple organ failure.
suicide rate by any means has also decreased markedly in China,
Vomiting and diarrhoea are severe, with considerable fluid
which has driven down global numbers (Mew 2017). However, this
loss. Renal failure, cardiac arrhythmias, coma, convulsions, and
global trend is unfortunately not reflected in all parts of the world,
oesophageal perforation lead to death (WHO IPCS 2009). This is
and the incidence of pesticide poisoning is increasing in some
known as fulminant poisoning (Vale 1987).
countries, such as Malaysia (Kamaruzaman 2019). Worldwide,
pesticide poisoning still represents a serious challenge to public Description of the intervention
health and accounts for an estimated 150,000 to 168,000 annual
deaths, representing around 20% of global suicides (Karunarathne There is no specific antidote for paraquat poisoning or standard
2020; Mew 2017). guidelines for treatment (Gawarammana 2011). Treatment often
involves decontamination using absorbents such as activated
Description of the condition charcoal or Fuller's earth; or elimination methods such as
haemodialysis, haemofiltration, or haemoperfusion; antioxidant
The mortality rate for paraquat poisoning is estimated to be
therapy; or immunosuppressive therapy (Lavergne 2018).
as high as 60% to 90%, but the exact mechanism of this
poisoning is still poorly understood (Gao 2020; Xu 2019). The Immunosuppressive therapy
damage done by paraquat is thought to be primarily due to
redox cycling and the subsequent generation of highly reactive Glucocorticoid drugs are usually used to reduce inflammation and
oxygen and nitrite species. The resultant oxidative stress results to suppress the immune response. Glucocorticoids are a class of
in mitochondrial toxicity, induced apoptosis, lipid peroxidation, corticosteroid hormones that bind to glucocorticoid receptors. The
and severe secondary inflammation. These mechanisms are activated glucocorticoid receptor-glucocorticoid complex leads to
thought to act synergistically to cause organ damage (Gao 2020; the up-regulation of the expression of anti-inflammatory proteins.
Gawarammana 2011). The glucocorticoids commonly used to treat paraquat poisoning
are methylprednisolone and dexamethasone.
Lung injury is one of the main features of paraquat poisoning:
paraquat molecules selectively accumulate in the lungs, and severe Cyclophosphamide is a broad-spectrum immunomodulatory drug,
inflammation and irreversible pulmonary fibrosis follows, which is usually used in the treatment of cancer and autoimmune disease.
known as 'paraquat lung' (Fukuda 1985; Smith 1975). This fibrosis
Since the 1970s, glucocorticoids and cyclophosphamide have been
leads to reduced ventilatory and lung diffusion capacity, resulting
used in combination as a means of suppressing the inflammation
in hypoxaemia, which is often lethal. A large proportion of patients
responsible for pulmonary fibrosis (Eddleston 2003), and were
appear asymptomatic until signs of breathing difficulty emerge; it
endorsed as a successful treatment by Addo and Poon-King (Addo
is difficult to predict the outcome of a patient who appears normal
1986). The effectiveness of this treatment combination for paraquat
but is actually suffering lung fibrosis (Eddleston 2003).
poisoning remains unclear.
Informed decisions.

How the intervention might work cyclophosphamide by themselves) or other combinations of

therapies.
In animal studies, glucocorticoids such as dexamethasone
have been shown to reduce lipid peroxidation and paraquat Types of outcome measures
accumulation in the lungs (Dinis-Oliveira 2006). Cyclophosphamide
has a broad immunomodulatory effect. • Mortality at 30 days following the ingestion of paraquat.
• All-cause mortality at the end of the maximum follow-up
Given the profound secondary inflammation seen in paraquat period recorded by investigators, as long as they are clinically
poisoning, particularly in the lungs where paraquat molecules compatible.
actively accumulate, immunosuppression would appear to be a
logical therapy for paraquat poisoning. Otherwise we reported findings in the short term (typically
at hospital discharge, which of necessity varied between
Why it is important to do this review participants), and longer term (three months after discharge in the
one trial that reported any follow-up), as appropriate.
Though it has been inferred from experimental, Lee 1984,
and clinical experience, Agarwal 2006, that immunosuppressive • New infections diagnosed within one week after initiation of
therapy might reduce deaths among paraquat-poisoned treatment.
patients, there is no consensus on the effectiveness of this
treatment. Considering the potential hazards associated with Search methods for identification of studies
immunosuppressive drugs (Winsett 2004) - such as making patients
more prone to infection - an update of the previous Cochrane In order to reduce publication and retrieval bias we did not restrict
Review on this topic was due in order to revisit assessments of our search by language, date, or publication status.
effectiveness and risk, support decision-making and inform further
Search strategies with notes for this update are listed in Appendix 1.
research.
Search methods and strategies for previous versions of the review
can be found in Appendix 2 and Appendix 3.
OBJECTIVES
Electronic searches
To assess the effects of glucocorticoid with cyclophosphamide for
moderate to severe oral paraquat poisoning. The Cochrane Injuries Group's Information Specialists searched:
METHODS • the Cochrane Central Register of Controlled Trials (CENTRAL)

(which contains the Cochrane Injuries Trials Register) in the
Criteria for considering studies for this review Cochrane Library (searched 21 September 2020);
Types of studies • Ovid MEDLINE(R), Ovid MEDLINE In-Process & Other Non-
Indexed Citations, Ovid MEDLINE Daily and Ovid OLDMEDLINE
We included randomised controlled trials (RCTs), including cluster- (1946 to 21 September 2020);
RCTs. We excluded any trials of a cross-over design as this is • Embase Classic + Embase (OvidSP) (1947 to 21 September 2020);
incompatible with our review question. For RCTs published after
• ISI Web of Science: Science Citation Index Expanded (SCI-
2010, we considered only prospectively registered RCTs to be
EXPANDED) (1970 to 21 September 2020);
eligible for this 2021 update, in accordance with Cochrane Injuries
Group policy. • ISI Web of Science: SCI-EXPANDED, SSCI, A&HCI, CPCI-S, CPCI-
SSH, ESCI (2017 to 22 September 2020)
Types of participants • US National Institutes of Health Ongoing Trials Register
Any person of any age, diagnosed with oral paraquat poisoning ClinicalTrials.gov (www.clinicaltrials.gov) (21 September 2020);
(although poisoning via inhalation or contact with skin is possible, • Current Controlled Trials (www.controlled-trials.com/) (20
these were not foci of the review). The review focuses on September 2020);
participants with moderate to severe poisoning, given that • World Health Organization International Clinical Trials Registry
generally in trials (as in clinical practice) individuals with mild Platform (apps.who.int/trialsearch/) (21 September 2020).
cases of paraquat poisoning (which tend to resolve on their own)
and those with extremely severe intoxication (see definition of The following Chinese databases were searched by one review
'fulminant' in both Description of the condition and Included author (LL):
• China National Knowledge Infrastructure (CNKI 数据库)

studies) may not be treated; or, if they are treated, may not be
analysed with other groups of participants, due to prognosis.
(12 November 2020);
Types of interventions • Wanfang Data (万⽅数据库) (12 November 2020);
• Intervention: glucocorticoid with cyclophosphamide, in • VIP (维普数据库) (12 November 2020).
combination. Searching other resources
• Eligible comparators: placebo; standard care alone; or any
alternative therapy in addition to standard care. We searched the internet through search engines google.com and
baidu.com on 12 November 2020 using the term 'clinical trial
We excluded studies that focused on any single & paraquat'. We also checked the reference lists of reports and
immunosuppressant (either a glucocorticoid drug or systematic/literature reviews on paraquat poisoning for potentially

Informed decisions.

relevant published or unpublished trials. We contacted the authors method of sequence generation. Review author CY moderated the
of the included trials for further information. discussion on inclusion of this trial, and in the last published
version of this review (Li 2014), it was agreed that the trial would
Data collection and analysis be included, but classified as being at high risk of bias. Following
statistical advice in 2021, the author team as a whole decided
Selection of studies
once again to retain the trial, whilst highlighting issues to do
For the present version of the review, three review authors (LL, with its interpretation, on the basis not only of design, but also
JD and HW) independently screened the search results from the due to investigator-acknowledged failure to assess participants via
English language databases. LL, CY and BC independently screened plasma concentration.
the results from the Chinese databases. We obtained and assessed
the full-text versions of potentially relevant trials. Duplicate reports We identified one new study that met our eligibility criteria
were identified and noted. (Gawarammana 2017). A flow chart of the study selection process
is shown in Figure 1.
Review authors LL and BC disagreed about the inclusion of the
Afzali 2008 study due to the use of alternate allocation as the

Figure 1. Study flow diagram for 2021 update

Informed decisions.

Data extraction and management Assessment of risk of bias in included studies

Three review authors (LL, BC, and JD) independently extracted the Four review authors (LL, BC, JD, and HW) independently evaluated
following data from the four included trials: the risk of bias for each included trial based on the following
domains: sequence generation, allocation concealment, blinding,
• study design; incomplete outcome data, selective reporting, and any other
• study setting (country, city, number of centres); sources of bias. We based our judgements on the criteria outlined
• inclusion and exclusion criteria for studies; in Chapter 8 of the Cochrane Handbook for Systematic Reviews
• number, gender, and severity of intoxication of participants in of Interventions (Higgins 2011). We assessed each domain as at
the intervention and control groups; low or high risk, or unclear risk if the information provided was
insufficient to permit a judgement on a particular area of bias, or
• outcome data including number of deaths and infection cases
if it was unclear in which direction a bias might lead. Risk of bias
assessed at each data collection point;
judgements are provided in the risk of bias tables in Characteristics
• loss to follow-up/withdrawals from analysis; of included studies, and summaries of the judgements are given in
• information on study conduct sufficient to assess risk of bias, Figure 2 and Figure 3.
including evidence of prospective trial registration for all studies
published from 2010 onwards.

Figure 2. Methodological quality graph: review authors' judgements about each methodological quality item
presented as percentages across all included studies
Random sequence generation (selection bias)

Allocation concealment (selection bias)
Blinding (performance bias and detection bias): All outcomes
Incomplete outcome data (attrition bias): All outcomes
Selective reporting (reporting bias)
Other bias
0% 25% 50% 75% 100%
Low risk of bias Unclear risk of bias High risk of bias


Informed decisions.

Figure 3. Methodological quality summary: review authors' judgements about each methodological quality item
for each included study
Blinding (performance bias and detection bias): All outcomes

Incomplete outcome data (attrition bias): All outcomes
Random sequence generation (selection bias)
Allocation concealment (selection bias)
Selective reporting (reporting bias)

Other bias
Afzali 2008 ? - + + ? -
Gawarammana 2017 + + + + + +
Lin 1999 + ? + + ? -
Lin 2006 + + + + ? +

Measures of treatment effect Unit of analysis issues
Where feasible, we calculated the risk ratio (RR) and 95% No unit of analysis issues arose within the current version of this
confidence interval (CI). We summarised data for all-cause review. We excluded cross-over studies as inappropriate for our
mortality at final follow-up in a meta-analysis using a fixed-effect review question (given our focus on mortality). We identified no
model. When assuming a 75% and 65% mortality rate for control cluster-RCTs and no trials with multiple eligible intervention (or
and treatment group respectively, with a level of significance of 5% comparator) arms.
and a power of 90%, we calculated that 439 participants would be
required in each arm of a study to detect the effect sufficiently. Should the issue of clustering arise in future updates, we plan to
use the methods detailed in the Cochrane Handbook for Systematic
Reviews of Interventions as applicable (Higgins 2021), based on

Informed decisions.

information provided by investigators, including: the number of Summary of findings and assessment of the certainty of the
clusters (or groups) randomised or the average (mean) size of evidence
each cluster; and an estimate of the intracluster (or intraclass)
We summarised our findings for all outcomes in Summary of
correlation coefficient (ICC).
findings 1.
Should the issue of multiple eligible arms or comparator groups
We assessed the certainty of the evidence for each outcome
arise, we will consider appropriate methods including pooling data
measure using the GRADE approach (Yan 2016). GRADE is used
from eligible arms, or conversely treating a multiple-arm study as
to assess the certainty of evidence according to the following
two or more studies. In the latter case, we will take care to reduce
five domains: imprecision, indirectness, inconsistency, risk of bias,
the number of participants in the arm(s) left intact, in order to
and publication bias. The certainty of evidence assessments are
preserve the independence of findings.
provided in Summary of findings 1 and were generated within the
Dealing with missing data online tool GRADEpro GDT (GRADEpro GDT 2020).
The amount of missing data in each trial was assessed, outcome by RESULTS
outcome, and any potential effects on results informed our risk of
bias assessment (see Risk of bias in included studies) and GRADE Description of studies
ratings.
Results of the search
Assessment of heterogeneity The study selection process is outlined in Figure 1.
We considered possible causes of clinical heterogeneity (e.g.
The English language electronic search retrieved a total of 2813
how studies measured severity of poisoning, treatment regimen,
records across all years. We identified eight potentially relevant
variation in standard care, length of follow-up) when making
trials, four of which were eligible for inclusion in this update.
decisions about whether to pool data.
Review author LL identified a total of 1010 reports through a
Where possible and appropriate, we examined statistical
Chinese language search on Chinese language databases; none of
heterogeneity using the Chi2 test and I2 statistic. For the outcome
these met our inclusion criteria.
of mortality, we considered an I2 value of more than 50% as
indicative of substantial heterogeneity, and acknowledged this in Review author LL identified one report whilst searching google.com
our interpretation of the data. using the term 'clinical trial & paraquat', which was later excluded
(Tsai 2009). We identified no additional eligible RCTs through
Assessment of reporting biases screening reference lists or literature reviews.
Had we identified 10 or more eligible trials for this update, we would
have investigated the possibility of reporting biases, including Included studies
publication bias, by assessing funnel plots for asymmetry where In this update, we included one new study (Gawarammana 2017),
10 or more studies reported on the same outcome (Egger 1997). which resulted in a total of four included studies, with a combined
We acknowledge that asymmetry could be due to publication bias total of 463 participants for analysis (Afzali 2008; Gawarammana
or to a genuine relationship between trial size and effect size. We 2017; Lin 1999; Lin 2006). See also Characteristics of included
also searched for evidence of prospective trial registration and trial studies.
protocols for all studies included in the review.
Design and setting
Data synthesis
All of the included studies were reported as parallel RCTs; however,
Where feasible, we analysed data using Review Manager 5 (Review following contact with one trialist, it became clear that one study
Manager 2020). Where this was not feasible, we reported results used quasi-randomised methods to generate the sequence (Afzali
narratively. 2008).
Subgroup analysis and investigation of heterogeneity Two studies were conducted at a single site (a hospital in Taiwan,
Had there been sufficient studies, we would have performed which featured a dedicated poison control centre) (Lin 1999; Lin
subgroup analysis based on the treatment regimen used to 2006); one study took place at single site in Hamadan, Iran; and one
investigate the impact on both mortality and adverse effects. study was a multisite RCT conducted at six district hospitals in Sri
Lanka (Gawarammana 2017).
Sensitivity analysis
Sample sizes
We performed post hoc sensitivity analyses to investigate the effect
of reporting bias in the Lin 1999 study, as well as on the effect No sample size calculation is mentioned in the text of Lin
of possible selection bias in the Afzali 2008 study (see Effects 1999, which reports data for 121 participants but retrospectively
of interventions). These two studies also merited exclusion from excluded those with fulminant poisoning who died within one week
sensitivity analyses because they were subject to bias based on lack of intoxication (n = 71), leaving 50 participants. The same trialists
of appropriate assessment of plasma paraquat levels on entry to reported for their later, smaller study (n = 23), which assessed a
the study, leading ultimately to our decision to present a sensitivity different treatment regimen, that an "a priori estimate of sample
analysis for the outcome 'all-cause mortality at final follow-up: at size for this study could not be performed because only case
hospital discharge' in Summary of findings 1. reports using the new treatment method were noted in previous

Informed decisions.

literature" (Lin 2006). No sample size calculation is mentioned in review. The authors of the Lin 1999 paper - the oldest in the review
the text of Afzali 2008 (n = 20). Investigators in Gawarammana - defended their choice to analyse as they did by explaining that
2017 (n = 299) did provide a sample size calculation, reporting: they were constrained at the time of recruitment "because no paper
"In order to be able to detect whether either regimen increases suggested an index to accurately predict the clinical severity of PQ
survival from 18% to 28%, with a significance level (alpha) of [paraquat]-poisoned patients up till now" (Buckley 2001).
5% and a power of 80%, a minimum of 295 patients must be
recruited to each arm of the trial (i.e. 590 patients in total)" (p 634). Demographics
This recruitment objective was manifestly unmet due to the trial Although inclusion criteria, where specified, admitted the
stopping early, which was sanctioned by the Data Monitoring and recruitment of adolescents from the age of 14, Gawarammana 2017,
Ethics Committee because of a "collapse" in recruitment due firstly or 15, Lin 2006, the included participants tended to be adults, with
to restrictions on, and subsequently the banning of, paraquat use means of 33 to 37 years of age in Lin 2006, 26 in Afzali 2008, 27
in Sri Lanka. in Gawarammana 2017, and 27 to 37 in the two groups reported
within Lin 1999. The majority of participants across trials (69%)
Participants were male.
Inclusion criteria
Diagnostic certainty/severity
All four included trials involved paraquat-intoxicated individuals
who had had their paraquat poisoning confirmed by sodium Two of the four included studies used the Severity Index of Paraquat
dithionite test. The included trials varied in the time period during Poisoning (SIPP), defined as plasma paraquat poisoning in mg/L
which participants could be recruited. In both Lin 1999 and Lin multiplied by the number of hours since paraquat ingestion, as
2006, participants had to present within 24 hours of ingestion a means of assessing severity on arrival to hospital (Sawada
of paraquat. No explicit criteria were given for presentation in 1988). By this measure, participants were comparable at baseline
the trial reported in Afzali 2008, but figures provided for mean within Lin 2006 (mean values between 14 and 16), but investigators
time between poisoning and admission to hospital never exceeded in Gawarammana 2017 reported a small but significant baseline
seven hours. The largest and most recent trial reported inclusion difference in participants allocated to intervention (mean 18.4)
criteria permitting recruitment up to 48 hours after ingestion compared to control (mean 13.1). Lin 1999, Lin 2006, and Afzali
(Gawarammana 2017). 2008 also reported urine colour (whether "dark blue" or "navy
blue") and found groups comparable, although there is consensus
As mentioned in Description of the condition and Types of in the field that only plasma tests are truly reliable. The authors of
participants, symptoms of paraquat ingestion are dose-dependent, Afzali 2008 state lack of such testing in the "limitations" section of
and intoxication is usually categorised as mild or 'low dose', their paper; the authors of Lin 1999 did not, but they included such
moderate, or severe or 'high dose'. Mild cases tend to resolve testing in their subsequent study (Lin 2006).
without further sequelae and therefore trialists typically exclude
such individuals when conducting trials of therapeutics. Similarly, Interventions
where they can be identified, 'fulminant' participants who have All studies compared the use of standard care alone versus
ingested a dose likely to kill within 48 to 72 hours due to multiple standard care and glucocorticoid (invariably methylprednisolone
organ failure are also typically excluded, if possible. Palliative care or dexamethasone, or both) with cyclophosphamide for individuals
tends to be the recommended course for people in this category. with paraquat poisoning. Gawarammana 2017 also provided
Methods for assessing degree of intoxication evolved between normal saline as a placebo within the control group, which no other
the conduct of the oldest study in this review, Lin 1999, and the trial did.
latest study, Gawarammana 2017. We considered these differential
criteria for defining severely poisoned patients to be a potential Standard care
source of clinical heterogeneity, and this was reflected in sensitivity Standard care given to both groups was defined in Lin
analysis for mortality and in our GRADE assessments. 1999 and Lin 2006 as including gastric lavage with normal saline
All of the included studies attempted to exclude severely poisoned followed by active charcoal added in magnesium citrate given
patients who were expected to die imminently with little chance through a nasogastric tube, courses of 8-hour active charcoal
of responding to therapy: Gawarammana 2017 excluded severely haemoperfusion therapy and administration of intravenous
poisoned patients with a Glasgow Coma Score of less than 8/15 or dexamethasone. "Conventional treatment" within the small Afzali
a systolic blood pressure less than 70 mmHg that did not respond 2008 study was similar, including gastric lavage with normal saline,
to 1 L of intravenous fluid; Afzali 2008 only provided analysis "charcoal-sorbitol lavage every two to four hours for three days,
for 20 of 45 participants screened for the trial, apparently not forced alkalinised diuresis in the first day of admission to the
randomising 15 participants whose sodium dithionite reactions hospital, and haemodialysis of four hours duration". This differed
tests indicated their cases were too mild, and 10 participants from Gawarammana 2017, where standard care was limited to
considered fulminant; they specifically listed their failure to intravenous fluid, activated charcoal, and pain relief.
conduct plasma assessments as a study limitation. Treatment regimens
The Lin 2006 trial only included participants with a predicted The combinations of cyclophosphamide and glucocorticoid
mortality of > 50% and ≤ 90% according to the Hart treatments in the four studies used pulse therapy administration
1984 formula. Lin 1999 excluded all participants who died within (high doses daily over a short period of time, to maximise
one week of poisoning in what appears to be a post hoc decision therapeutic effect whilst minimising toxicity).
(see Risk of bias in included studies), but which in fact leads to
a population broadly in line with other studies included in this
Informed decisions.

Following gastric lavage, active charcoal and charcoal Risk of bias in included studies
haemoperfusion therapy, the treatment arm in Lin 1999 received
infusions of cyclophosphamide for two hours per day for two We assessed Lin 2006 to be at low risk of bias across most
days, and methylprednisolone for two hours per day for three domains. Lin 1999 randomised all urine-positive patients, but
days; 10 mg dexamethasone was then administered every eight presented the outcomes for those who died within one week of
hours for 14 days. The treatment arm in Gawarammana 2017 was poisoning separately from those who survived longer. Presenting
similar, except that the methylprednisolone was infused over the data separately to exclude very severely poisoned people
one hour rather than two, and participants were given 8 is reasonable given the specific clinical features of paraquat
mg dexamethasone daily for 14 days. The authors of Afzali poisoning, but this post hoc decision introduces the risk of
2008 reported a similar regimen (but without administration of selective reporting. The small trial conducted by Afzali 2008 was
dexamethasone), with cyclophosphamide infused in two hours poorly reported, but personal communication with the authors
for two days with methylprednisolone also infused for four permitted some judgements with respect to risk of bias, which
hours, repeated over three consecutive days. Participants in the was frequently assessed as high risk because of the method of
Afzali 2008 and Gawarammana 2017 trials also received MESNA sequence generation. Gawarammana 2017 was a well-conducted,
(2-mercaptoethane sulfonate sodium (Na)) to mitigate the effects preregistered multicentre RCT that was underpowered due to
of cyclophosphamide. stopping early. Our risk of bias judgements are recorded in the risk
of bias tables in Characteristics of included studies and displayed
Lin 2006 had a more complex treatment regimen. Following in Figure 2 and Figure 3.
the same initial pulse therapy of cyclophosphamide and
methylprednisolone described in Lin 1999, 5 mg dexamethasone Allocation
was administered every six hours until the participant's partial Lin 2006 reported an appropriate method of sequence generation
pressure of oxygen (PaO2) reached a certain threshold. Another and allocation concealment using a sequence of labelled cards
pulse of methylprednisolone and cyclophosphamide (the latter in sealed envelopes that were prepared by a statistical advisor.
only for one day, and only if it was more than two weeks since the Gawarammana 2017 used computer programs to generate the
previous cyclophosphamide dose) was repeated if physiological random sequence by an IT consultant, and the allocations were
parameters fell below a certain threshold, indicating a poor conducted by the pharmacist and concealed from other members
outcome. Dexamethasone was then continued until the desired of the team. Lin 1999 generated the randomisation sequence using
PaO2 was reached. a random numbers table, but there was no mention of allocation
concealment, so we judged this to be at unclear risk of bias.
Outcomes
All of the included studies reported the primary outcome, Afzali 2008 used alternate allocation as a method of sequence
mortality, at hospital discharge (Afzali 2008; Gawarammana 2017; generation, which obviates concealment of allocation and under
Lin 1999; Lin 2006), and in one case, at three months after normal circumstances would call for an assessment of high risk
hospital discharge (Gawarammana 2017). No study reported on the of bias and exclusion of the study in sensitivity analysis, if the
outcome of infections (a known sequela of immunosuppression trial were to be included in the review at all. After discussion with
therapy), although two studies considered the related outcome the Cochrane Injuries statistician, we revised our assessment to
of leukopenia (Lin 1999; Lin 2006). Afzali 2008 did not report on unclear risk, given that the pace of recruitment (just 20 eligible
infections or on adverse events per se, but did confirm that no participants across 25 months) reduced the risk of manipulation of
complications of treatment appeared by the time participants the allocation of any given participant to a particular group.
had been discharged. The fourth trial assessed participants for
Blinding
two specific potential adverse effects of treatment (haematuria,
bladder pain) attributable to cyclophosphamide (Gawarammana In Gawarammana 2017, the pharmacist allocated participants
2017). and prepared identical treatment packs of both active treatment
and placebo, protecting team members and participants from
Study funding sources awareness of allocation.
Two studies did not report any source of funding (Afzali 2008; Lin
In Lin 1999 and Lin 2006, the statistician who contributed to the
1999); one reported funding from a governmental body in Taiwan
trial report was blinded to the allocation. Treating physicians and
(Lin 2006). One study reported mixed sources of funding including
participants were not blinded, nor could they be, as there was no
charitable and research grants from the UK and Australia as well
placebo. Given that the primary outcome is death, this may not
as support from Syngenta, a manufacturer of herbicides including
constitute a significant risk of bias. We therefore judged the risk of
paraquat (Gawarammana 2017).
bias for blinding as low in this version of the review, instead of high
Excluded studies as in previous versions, but considered the lack of blinding to be an
issue when performing the GRADE assessment for the outcome of
We excluded four trials: one trial with uncertain sequence infection.
generation that assessed the effects of methylprednisolone only
(Tsai 2009); one trial that used a historical control (Perriens 1992); Investigators involved in the Afzali 2008 trial confirmed by personal
and two potentially eligible studies that were not preregistered as correspondence that no blinding was attempted of any of the staff
required by the Cochrane Injuries Group for studies published after involved in the trial (nor could they be, with no placebo in use).
the year 2010 (Chen 2014; Ghorbani 2015). As with Lin 1999 and Lin 2006, we believe this does not constitute
a high risk of bias for the objective outcome used in our review
(mortality at discharge).

Informed decisions.

Incomplete outcome data Afzali 2008 was a small study reported in a very short paper, with
virtually no information on study conduct and a lack of clarity on
Our main outcome of interest was mortality, which was reported
the screening and allocation process. Some relevant information
in full in all studies either at hospital discharge or at six weeks.
was acquired by personal correspondence to which an author on a
In Gawarammana 2017, the primary result, in-hospital death, was
previous version of this review has lost access. In view of this, and
reported for all 299 randomised participants, as well as follow-up
in the absence of a prospective registration or a trial protocol, and
at three months after hospital discharge, by which time only six
in compliance with the Cochrane Injuries Group policy, we must
participants had been lost to follow-up, two out of 152 (1.3%) from
assess the risk of selective outcome reporting as at best unclear
the control group and four out of 147 (2.7%) from the treatment
whenever no prospective plans are made public.
group. Given the high risk of death (65% in treatment group and
75% in control group), we judged the potential impact of the loss of Other potential sources of bias
outcome data on the risk of attrition bias to be small.
The Gawarammana 2017 trial was stopped early "after consultation
Selective reporting with the data monitoring and ethics committee due to a collapse
in recruitment" following the phasing out and eventual banning
We identified no suggestion of selective reporting in Gawarammana
of paraquat in Sri Lanka, where the trial was being conducted.
2017, as it was prospectively registered and appears to have
However, since the reason for stopping early was unrelated to the
reported methods and data in full as planned, up to the time of
observed intervention effect, we do not deem this to be a source
stopping prematurely (see below).
of bias. The same trialists reported a small baseline difference in
Lin 2006 appears to have reported on all relevant outcomes, but the median SIPP score of participants allocated to intervention
in the absence of a prospective registration or a trial protocol, and (median 18.4) compared to control (median 13.1). As the method
in compliance with the Cochrane Injuries Group policy (2015), we of randomisation was unlikely to have been compromised (it was
must assess the risk of selective outcome reporting as (at best) done using purpose-designed software), this difference is likely to
unclear whenever no prospective plans are made public. have been caused by chance alone, and is therefore not considered
to be a source of bias.
Lin 1999 randomised 121 participants, but made what appears to
be a post hoc decision to exclude 71 of these participants following Effects of interventions
treatment, based on the very severe nature of their poisoning. See: Summary of findings 1 Glucocorticoid with
Although the clinical decision to exclude these fulminant patients cyclophosphamide plus usual care compared to usual care (with or
seems sensible given the predictable course of very severe without placebo) for oral paraquat poisoning
paraquat poisoning - which was done in all four included studies
- the criteria by which these patients were defined (death within Mortality at 30 days following the ingestion of paraquat
a week) and the post hoc nature of the decision leaves the study
open to reporting bias. Nevertheless, we believe we have obviated None of the included studies assessed this outcome.
the risk of bias from selective reporting by presenting data for both
All-cause mortality at the end of the follow-up period
groups and by using sensitivity analysis. However, as there is no
evidence of prospective reporting and no identifiable protocol, we See Figure 4; Figure 5; Figure 6
are again obliged to report the risk of bias for this domain to be
unclear.


Informed decisions.

Figure 4. Forest plot of comparison: 1 All-cause mortality, outcome: 1.1 All-cause mortality
Glucoc & cyclophosphamide Standard care Risk Ratio Risk Ratio

Study or Subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI
1.1.1 Mortality at hospital discharge

Afzali 2008 3 9 9 11 6.6% 0.41 [0.16 , 1.07]
Gawarammana 2017 78 147 94 152 75.2% 0.86 [0.70 , 1.04]
Lin 1999 4 22 16 28 11.5% 0.32 [0.12 , 0.82]
Lin 2006 5 16 6 7 6.8% 0.36 [0.17 , 0.80]
Subtotal (95% CI) 194 198 100.0% 0.73 [0.61 , 0.88]
Total events: 90 125
Heterogeneity: Chi² = 9.92, df = 3 (P = 0.02); I² = 70%
Test for overall effect: Z = 3.29 (P = 0.001)
1.1.2 Mortality at 3 months

Gawarammana 2017 101 143 108 150 100.0% 0.98 [0.85 , 1.13]
Subtotal (95% CI) 143 150 100.0% 0.98 [0.85 , 1.13]
Heterogeneity: Not applicable
Test for subgroup differences: Chi² = 5.88, df = 1 (P = 0.02), I² = 83.0% 0.2 0.5 1 2 5
Favours glucocorticoid with cyclophosphamide Favours control

Figure 5. Forest plot of comparison: 1.2 Sensitivity analysis: all-cause mortality at final follow-up: including
fulminant (and excluding Afzali 2008 due to risk of selection bias)


Afzali 2008 3 9 9 11 5.1% 0.41 [0.16 , 1.07]
Gawarammana 2017 78 147 94 152 58.5% 0.86 [0.70 , 1.04]
Lin 1999 38 56 53 65 31.1% 0.83 [0.67 , 1.03]
Lin 2006 5 16 6 7 5.3% 0.36 [0.17 , 0.80]
Subtotal (95% CI) 228 235 100.0% 0.80 [0.69 , 0.93]
Test for subgroup differences: Not applicable 0.2 0.5 1 2 5



Informed decisions.

Figure 6. Forest plot of comparison: 1 All-cause mortality, outcome: 1.3 Sensitivity analysis: all-cause mortality:
excluding studies without plasma paraquat assessment at baseline


Gawarammana 2017 78 147 94 152 91.7% 0.86 [0.70 , 1.04]
Lin 2006 5 16 6 7 8.3% 0.36 [0.17 , 0.80]
Subtotal (95% CI) 163 159 100.0% 0.82 [0.68 , 0.99]


All-cause mortality (in hospital) received standard care alone (RR 0.98, 95% CI 0.85 to 1.13) (Analysis
1.1). We downgraded the certainty of the evidence for imprecision
The primary analysis suggests that participants who receive
(the total number of participants across the included studies fell
glucocorticoids with cyclophosphamide in addition to standard
short of the optimal information size) and for indirectness (which
care may have lower mortality in hospital than those who received
we defined as uncertainty with regard to the interpretation of long-
standard care alone (risk ratio (RR) 0.73, 95% CI 0.61 to 0.88; 4
term mortality data in a population where injury is sustained as a
studies; 392 participants). Statistical heterogeneity for this result
result of self-harm).
was substantial (I2 = 70%) (Analysis 1.1). There was also some
clinical heterogeneity relating to participant inclusion criteria: Lin New infection diagnosed within one week after initiation of
1999 randomised 121 participants, but made what appears to be a treatment
post hoc decision to exclude 71 following treatment, given that they
died within a week due to the very severe nature of their poisoning. Two of the four included trials assessed the outcome of infection
(Lin 1999; Lin 2006). Neither study reported any new infections
We thus had serious concerns about the possibility of selection diagnosed within one week after initiation of methylprednisolone
bias, and to investigate, we performed an exploratory post hoc and cyclophosphamide. Lin 1999 reported the related outcome of
sensitivity analysis to re-include the 71 fulminant participants from leukopenia in 8 of 22 participants (36.4%) in the intervention arm.
Lin 1999. With the fulminant participants included, the estimated These participants spontaneously recovered one week later, with
effect of glucocorticoid with cyclophosphamide in addition to no mortality. Lin 2006 reported leukopenia in 6 of 16 participants
standard care was reduced to a less clear effect (RR 0.80, 95% CI (37.4%) in the intervention arm, all of whom recovered within one
0.69 to 0.93; 463 participants). (Analysis 1.2). This was expected, to two weeks. There were no reports of leukopenia in the control
given that fulminant patients will almost inevitably die irrespective groups.
of treatment (WHO IPCS 2009; Gawarammana 2017). Heterogeneity
remained substantial (I2 = 53%). Afzali 2008 and Gawarammana 2017 measured neither infection nor
leukopenia. As this outcome was reported by only two small trials,
We also elected to perform a second previously unplanned and since we have concerns about reporting bias in the larger of the
sensitivity analysis, excluding all participants from both the Lin two trials, we downgraded for serious imprecision and risk of bias
1999 and Afzali 2008 trials, on the grounds that they stood out for this outcome, assessing the evidence to be of very low certainty.
amongst the included trials by not using SIPP methods at baseline.
The results left two studies with combined results suggesting that DISCUSSION
there may be a small benefit of the intervention on the outcome of
mortality (RR 0.82, 95% CI 0.68 to 0.99; participants = 322) (Analysis
Summary of main results
1.3). The I2 of 77% indicates that only the smaller study (Lin 2006, This systematic review update includes four trials with a combined
n = 23) with just seven participants in the control group showed total of 463 participants who had moderate to severe paraquat
benefit, whilst the larger trial (Gawarammana 2017) found none. poisoning. Low-certainty evidence from two trials indicates that
We chose to present the latter results as our main analysis for the participants who received glucocorticoids with cyclophosphamide
purposes of GRADE, and judged the certainty of the evidence to be in addition to standard care may have a slightly lower risk of death
low, downgrading twice for imprecision and inconsistency (Analysis than those receiving standard care alone in the short term (at
1.3; Summary of findings 1). hospital discharge). We chose the results of a sensitivity analysis
as the main analysis for this outcome, conceding that the lack
All-cause mortality (long term - three months after discharge of plasma testing at baseline rendered the results of two studies
from hospital) unreliable. Heterogeneity was high, as one small study found a
Low-certainty evidence suggests that there may be little or no large benefit whilst the other (the largest and best conducted within
difference in mortality at three months after discharge from the review) found none. The same single, large study provided low-
hospital between participants who received glucocorticoids with certainty evidence of little or no effect of treatment on mortality at
cyclophosphamide in addition to standard care and those who three-month follow-up after discharge from hospital.

Informed decisions.

We are uncertain about the effects of glucocorticoids with Our meta-analyses revealed substantial statistical heterogeneity,
cyclophosphamide in addition to standard care on infection risk therefore we downgraded the certainty of evidence accordingly.
within one week of treatment. Two trials recorded leukopenia in There were several possible sources for this heterogeneity, as
just over one-third of participants in the treatment arm, all of whom follows.
recovered within two weeks and none progressed to a diagnosed
infection. We assessed the evidence for this outcome to be of very • Differing inclusion criteria regarding severity: Lin 1999 did
low certainty. not employ plasma testing, and also retrospectively excluded
participants who died within a week. Lin 2006 only included
Overall completeness and applicability of evidence participants with a predicted mortality of > 50% and ≤ 90%
according to the Hart 1984 formula. Afzali 2008 did not use
The completeness of the evidence for the effects of glucocorticoid SIPP values and provided limited detail on inclusion criteria
with cyclophosphamide for paraquat-poisoned patients is limited other than that "mild" and "fulminant" cases were excluded.
by the fact that we were only able to include four RCTs, three of Gawarammana 2017 excluded severely poisoned people with
which were small, and the largest of which was still underpowered a Glasgow Coma Score of less than 8/15 or a systolic blood
due to trial termination. Differing inclusion criteria regarding pressure less than 70 mmHg.
severity of paraquat poisoning, leading to concerns about clinical
• Variation between treatment arms across trials: dosages and
heterogeneity, was also an issue (see Quality of the evidence).
infusion times varied; the Afzali 2008 regimen did not include
With regard to representation of populations most affected by dexamethasone, unlike the other three studies.
this condition, trials are relatively restricted, having only been • Variation between standard care arms across trials:
conducted in China (two studies), Iran (one study), and Sri Lanka Gawarammana 2017 limited standard care to intravenous
(one study), and the context of recruitment may well have changed fluid, activated charcoal, and pain relief, whilst Lin 1999
since the time when these studies were conducted (range 1997 and Lin 2006 included gastric lavage with normal saline
to 2010). The mean age of participants was in the upper 20s, and followed by active charcoal added in magnesium citrate given
the vast majority of participants were male. Since poisoning as a through a nasogastric tube, courses of 8-hour active charcoal
means of attempting suicide is most prevalent in older age groups haemoperfusion therapy and administration of intravenous
in many countries (e.g. China and Korea), this review is limited in dexamethasone. It is possible that there was a synergistic
its applicability to this population (Wang 2019). This is particularly interaction between one of these elements of standard care
important since age can increase physiological susceptibility to the in the Lin trials, when combined with glucocorticoids and
effects of pesticides (Ginsberg 2005). cyclophosphamide, as suggested by Xu 2019.
One consideration of our results is that this effect estimate applies The sample size for this review fell short of the 429 participants
only to moderate to severe cases of poisoning: mild cases with a required in each arm of the study, therefore we downgraded the
negative urine dithionite test were excluded, as were very severe certainty of evidence to reflect this imprecision.
cases. This makes clinical sense, as mild cases tend to make a
complete recovery with only mild symptoms, whilst very severe Potential biases in the review process
cases of fulminant poisoning are expected to die imminently and This review was conducted according to predefined inclusion
have little chance of responding to therapy (WHO IPCS 2009). criteria and methodology to select and appraise eligible studies.
The search for trials was extensive, and was conducted on both
Quality of the evidence
English and Chinese language databases. Publication bias is a
We judged two of the four included studies to be at low risk consideration in any systematic review. Although only four trials
of bias for key domains, including the largest of the three trials were included in the review, we believe that given the extent of
(Gawarammana 2017, n = 299). Although three trials, Afzali 2008, the search for trials, these were the only eligible RCTs addressing
Lin 1999 and Lin 2006, did not use a placebo, we did not consider this research question at the time of the search. We excluded
this to be a major source of bias since our main outcome of interest two studies because they were not prospectively registered and
was mortality. We assessed Lin 1999 to be at unclear risk of bias therefore did not meet the Cochrane Injuries Group inclusion
for allocation concealment, as this was not mentioned in the trial criteria for studies conducted after the year 2010.
report. This study randomised 121 participants regardless of the
severity of poisoning, and data on mortality were presented in full. We differed from the authors of previous updates and used the
However, the trialists made what appears to be a post hoc decision data as analysed in Lin 1999, in which the trialists, following
to exclude very severely poisoned participants ("fulminant", n = completion of the trial, excluded participants with very severe
71) from the final analysis. We performed a sensitivity analysis to fulminant poisoning from their analysis. Whilst we acknowledge
investigate the effect of including these participants in the final that this introduces a high risk of selection bias, we considered
analysis. The effect estimate was reduced, as would be expected that it made better sense clinically to exclude these participants,
given that people with fulminant paraquat poisoning are likely to since people with fulminant poisoning are very unlikely to respond
die imminently irrespective of treatment. The Afzali 2008 trial was to any therapy, and the other studies also excluded these
small and poorly reported, but considering that its use of alternate participants. We accounted for this high risk of selection bias in
allocation was unlikely to seriously bias results given the slow our GRADE assessment, and undertook sensitivity analysis. If there
pace of recruitment, we included data from this study. We further are sufficient data in future updates of this review, we will perform
conducted sensitivity analysis to consider the impact of diagnostic subgroup analysis according to severity of poisoning.
uncertainty, as noted below.

Informed decisions.

Agreements and disagreements with other studies or effect on mortality in people with moderate to severe paraquat
reviews poisoning during hospitalisation, however, there appears to be
little or no effect at three months post hospital discharge. Our
The authors of the largest and best-conducted study included in confidence in the evidence is limited by the small number of
this review, Gawarammana 2017, appear to be sceptical of any participants contributing reliable data and the heterogeneity of
real benefit of the treatment regimen they studied. Based on results across studies. We are uncertain about the potential of
their experience and their assessment of other literature in the glucocorticoid with cyclophosphamide to increase the risk of
field, they concluded that it is likely that "any possible benefit" infection due to limited evidence.
of the glucocortoid/cyclophosphamide combined treatment "is
due to the dexamethasone component rather than high-dose Implications for research
immunosuppression. We believe clinical research efforts would be
best spent on exploring the optimal dose of dexamethasone and To enable further study of the effects of glucocorticoid with
other inexpensive and low toxicity antidotes with favourable effects cyclophosphamide for individuals with moderate to severe
in animal studies (for example acetylcysteine)" (Gawarammana paraquat poisoning, hospitals may provide this treatment
2017, p 639, emphasis added). as part of a randomised controlled trial with appropriate
allocation concealment. All future research should be prospectively
This is in considerable contrast to the findings of two recently registered, CONSORT-compliant, and investigators should make
published systematic reviews, whose authors conclude that rigorous efforts to ensure an adequate sample size as well as seek
"immunosuppressive pulse therapy can efficiently reduce the long-term follow-up of participants, including on adverse events.
mortality of PQ [paraquat] poisoning and is relatively safe" (Xu In research contexts, no participant should be recruited without
2019, p 588), and that immunosuppressive drugs generally "may an assessment of plasma for time-adjusted paraquat concentration
reduce the mortality and incidence rate of MODS [multiple organ being conducted.
dysfunction syndrome] in moderate to severe PQ poisoning
patients, and severe PQ poisoning patients might benefit more Since the drafting of the protocol for this review in 2009, other
from ISDs [immunosuppressant drugs]" (Gao 2020). However, both trials have been conducted to assess the effects of glucocorticoids
systematic reviews had markedly different inclusion criteria to our in combination with other treatments, such as haemoperfusion.
review, including results from many non-randomised studies, as Use of other treatment options such as therapy with high-dose,
well as one recent RCT that we excluded because there was no long-term antioxidant free radicals is also increasing. We aim in
evidence of preregistration. Furthermore, Xu 2019 unaccountably future updates to conduct a network meta-analysis in an effort to
omits results from the largest study included within this review determine the most effective therapy - or combination of therapies
(Gawarammana 2017). - for paraquat poisoning.
The findings of our review align more closely with those of ACKNOWLEDGEMENTS
Gawarammana 2017 itself, for obvious reasons. We believe
We thank Karen Blackhall for providing the search strategy and
their trial, and our review, provide low-certainty evidence that
the search of English language databases for previous versions
participants who received glucocorticoid with cyclophosphamide
of this review. We thank Xi Lv for searching Chinese language
in addition to standard care may have little reduction in mortality
databases in 2012. We thank Iris Gordon for help with search
compared to those receiving standard care alone at hospital
processing with English language databases. We thank Marialena
discharge, and little or no effect at three months after discharge
Trivella for statistical peer review, and Elizabeth Royle for managing
from hospital. We considered longer-term data as more equivocal
the editorial process. We thank peer reviewers Dr James Coulson
than shorter-term data, believing it to be possible that the special
(Cardiff University/Cardiff & Vale University Health Board, UK), Prof
circumstances of this type of injury (paraquat is normally ingested
Narcisse Elenga (University of French West Indies/Cayenne General
in quantities seen in the included studies as a method of self-
Hospital, French Guiana), Harrison Davies, and Danial Sayyad for
poisoning) render any interpretation of longer-term data difficult,
their constructive comments, which we believe improved this
as factors other than the intervention may have a greater impact on
manuscript profoundly.
the long-term well-being and survival of patients.
This project was supported by the UK National Institute for Health
AUTHORS' CONCLUSIONS
Research, through Cochrane Infrastructure funding to the Cochrane
Implications for practice Injuries Group. The views and opinions expressed are those of
the authors and do not necessarily reflect those of the Systematic
Low-certaintly evidence indicates that glucocorticoid with Reviews Programme, NIHR, NHS or the Department of Health.
cyclophosphamide in addition to standard care may have a small

Informed decisions.

REFERENCES

References to studies included in this review Additional references
Afzali 2008 {published data only} Addo 1986
Afzali S, Gholyaf M. The effectiveness of combined treatment Addo E, Poon-King T. Leucocyte suppression in treatment of 72
with methylprednisolone and cyclophosphamide in patients with paraquat poisoning. Lancet 1986;1:1117–20.
oral paraquat poisoning. Archives of Iranian Medicine
2008;11(4):387-91. Agarwal 2006
Agarwal R, Srinivas R, Aggarwal AN, Gupta D. Experience with
Gawarammana 2017 {published and unpublished data} paraquat poisoning in a respiratory intensive care unit in North
Gawarammana I, Buckley NA, Mohamed F, Naser K, India. Singapore Medical Journal 2006;47(12):1037.
Jeganathan K, Ariyananada PL, et al. High-dose
immunosuppression to prevent death after paraquat Buckley 2001
self-poisoning - a randomised controlled trial. Clinical Buckley N. Letter to Editor: Pulse corticosteroids and
Toxicology 2017;56(7):633-9. [CTG: ISRCTN85372848] [DOI: cyclophosphamide in paraquat poisoning (includes response by
10.1080/15563650.2017.1394465] authors Lin 1999). American Journal of Respiratory and Critical
Care Medicine 2001;163:585.
Lin 1999 {published data only}
Lin JL, Leu ML, Liu YC, Chen GH. A prospective clinical trial Dawson 2010
of pulse therapy with glucocorticoid and cyclophosphamide Dawson H, Eddleston M, Senarathna L, Mohamed F,
in moderate to severe paraquat-poisoned patients. Gawarammana I, Bowe SJ, et al. Acute human lethal toxicity
American Journal of Respiratory and Critical Care Medicine of agricultural pesticides: a prospective cohort study. PLOS
1999;159(2):357-60. Medicine 2010;7(10):e1000357.
Lin 2006 {published data only} Dinis-Oliveira 2006

Lin JL, Lin-Tan DT, Chen KH, Huang WH. Repeated pulse of Dinis-Oliveira RJ, Duarte JA, Remião F, Sánchez-Navarro A,
methylprednisolone and cyclophosphamide with continuous Bastos ML, Carvalho F. Single high dose dexamethasone
dexamethasone therapy for patients with severe paraquat treatment decreases the pathological score and increases
poisoning. Critical Care Medicine 2006;34(2):368-73. the survival rate of paraquat-intoxicated rats. Toxicology
2006;227(1-2):73-85.

References to studies excluded from this review Eddleston 2003
Chen 2014 {published data only} Eddleston M, Wilks MF, Buckley NA. Prospects for treatment
of paraquat-induced lung fibrosis with immunosuppressive
Chen YP. Effect of early intensive immunosuppressive therapy
on moderate to severe paraquat poisoning [早期强化免疫抑制
drugs and the need for better prediction of outcome: a
治疗对中重度百草枯中毒的疗效 [Chinese]]. Chinese Journal of
systematic review. QJM: An International Journal of Medicine
2003;96:809-24.
Primary Medicine and Pharmacology 2014;21:3305-6.
Egger 1997
Ghorbani 2015 {published data only}
Egger M, Davey Smith G, Schneider M, Minder C. Bias in
Ghorbani A, Masoumi K, Forouzan A, Rahmani AH, Rahim F,
meta-analysis detected by a simple, graphical test. BMJ
Taeybi B, et al. Effect of pulse therapy with glucocorticoids and
1997;315(7109):629-34. [DOI: 10.1136/bmj.315.7109.629]
cyclophosphamide in patients with paraquat poisoning. Hong
Kong Journal of Emergency Medicine 2015;22(4):235-40. Fukuda 1985
Perriens 1992 {published data only} Fukuda Y, Ferrans VJ, Schoenberger CI, Rennard SI, Crystal RG.
Patterns of pulmonary structural remodeling after experimental
Perriens JH, Benimadho S, Kiauw IL, Wisse J, Chee H. High-dose
paraquat toxicity. The morphogenesis of intraalveolar fibrosis.
cyclophosphamide and dexamethasone in paraquat poisoning:
American Journal of Medicine 1985;118(3):452-75.
a prospective study. Human and Experimental Toxicology
1992;11(2):129-34. Gao 2020
Tsai 2009 {published data only} Gao L, Yuan H, Xu E, Junting L. Toxicology of paraquat
and pharmacology of the protective effect of 5-hydroxy-1-
Tsai J, Lee R, Wang C, Fang T, Hsu B. A clinical study of
methylhydantoin on lung injury caused by paraquat based on
prognosis and glucocorticoid pulse treatment in patients
metabolomics. Nature Scientific Reports 2020;10:1790.
with acute paraquat intoxication. Tzu Chi Medical Journal
2009;21(2):156-60. Gawarammana 2011
Gawarammana I, Buckley NA. Medical management of
paraquat ingestion. British Journal of Clinical Pharmacology
2011;72(5):1365-2125.

Informed decisions.

Ginsberg 2005 Lee 1984

Ginsberg G, Hattis D, Russ A, Sonawane B. Pharmacokinetic Lee SS, Kawanami O, Aihara K, Chung CH, Hirohata Y.
and pharmacodynamic factors that can affect sensitivity to Paraquat intoxication on rat lung. Korean Journal of Pathology
neurotoxic sequelae in elderly individuals. Environmental 1984;18(4):333.
Health Perspectives 2005;113(9):1243-9.
Lefebvre 2011
GRADEpro GDT 2020 [Computer program] Lefebvre C, Manheimer E, Glanville J. Chapter 6: Searching for
McMaster University GRADEpro Guideline Development Tool. studies. In: Higgins JPT, Green S, editor(s). Cochrane Handbook
Evidence Prime Inc, Version Available from gradepro.org. for Systematic Reviews of Interventions Version 5.1.0 (updated
McMaster University, 2020. March 2011). The Cochrane Collaboration, 2011. Available from
training.cochrane.org/handbook/archive/v5.1/.
Gunnell 2003
Gunnell D, Eddleston M. Suicide by intentional ingestion of Mew 2017
pesticides: a continuing tragedy in developing countries. Mew EJ, Padmanathan P, Konradsen F, Eddleston M, Chang SS,
International Journal of Epidemiology 2003;32(6):902-9. [DOI: Phillips MR, et al. The global burden of fatal self-poisoning
10.1093/ije/dyg307] with pesticides 2006-15: systematic review. Journal of Affective
Disorders 2017;219:93–104.
Hart 1984
Hart TB, Nevitt A, Whitehead A. A new statistical approach to Review Manager 2020 [Computer program]
the prognostic significance of plasma concentrations. Lancet Nordic Cochrane Centre, The Cochrane Collaboration Review
1984;324(8413):1222-3. Manager (RevMan). Version 5.4. Copenhagen: Nordic Cochrane
Centre, The Cochrane Collaboration, 2020.
Higgins 2011
Higgins JT, Green S, editor(s). Cochrane Handbook for Sawada 1988
Systematic Reviews of Interventions Version 5.1.0 (updated Sawada Y, Yamamoto I, Hirokane T, Nagai Y, Satoh Y,
March 2011). The Cochrane Collaboration, 2011. Available from Ueyama M. Severity index of paraquat poisoning. Lancet
training.cochrane.org/handbook/archive/v5.1/. 1988;331(8598):1333.
Higgins 2021 Smith 1975

Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Smith P, Heath D. The pathology of the lung in paraquat
Welch VA, editor(s). Cochrane Handbook for Systematic Reviews poisoning. Journal of Clinical Pathology 1975;9:81-93.
of Interventions Version 6.2 (updated February 2021). Cochrane,
2021. Available from training.cochrane.org/handbook. Tomlin 1994
Tomlin C. The Pesticide Manual. 10th edition. Farnham (UK):
Kamaruzaman 2019 British Crop Protection Council and the Royal Society of
Kamaruzaman NA, Leong Y, Jaafar MH, Khan HR, Rani NA, Chemistry, 1994.
Razali MF, et al. Epidemiology and risk factors of pesticide
poisoning in Malaysia: a retrospective analysis by the Vale 1987
National Poison Centre (NPC) from 2006 to 2015. BMJ Open Vale JA, Meredith TJ, Buckley BM. Paraquat poisoning: clinical
2020;10:e036048. features and immediate general management. Human
Toxicology 1987;6(1):41-7.
Karunarathne 2020
Karunarathne A, Gunnell D, Konradsen F, Eddleston M. How Wang 2019
many premature deaths from pesticide suicide have occurred Wang N, Jiang Q, Han L, Zhang H, Zhu B, Liu X. Epidemiological
since the agricultural Green Revolution? Clinical Toxicology characteristics of pesticide poisoning in Jiangsu Province,
2020;58(4):227-32. [DOI: 10.1080/15563650.2019.1662433] China, from 2007 to 2016. Nature Scientific Reports 2019;9:8604.
[DOI: 10.1038/s41598-019-44986-7]
Knipe 2014
Knipe DW, Metcalfe C, Fernando R, Pearson M, Konradsen F, WHO IPCS 2009
Eddleston M, et al. Suicide in Sri Lanka 1975–2012: age, period World Health Organization - International Programme for
and cohort analysis of police and hospital data. BMC Public Chemical Safety. Poison information monograph 399. Paraquat.
Health 2014;14:839. www.inchem.org/documents/pims/chemical/pim399.htm
(accessed 28 June 2021).
Lavergne 2018
Lavergne V, Gosselin S, Ghannoum M, Hoegberg LC, Roberts DM. Winsett 2004
Extracorporeal blood purification for treating acute paraquat Winsett RP, Arheart K, Stratta RJ, Alloway R, Wicks MN,
poisoning. Cochrane Database of Systematic Reviews 2018, Issue Gaber AO, et al. Evaluation of an immunosuppressant side
9. Art. No: CD013112. [DOI: 10.1002/14651858.CD013112] effect instrument. Progress in Transplantation 2004;14(3):210-6.

Informed decisions.

Xu 2019 Available at cccrg.cochrane.org/author-resources (accessed

Xu YG, Lu YQ. Systematic review and meta-analysis of the prior to 22 June 2021).
efficacy and safety of immunosuppressive pulse therapy in the

treatment of paraquat poisoning. Journal of Zhejiang University-
Science B 2019;20(7):588-97. [DOI: 10.1631/jzus.B1800640] References to other published versions of this review
Li 2014
Yan 2016
Li LR, Sydenham E, Chaudhary B, Beecher D, You C.
Yan R, Hill S. How to GRADE the quality of the evidence. Glucocorticoid with cyclophosphamide for paraquat-
Cochrane Consumers and Communication Group. Version 3.0. induced lung fibrosis. Cochrane Database of Systematic
Reviews 2014, Issue 8. Art. No: CD008084. [DOI:
10.1002/14651858.CD008084.pub4]

CHARACTERISTICS OF STUDIES
Characteristics of included studies [ordered by study ID]

Afzali 2008
Study characteristics
Methods Study design: quasi-randomised controlled trial (alternate allocation)
Duration of study: 3 months
Number of centres: 1
Location: Iran
Study setting: recruitment in ED of Sina Hospital, Hamadan
Date of study: September 2003 to October 2005
Participants Study inclusion criteria: no formal selection criteria were given; investigators apparently intended
to limit recruitment to patients with moderate to severe (but not fulminant) poisoning by the follow-
ing means (emphasis added): "Sodium dithionite reaction test was done on the urine samples of all
patients as soon as possible. ... Navy blue (NB) or dark blue (DB) colors usually indicate significant PQ
poisoning. The studied patients were divided into three groups on the basis of clinical manifesta-
tions and test results: 1) Fulminant poisoning: The urine test was NB color. All of these patients died
during the first few days (three to four days) after poisoning because of multiorgan involvement, such
as acute tubular necrosis, myocarditis, hepatic necrosis, and pulmonary bleeding. 2) Moderate to se-
vere poisoning: The urine test color was NB or DB and the clinical manifestations were oropharyngeal
burns, pharyngeal pseudomembranes, vomiting, severe diarrhoea, and acute renal and hepatic failure
3) Mild poisoning: The urine test was colourless or light blue and the clinical manifestations includ-
ed transient diarrhoea, vomiting, and buccal hyperemia, which mostly resolved without further seque-
lae. Of the 45 patients assessed in this study, 15 patients with mild and 10 patients with fulminant poi-
soning were excluded. So, 20 patients remained. Of them 11 patients received 'conventional treatmen-
t' (group 1) and nine patients received 'conventional' treatment plus 'new treatment' (group 2) ran-
domly" (Afzali 2008, p 388).
Study exclusion criteria: no formal criteria given bar those implied above
Number of participants: 45 screened; 15 excluded due to mild and 10 due to fulminant status. 20 were
randomised (9 to intervention group, 11 to conventional treatment).
Age in years, mean (∓): intervention: 27 years (∓ 10); control: 25 years (∓ 10)
Gender, n (%): intervention: female 1 (11.1%), male 8 (88.9%); control: female 3 (27.3%), male 8
(72.7%)
Severity (urine colour): intervention group: 3 NB, 6 DB; control group: 4 NB, 7 DB

Informed decisions.

Afzali 2008 (Continued)
"In our study, all patients used PQ in order to commit suicide and there were no cases of homicide or
accidental use" (Afzali 2008, p 390)
Interventions Intervention: "Conventional treatment included fixation of a nasogastric tube, gastric lavage with
normal saline, charcoal-sorbitol gavage [sic] every two to four hours for three days, forced alkalin-
ized diuresis in the first day of admission to the hospital, and haemodialysis of four hours duration
for both groups. In addition, 15 mg/kg of CP [cyclophosphamide] in dextrose saline (200 mL) was in-
fused in two hours for two days in [the intervention group] ... MP (methylprednisolone), one gram in
200 mL dextrose saline was also infused for four hours and was repeated for three consecutive days ...
as well. Meanwhile, 15 mg/kg of MESNA was prescribed (for four days) in order to avoid the side effects
of CP" (Afzali 2008, p 388).
Control: "Conventional treatment" (as above) alone
Outcomes Primary outcome: not stated. Would appear to be mortality in-hospital, prior to discharge (referred to
in document as "during the admission")
Secondary outcomes: none specifically stated. Investigators reported that creatinine > 1.4 mg/dL, ala-
nine aminotransferase (ALT) and aspartate aminotransferase (AST) > 70 IU/L, and PaO2 < 70 mmHg (in
the room air) were considered asacute renal failure (ARF), hepatitis, and hypoxia, respectively (em-
phasis added).
Information was provided concerning daily in-hospital assessments: "laboratory tests included liver
function tests (LFT), complete blood count (CBC), arterial blood gas (ABG), blood urea nitrogen (BUN),
and creatinine"; chest radiography was repeated every 2 days. Finally, autopsies were conducted for
"all expired patients".
Data on adverse effects do not appear to have been rigorously sought, but investigators mention that
there were "no severe complications ... in the new treatment group" (Afzali 2008, p 390).
Notes Funding: none reported.
Registration: none reported (though it is noted that "this survey [sic] was approved by the Ethics Com-
mittee of Hamadan University of Medical Sciences") (Afzali 2008, p 388).
Interests: none declared.
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence genera- Unclear risk "11 patients received 'conventional treatment' ... and nine patients received
tion (selection bias) 'conventional' treatment plus 'new treatment' ... randomly" (Afzali 2008, p
388)
We contacted the first author of the study, who informed us that alternate allo-
cation was used (Afzali 2010 - personal correspondence between lead investi-
gator and previous review author Emma Sydenham. Correspondence details
are no longer available). Whilst we initially considered this domain as at high
risk of bias, given the long period of the study (25 months, in which only 45 pa-
tients were screened), we reconsidered that the danger of manipulation of the
allocation of successive participants was not serious.
Allocation concealment High risk Concealment of allocation is not mentioned in the paper, and is not possible
(selection bias) with use of alternate allocation.
Blinding (performance Low risk Blinding was not reported, and according to the author was not done (Afzali
bias and detection bias) 2010 - personal correspondence - see above).
All outcomes
Given that the main outcome of interest (mortality) is objective, we did not
consider it likely that the lack of blinding introduced bias.

Informed decisions.

Afzali 2008 (Continued)
Incomplete outcome data Low risk The main outcome was death at discharge from hospital, which appears to be
(attrition bias) reported in full.
All outcomes
Selective reporting (re- Unclear risk The study reported the main outcome for this review (mortality) in full; howev-
porting bias) er, in the absence of a trial protocol or evidence of prospective registration, we
are obliged to report the risk of bias as unclear.
Other bias High risk Investigators acknowledged the "limitation" that they "did not check plasma
PQ" in participants on entry to the study (or at any other point).

Gawarammana 2017
Methods Study design: randomised controlled trial
Duration of study: 3 months
Location: Sri Lanka
Study setting: district hospitals
Date of study: March 2007 to November 2010
Participants Study inclusion criteria: individuals aged 14 years and over with positive urine dithionite test who pre-
sented within 48 h of paraquat ingestion and received no other specific treatments such as haemodial-
ysis or haemoperfusion
Study exclusion criteria: people "with a systolic blood pressure < 70mmHg that did not respond to 1 L
of intravenous fluid or a Glasgow Coma score less than 8/15 were excluded" (Gawarammana 2017, p 2)
Number of participants: 604 screened; 305 excluded due to negative urine test, refusing consent,
death before randomisation, GCS < 8, pregnant, aged under 14; late admission. 299 participants were
randomised (147 to intervention group, 152 to placebo). 143 received intervention as allocated; 147 re-
ceived placebo as allocated.
Age in years, median (IQR): intervention: 27 (21 to 38); control: 27 (21 to 36)
Gender, n (%): intervention: female 47 (32%), male 100 (68%); control: female 39 (26%), male 113
(74%)
Median (IQR) and range of SIPP: intervention group (data for 125): 18.4 (2.7 to 56.2); control group
(data for 127): 13.1 (3.1 to 58.4). Investigators reported that the SIPP had a small but significant differ-
ence at baseline.
Interventions Intervention: on day 0 and 1 (i.e. shortly following randomisation and 24 h later):
1. MESNA (400 mg) (Baxter Oncology GmbH, Halle, Germany) or normal saline, given intravenously (IV)
2. Cyclophosphamide 15 mg/kg (CYPHOS 200 mg, Get Well Pharmaceutical, Haryana, India) (based on
actual or estimated weight*)
3. Methylprednisolone 1 g given over 1 h by infusion (Medixon 500 mg, PT Phapros, Semerang for PT Dexa
Medica, Titan Center, Tangerang, Indonesia)
4. Second dose MESNA given IV 4 h after the start of CP infusion
5. Third dose MESNA given IV 8 h after the start of CP infusion
On day 2: 1 g methylprednisolone given over 1 hour by infusion

Informed decisions.

Gawarammana 2017 (Continued)
From day 3 onwards, participants were given IV or oral dexamethasone 8 mg (State Pharmaceutical
Manufacturing Corporation, Colombo, Sri Lanka) 3 times daily for 2 weeks or placebo.
Comparison: normal saline (State Pharmaceutical Manufacturing Corporation, Colombo, Sri Lanka)
was used as placebo for all IV drugs.
* "The protocol suggests a possibility of using major rounding of doses (to 750 mg or 1000 mg) if an ac-
tual weight could not be measured but this might have meant doses of up to 20 mg/kg for some indi-
viduals. This was not done and 15 mg/kg based on estimated weight was used." (Supplemental Box 1)
Outcomes Primary outcome: mortality in-hospital
Secondary outcomes: mortality at 3 months; lung function (formal lung functions and a high resolu-
tion CT scan (HRCT) at long-term follow-up); however, this was not possible for "a meaningful number
of patients" (Gawarammana 2017, p 634)
Data on adverse effects, specifically haematuria and bladder pain, were sought in hospital.
Notes Funding: Syngenta Crop Protection AG; Wellcome Trust/National Health and Medical Research Council
International Collaborative Research Grant (ICRG 071669); Australian Leadership Award (ALA00379)
Registration: ISRCTN85372848
Interests: "MW [10th author] was an employee of Syngenta at the time of trial inception and until 2009.
AHD [12th author] received funding to attend DMEC meeting from Syngenta. ME [11th author] have [sic]
received travel expenses from Syngenta to attend meetings of a scientific advisory group in relation to
studies of new paraquat formulations. No other conflicts declared" (Gawarammana 2017, p 639)
Risk of bias
Random sequence genera- Low risk Randomisation was done using purpose-designed computer software.
tion (selection bias)
Allocation concealment Low risk "The random sequence and allocation were concealed prior to randomiza-
(selection bias) tion ... The allocation sequences were generated and encrypted independently
by an IT consultant who had no role in patient recruitment, treatment and as-
sessment." (Gawarammana 2017, p 634)
Blinding (performance Low risk "The allocation was only known to the pharmacists who had no other role in
bias and detection bias) patient management and data collection." (Gawarammana 2017, p 634)
All outcomes
Incomplete outcome data Low risk "We were able to follow up all but six patients post discharge." 4 losses to fol-
(attrition bias) low-up were in the intervention group and 2 in the control group.
All outcomes
Selective reporting (re- Low risk Study was prospectively registered (ISRCTN85372848). Ethics were granted
porting bias) a year before recruitment began. Methods and results seem to be in concert
with planned analyses; reasons for lack of presentation of planned lung func-
tion results are reported, as are grounds for early stopping.
Other bias Low risk This RCT was well-conducted; however, it was stopped early "after consulta-
tion with the data monitoring and ethics committee due to a collapse in re-
cruitment". The collapse followed regulatory decisions which saw the phas-
ing out of paraquat use "from Sri Lankan agricultural practice through 2008
and 2009"; the herbicide was "finally banned in August 2010" (Gawarammana
2017, p 635).

Informed decisions.

Gawarammana 2017 (Continued)
Gawarammana 2017 reported a small difference in the median SIPP score of
participants allocated to intervention (median 18.4) compared to control (me-
dian 13.1). As the method of randomisation was unlikely to have been compro-
mised (it was done using purpose-designed software), this difference is likely
to have been caused by chance alone, therefore we do not consider this to be a
source of bias.

Lin 1999
Duration of study: hospital admission to discharge (maximum length not stated)
Location: Taipei, Taiwan, Republic of China
Study setting: recruitment in ED; treatment in specialised poison centre
Date of study: January 1999 to December 2003
Participants Study inclusion/exclusion criteria: people who had ingested paraquat within the previous 24 h, and
had a urine sample that resulted in a navy blue or dark blue reaction to a sodium dithionite test, were
screened.
Number of participants: 142 screened. Investigators reported that they excluded 71 people with "ful-
minant" cases and 21 with "mild" cases, and "recruited" only 50 participants; however, they reported
data for the "fulminant" and the "recruited" cases and excluded only "mild" cases, for which data were
not presented.
N at baseline:
"Fulminant": intervention group (n = 34); control group (n = 37)
"Recruited": intervention group (n = 22); control group (n = 28)
Age in years:
"Fulminant" group (n = 71): intervention group: mean 35.7 years (SD 13.6); control group: mean 38.2
years (SD 16.3)
"Recruited" group (n = 50): intervention group: mean 28.2 years (SD 11.1); control group: mean 26.8
years (SD 10.1)
Gender, n:
"Fulminant" group' (n = 71): intervention group: 14 F, 20 M; control group: 10 F, 27 M
"Recruited" group (n = 50): intervention group: 9 F, 13 M; control group: 10 F, 18 M
Severity (urine colour):
"Fulminant" group (n = 71): intervention group: 2 NB, 35 DB; control group: 2 NB, 32 DB
"Recruited" group (n = 50): intervention group: 8 NB, 14 DB; control group: 9 NB, 19 DB
Interventions All participants received: "... active charcoal added in magnesium citrate ... given through a nasogas-
tric tube after gastric lavage with normal saline. All patients received two courses of 8-h active charcoal
haemoperfusion therapy in the emergency room (ER), and dexamethasone 10 mg intravenous injection
every 8 h was given for 14 d after admission" (Lin 1999, p 357)
Informed decisions.

Lin 1999 (Continued)
Intervention: the intervention group also received: "pulse therapy after haemoperfusion at ER. Pulse
therapy included 15 g/kg of CP in 5% glucose saline 200 ml and 1 g MP in the other 200 ml 5% glucose
saline intravenously infused for 2 h/d. CP was infused for 2 d and MP for 3 d" (Lin 1999, p 357).
Control: no pulse therapy
Outcomes • Mortality (at hospital discharge)

• Leukopenia (in hospital, as a major complication of treatment)
Notes Funding: this trial was "supported by the National Science Council Foundation, Taiwan, ROC
(NSC-88-2314-B-182A-049)" (Lin 1999, p 358).
Registration: not identified
Interests: no interests declared
Risk of bias
Random sequence genera- Low risk "... according to random digit methods" (Lin 1999, p 357)
tion (selection bias)
Allocation concealment Unclear risk There was no mention of allocation concealment.

(selection bias)
Blinding (performance Low risk Since no placebo was used, blinding of the treating physician and participants
bias and detection bias) was not possible. However: "At the end of this study, to avoid bias, the da-
All outcomes ta were collected and analyzed by other doctors who were not aware of the
study" (Lin 1999, p 357). Given that the main outcome of interest (mortality)
is objective, we did not consider it likely that the lack of blinding introduced
bias.
Incomplete outcome data Low risk The main outcome was death at final follow-up, which was reported in full.
(attrition bias)
All outcomes
Selective reporting (re- Unclear risk In the absence of a trial protocol or evidence of prospective registration, we
porting bias) are obliged to report the risk of bias as unclear.
Furthermore, 121 participants were randomised into the study regardless of

the severity of poisoning, and information on the primary outcome (mortali-
ty) was presented in full. However, the trialists made what appears to be a post
hoc decision to exclude very severely poisoned patients ("fulminant"; n = 71)
from the final analysis. We have chosen to report mortality data both in aggre-
gate and separately in order to consider the impact of all data provided and to
minimise the potential for selective reporting.
Other bias High risk Methods of diagnosing poisoning severity did not include plasma assess-
ment (likely leading to the post protocol decision to exclude participants later
known to have been fulminant on admission).

Lin 2006

Informed decisions.

Duration of study: 42 days
Location: Taipei, Taiwan, Republic of China
Study setting: recruitment in ED; treatment in specialised poison centre
Date of study: January 1992 to December 1997
Participants Study inclusion criteria: people who:
• arrived at the emergency department within 24 hours of ingesting paraquat;

• were age 15 years or older;
• had paraquat plasma levels between the predictive mortality of > 50% and ≤ 90% according to the
Hart 1984 formula; and
• had urine sodium dithionite tests showing the colour dark blue or navy blue.
Study exclusion criteria: "exposure to paraquat via skin or intravascular [sic] injection; absence of
paraquat levels in biological fluids; or had ingested paraquat due to major systemic diseases including
cancer and heart, lung, renal, and liver diseases; or did not give informed consent" (Lin 2006, p 369)
Number of participants: 96 screened. Investigators excluded 13 for arriving > 24 h after ingestion; 6 for
exposure to paraquat other than that stated in the inclusion criteria; 21 for excessive predicted mortali-
ty (> 90%); 18 for insufficient predicted mortality (< 50%); 15 for having no paraquat detectable in urine.
23 participants remained, of whom 16 were assigned to the intervention and 7 to control.
Age in years: intervention group: mean 33.6 years (SD 14.3); control group: mean 37.0 years (SD 13.8)
Gender, n: intervention group: 5 F, 11 M; control group: 2 F, 5 M
Severity (urine colour): intervention group: 5 NB, 11 DB; control group: 1 NB, 6 DB
Mean SIPP: intervention group: 14.6 (SD 9.1); control group: 15.5 (SD 7.0)
Interventions All participants received: "To prevent absorption of paraquat from the gastrointestinal tract, acti-
vated charcoal 1 g/kg added to 250 mL of magnesium citrate was given through a nasogastric tube af-
ter gastric lavage with normal saline. In addition, all patients received two doses of 8-hr active char-
coal-containing haemoperfusion therapy in the emergency room. After hemoperfusion therapy, the
control group received dexamethasone 5 mg in an intravenous injection every 6 hrs until their arterial
blood gas showed PaO2 11.5 kPa (80mm Hg) or they died" (p 369)
Intervention: "... pulse therapy with 15 mg/kg cyclophosphamide in 5% glucose saline 200 mL and 1
g of methylprednisolone in the other 200 mL of 5% glucose saline intravenously infused for 2 hrs per
day. Cyclophosphamide was infused for 2 days and methylprednisolone for 3 days simultaneously. Pre-
ceding dexamethasone, a 5-mg intravenous injection every 6 hrs was given until the arterial blood gas
showed PaO2 11.5 kPa (80 mm Hg). Repeated pulse therapy with 1 g of methylprednisolone in the oth-
er 200 mL of 5% glucose saline intravenously infused for 2 hrs per day for 3 days was given again if PaO2
was 8.64 kPa (60 mm Hg). In addition, 15mg/kg/day cyclophosphamide was infused for 1 day again if
patients’ white cell counts were 3000/m3 and the duration was 2 wks after initial cyclophosphamide
pulse therapy to avoid a severe leukopenia episode" (Lin 2006, p 369)
Control: no pulse therapy, but as in intervention group: "... dexamethasone 5 mg in an intravenous in-
jection every 6 hrs until their arterial blood gas showed PaO2 11.5 kPa (80mm Hg) or they died" (Lin
2006, p 369)
Outcomes • Mortality (at minimum of 6 weeks)

• Leukopenia (in hospital)
Notes Funding: not specified
Registration: not identified

Informed decisions.

Interests: "None of the authors have any financial interests to disclose" (Lin 2006, p 368)
Risk of bias
Random sequence genera- Low risk "All study patients were randomly allocated to control and study groups in the
tion (selection bias) proportion of 1:2 by means of a sequence of labelled cards contained in sealed
numbered envelopes that were prepared by a statistical adviser." (Lin 2006, p
369)
Allocation concealment Low risk "... [the envelope was] opened by the researcher in the presence of pa-
(selection bias) tients." (Lin 2006, p 369)
Blinding (performance Low risk "Neither stratification nor blinding was made [sic] in this study." (Lin 2006, p
bias and detection bias) 369) "The data were collected and analyzed by other doctors not familiar with
All outcomes the study." (Lin 2006, p 369)
Given that the main outcome of interest (mortality) is objective, we did not
consider it likely that the lack of blinding introduced bias.
Incomplete outcome data Low risk The main outcome was death at final follow-up, which was reported in full.
(attrition bias)
All outcomes
Selective reporting (re- Unclear risk The study reported the main outcome (death) at final follow-up in full; howev-
porting bias) er, in the absence of a trial protocol or evidence of prospective registration, we
are obliged to report the risk of bias as unclear.
Other bias Low risk We did not identify any other areas of bias.
Abbreviations
CP = cyclophosphamide; CT = computed tomography; DB = dark blue; DMEC = Data Monitoring and Ethics Committee; ED = Emergency
Department; F = female; GCS = Glasgow Coma Score; HRCT = high resolution CT scan; IQR = interquartile range; M = male; MESNA
= 2-mercaptoethane sulfonate sodium (Na); MP = methylprednisolone; NB = navy blue; PaO2 = partial pressure of oxygen; PQ = paraquat;
SD = standard deviation; SIPP = Severity Index of Paraquat Poisoning

Characteristics of excluded studies [ordered by study ID]

Study Reason for exclusion
Chen 2014 This study was characterised as an RCT within Gao 2020; however, the paper is very brief, has 1 au-
thor, no detail on methods (including randomisation), and lacked prospective registration.
Ghorbani 2015 This RCT lacked evidence of prospective registration, which is required for the inclusion of studies
conducted after the year of 2010, according to Cochrane Injuries Group policy (2015).
Perriens 1992 This comparative study used an historic control group. "Patients admitted before October 10, 1986
received the standard treatment only, because IV cyclophosphamide and IV dexamethasone were
not available in Suriname until that time. Patients presenting after October 10, 1986 received high-
dose cyclophosphamide and dexamethasone treatment, in addition to standard treatment" (Per-
riens 1992, p 130)
Tsai 2009 This study focused on methylprednisolone (an eligible glucocorticoid) only, without cyclophos-
phamide; moreover, there was no mention of randomisation.

Informed decisions.

IV = intravenous; RCT = randomised controlled trial

DATA AND ANALYSES

Comparison 1. All-cause mortality
Outcome or subgroup title No. of studies No. of partici- Statistical method Effect size
pants
1.1 All-cause mortality 4 Risk Ratio (M-H, Fixed, Subtotals only

95% CI)
1.1.1 Mortality at hospital discharge 4 392 Risk Ratio (M-H, Fixed, 0.73 [0.61, 0.88]
95% CI)
1.1.2 Mortality at 3 months 1 293 Risk Ratio (M-H, Fixed, 0.98 [0.85, 1.13]
95% CI)
1.2 Sensitivity analysis: all-cause mortali- 4 Risk Ratio (M-H, Fixed, Subtotals only
ty re-including fulminant participants 95% CI)
95% CI)
1.3 Sensitivity analysis: all-cause mor- 2 Risk Ratio (M-H, Fixed, Subtotals only
tality: excluding studies without plasma 95% CI)
paraquat assessment at baseline
95% CI)

Analysis 1.1. Comparison 1: All-cause mortality, Outcome 1: All-cause mortality


Afzali 2008 3 9 9 11 6.6% 0.41 [0.16 , 1.07]
Gawarammana 2017 78 147 94 152 75.2% 0.86 [0.70 , 1.04]
Lin 1999 4 22 16 28 11.5% 0.32 [0.12 , 0.82]
Lin 2006 5 16 6 7 6.8% 0.36 [0.17 , 0.80]
Subtotal (95% CI) 194 198 100.0% 0.73 [0.61 , 0.88]
1.1.2 Mortality at 3 months

Gawarammana 2017 101 143 108 150 100.0% 0.98 [0.85 , 1.13]
Subtotal (95% CI) 143 150 100.0% 0.98 [0.85 , 1.13]
Heterogeneity: Not applicable
Test for subgroup differences: Chi² = 5.88, df = 1 (P = 0.02), I² = 83.0% 0.2 0.5 1 2 5

Informed decisions.


Analysis 1.2. Comparison 1: All-cause mortality, Outcome 2: Sensitivity
analysis: all-cause mortality re-including fulminant participants


Afzali 2008 3 9 9 11 5.1% 0.41 [0.16 , 1.07]
Gawarammana 2017 78 147 94 152 58.5% 0.86 [0.70 , 1.04]
Lin 1999 38 56 53 65 31.1% 0.83 [0.67 , 1.03]
Lin 2006 5 16 6 7 5.3% 0.36 [0.17 , 0.80]
Subtotal (95% CI) 228 235 100.0% 0.80 [0.69 , 0.93]


Analysis 1.3. Comparison 1: All-cause mortality, Outcome 3: Sensitivity analysis: all-
cause mortality: excluding studies without plasma paraquat assessment at baseline


Gawarammana 2017 78 147 94 152 91.7% 0.86 [0.70 , 1.04]
Lin 2006 5 16 6 7 8.3% 0.36 [0.17 , 0.80]
Subtotal (95% CI) 163 159 100.0% 0.82 [0.68 , 0.99]


APPENDICES
Appendix 1. Search strategies for update searches run in 2017 and 2020
Ovid MEDLINE(R) and Epub Ahead of Print, In-Process & Other Non-Indexed Citations, Daily and Versions(R) 1946 to September 21,
2020
1 exp Herbicides/
2 exp Paraquat/
3 (Paraquat or (methyl adj3 viologen) or Dimethyl* or gramoxone or grammoxone or paragreen or Herbicide* or Pyridinium compound*
or pathclear or weedol).mp.
4 1 or 2 or 3
5 exp Glucocorticoids/
6 glucocorticoid*.ab,ti.
7 exp Cyclophosphamide/
8 (cyclophosphamid* or carloxan or clafen or cycloblastin or cycloblastine or cyclofos amide or cyclofosfamid or cyclofosfamide or
cyclophosphan*or cycloxan or cyphos or cytophosphan* or cytoxan or endocyclo phosphate or endoxan* or enduxan or genoxalor mitoxan
or neosan or neosar or noristan or nsc 26271 or nsc 2671 or b-518 or procytox* or semdoxan or sendoxan).ab,ti.
9 5 or 6 or 7 or 8
10 4 and 9
11 randomi?ed.ab,ti.
12 randomized controlled trial.pt.
Informed decisions.

13 controlled clinical trial.pt.

14 placebo.ab.
15 clinical trials as topic.sh.
16 randomly.ab.
17 trial.ti.
18 Comparative Study/
19 11 or 12 or 13 or 14 or 15 or 16 or 17 or 18
20 (animals not (humans and animals)).sh.
21 19 not 20
22 10 and 21
EMBASE 1980 to 2020 Week 38

1 exp Herbicide/
2 exp Paraquat/
3 exp pyridinium derivative/
4 (Paraquat or (methyl adj3 viologen) or Dimethyl* or gramoxone or grammoxone or paragreen or Herbicide* or Pyridinium compound*
5 1 or 2 or 3 or 4
6 exp Glucocorticoids/
7 glucocorticoid*.ab,ti.
8 exp Cyclophosphamide/
9 exp cyclophosphamide derivative/
10 (cyclophosphamid* or carloxan or clafen or cycloblastin or cycloblastine or cyclofos amide or cyclofosfamid or cyclofosfamide or
11 6 or 7 or 8 or 9 or 10
12 5 and 11
13 exp Randomized Controlled Trial/
14 exp controlled clinical trial/
15 exp controlled study/
16 comparative study/
17 randomi?ed.ab,ti.
18 placebo.ab.
19 *Clinical Trial/
20 exp major clinical study/
21 randomly.ab.
22 (trial or study).ti.
23 13 or 14 or 15 or 17 or 18 or 19 or 20 or 21 or 22
24 exp animal/ not (exp human/ and exp animal/)
25 23 not 24
26 5 and 11 and 25
27 limit 26 to embase
28 limit 27 to yr="2017 - 2021"
WEB OF SCIENCE. SCI-EXPANDED, SSCI, A&HCI, CPCI-S, CPCI-SSH, ESCI (2017 to 22 September 2020)SCI-EXPANDED, SSCI, A&HCI,
CPCI-S, CPCI-SSH, ESCI (all years to 21 Sept 2020)
# 3#2 AND #1 Indexes=SCI-EXPANDED, CPCI-S Timespan=2017-2020
# 2TS=(cyclophosphamid* or carloxan or clafen or cycloblastin or cycloblastine or cyclofos amide or cyclofosfamid or cyclofosfamide or
or neosan or neosar or noristan or nsc 26271 or nsc 2671 or b-518 or procytox* or semdoxan or sendoxan) Indexes=SCI-EXPANDED, CPCI-
S Timespan=2017-2020
# 1TS=(Paraquat or Dimethyl* or gramoxone or grammoxone or paragreen or Herbicide* or pathclear or weedol or Pyridinium compound*)
or TS=(methyl same viologen) Indexes=SCI-EXPANDED, CPCI-S Timespan=2017-2020
CENTRAL (searched 21 September 2020)

#1 MeSH descriptor: [Herbicides] explode all trees
#2 MeSH descriptor: [Paraquat] explode all trees
#3 (Paraquat or (methyl near/3 viologen) or Dimethyl* or gramoxone or grammoxone or paragreen or Herbicide* or Pyridinium compound*
or pathclear or weedol):ti,ab,kw (Word variations have been searched)
#4 #1 or #2 or #3
#5 MeSH descriptor: [Glucocorticoids] explode all trees

Informed decisions.

#6 (cyclophosphamid* or carloxan or clafen or cycloblastin or cycloblastine or cyclofos amide or cyclofosfamid or cyclofosfamide or

or neosan or neosar or noristan or nsc 26271 or nsc 2671 or b-518 or procytox* or semdoxan or sendoxan):ti,ab,kw (Word variations have
been searched)
#7 MeSH descriptor: [Cyclophosphamide] explode all trees
#8 #5 or #6 or #7
#9 #4 and #8
Chinese databases were searched using the general strategy: Paraquat AND cyclophosphamide AND lung in abstracts， with synonym
expansion, and publication time from 2014 to 2020.
CNKI: 检索范围： (摘要%百草枯) AND (摘要%环磷酰胺) AND (摘要%肺) 资源范围: 总库；同义词扩展；时间范围：发表时

间:2014-04-01到2020-11-12；
WAN FANG DATA: 更新时间: 不限万⽅：检索表达式：百草枯 * 环磷酰胺 * 肺 * Date:2014-2020
VIP: ⽂摘=百草枯AND⽂摘=环磷酰胺AND⽂摘=肺 AND 年份：2014-2020
Translation:
China National Knowledge Infrastructure (CNKI): (Abstract % paraquat ) AND (Abstract% cyclophosphamide ) AND (Abstract % lung). All
databases; synonym expand; Time: from 2014-04-01 to 2020-11-12.
WAN FANG DATA: paraquat*cyclophosphamide*lung*Date: 2014-2020.
VIP: Abstract=paraquat AND Abstract=cyclophosphamide AND Abstract=lung AND year: 2014-2020.
Appendix 2. Search strategies 2014 update

For this update the search strategies were modified: the terms relating to lung were removed and the RCT filters were added. The strategies,
as they were, did not retrieve the included studies even though they were indexed in MEDLINE and /or Embase. The included studies may
have been retrieved either by screening reference lists.The added study filter is a modified version of the Ovid MEDLINE Cochrane Highly
Sensitive Search Strategy for identifying randomised trials (Lefebvre 2011); for Embase we added to the search strategy study design terms
as used by the UK Cochrane Centre (Lefebvre 2011).
Other strategies for databases in the English language were not modified.
Cochrane Injuries Group's Specialised Register
#1 ((Paraquat or (methyl and viologen) or Dimethyl* or gramoxone or grammoxone or paragreen or Herbicide* or "Pyridinium compound"
or pathclear or weedol)) AND ((cyclophosphamid* or carloxan or clafen or cycloblastin or cycloblastine or "cyclofos amide" or
cyclofosfamid or cyclofosfamide or cyclophosphan* or cycloxan or cyphos or cytophosphan* or cytoxan or "endocyclo phosphate" or
endoxan* or enduxan or "genoxalor mitoxan" or neosan or neosar or noristan or "nsc 26271" or "nsc 2671" or b-51)) [REFERENCE]
[STANDARD]
MEDLINE (OvidSP)
1. exp Herbicides/
2. exp Paraquat/
3.(Paraquat or (methyl adj3 viologen) or Dimethyl* or gramoxone or grammoxone or paragreen or Herbicide* or Pyridinium compound*
4. 1 or 2 or 3
5. exp Lung Diseases/
6. exp Pulmonary Fibrosis/
7. ((Pulmonary or lung) adj3 (fibrosis or fibroses)).ab,ti.
8. ((Alveolitis or alveolitides) adj3 fibrosing).ab,ti.
9. 5 or 6 or 7 or 8
10. exp Glucocorticoids/
11. glucocorticoid*.ab,ti.
12. exp Cyclophosphamide/
13. (cyclophosphamid* or carloxan or clafen or cycloblastin or cycloblastine or cyclofos amide or cyclofosfamid or cyclofosfamide or
14. 10 or 11 or 12 or 13
15. 4 and 9 and 14
16. (animals not (humans and animals)).sh.

Informed decisions.

17. 15 not 16
Embase (Ovid)
1. exp Herbicide/
2. exp Paraquat/
3. exp pyridinium derivative/
4. (Paraquat or (methyl adj3 viologen) or Dimethyl* or gramoxone or grammoxone or paragreen or Herbicide* or Pyridinium compound*
5. 1 or 2 or 3 or 4
6. exp Lung Disease/
7. exp lung fibrosis/
10. 6 or 7 or 8 or 9
14. exp cyclophosphamide derivative/
16. 11 or 12 or 13 or 14 or 15
17. 5 and 10 and 16
18. exp animal/ not (exp human/ and exp animal/)
19. 17 not 18
ISI Web of Science: Science Citation Index Expanded (SCI-EXPANDED) and ISI Web of Science: Conference Proceedings Citation
Index-Science (CPCI-S)
1. (Paraquat or Dimethyl* or gramoxone or grammoxone or paragreen or Herbicide* or pathclear or weedol or Pyridinium compound*) or
(methyl same viologen)
or neosan or neosar or noristan or nsc 26271 or nsc 2671 or b-518 or procytox* or semdoxan or sendoxan)
3. ((Pulmonary or lung) SAME (fibrosis or fibroses)) OR ((Alveolitis or alveolitides) SAME fibrosing) OR (lung* SAME disease*)
4. 1 and 2 and 3
Clinical trials registries

Search terms: paraquat
Study results: All studies
Chinese databases
The databases orginially searched in 2012 are now incorporated in a new Government sponsored database:
China National Knowledge Infrastructure (CNKI 数据库）
The authors also searched the following which were not included in the 2012 search:
WAN FANG DATA（万⽅数据库）(维普数据库)
The search string was limited to, Paraquat AND lung AND cyclophosphomide, due to the difficulties in using the search interfaces.
Appendix 3. Search methods for previous versions of the review

Cochrane Injuries Group's Specialised Register
1. (Paraquat or (methyl and viologen) or Dimethyl* or gramoxone or grammoxone or paragreen or Herbicide* or "Pyridinium compound"
or pathclear or weedol)
2. (cyclophosphamid* or carloxan or clafen or cycloblastin or cycloblastine or "cyclofos amide" or cyclofosfamid or cyclofosfamide or
cyclophosphan* or cycloxan or cyphos or cytophosphan* or cytoxan or "endocyclo phosphate" or endoxan* or enduxan or "genoxalor
mitoxan" or neosan or neosar or noristan or "nsc 26271" or "nsc 2671" or b-51
3. 1 and 2
CENTRAL (The Cochrane Library)

#1 MeSH descriptor Herbicides explode all trees
#2 MeSH descriptor Paraquat explode all trees
#3 (Paraquat or (methyl near3 viologen) or Dimethyl* or gramoxone or grammoxone or paragreen or Herbicide* or Pyridinium compound*
or pathclear or weedol)
Informed decisions.

#4 (#1 OR #2 OR #3)
#5 MeSH descriptor Glucocorticoids explode all trees
#6 MeSH descriptor Cyclophosphamide explode all trees
#7 (cyclophosphamid* or carloxan or clafen or cycloblastin or cycloblastine or cyclofos amide or cyclofosfamid or cyclofosfamide or
or neosan or neosar or noristan or nsc 26271 or nsc 2671 or b-518 or procytox* or semdoxan or sendoxan)
#8 (#5 OR #6 OR #7)
#9 (#4 AND #8)
Ovid MEDLINE
1. exp Herbicides/
2. exp Paraquat/
3. (Paraquat or (methyl adj3 viologen) or gramoxone or paragreen or Herbicide* or Pyridinium Compound*).mp.
4. 1 or 2 or 3
5. exp Lung Diseases/
6. exp Pulmonary Fibrosis/
9. 5 or 6 or 7 or 8
13. (Cyclophosphamide* or cytophosphan or cyclophosphane or procytox or sendoxan or b-518 or neosar or cytoxan or endoxan or
nsc-26271).ti,ab.
14. 10 or 11 or 12 or 13
15. randomi?ed.ab,ti.
16. randomized controlled trial.pt.
17. controlled clinical trial.pt.
18. placebo.ab.
19. clinical trials as topic.sh.
20. randomly.ab.
21. trial.ti.
22. 15 or 16 or 17 or 18 or 19 or 20 or 21
23. (animals not (humans and animals)).sh.
24. 22 not 23
25. 4 and 9 and 14 and 24
CBM (1978 to April 2012), CMCC (1995 to April 2012), CMAC (1994 to April 2012)
1. exp herbicide/
2. exp paraquat/
3. 1 or 2
4. exp lung disease/
5. exp pulmonary fibrosis/
6. 4 or 5
7. exp corticosteroids/
8. steroids*.ab,ti.
9. exp cyclophosphamide/
10. 7 or 8 or 9 11.(randomised).ab,ti.
12. randomized controlled study.pt.
13. clinical controlled study.pt.
14. randomly.ab.
15. trial.ti.
16. 11 or 12 or 13 or 14 or 15.
17. 3 and 6 and 10 and 16.
WHAT'S NEW


Informed decisions.

Date Event Description
29 June 2021 New citation required and conclusions Review updated with a search date of 21 September 2020 and
have changed one new study added.
12 November 2020 New search has been performed The search has been updated to 21 September 2020.
This update includes four studies. One study has been added
since the previous published version of this review (Gawaram-
mana 2017).

HISTORY
Protocol first published: Issue 4, 2009
Review first published: Issue 6, 2010

Date Event Description
12 November 2020 Amended Jane Dennis has been added as an author for this update.
12 December 2019 Amended Helen Wakeford was added as an author for this update.
12 December 2019 Amended Helen Wakeford was added as an author for this update.
20 July 2019 Amended We added the outcome 'new infections within one week after ini-
tiation of the treatment' (adverse effect of treatment).
20 July 2019 Amended Emma Sydenham is no longer an author for this update.
20 July 2019 Amended Deirdre Beecher is no longer an author for this update.
25 May 2012 New search has been performed The search has been updated to 1 February 2012. No new studies
were identified. The results and conclusions remain the same.
24 May 2012 New citation required but conclusions The search has been updated, but no new studies were identi-
have not changed fied. The results and conclusions remain the same.

CONTRIBUTIONS OF AUTHORS
For the 2021 version of this review: Iris Gordon and Jane Dennis (JD) re-ran the searches in English language databases (in 2017 and 2020
respectively). LL, Helen Wakeford (HW) and JD screened the search results. HW updated the Background section. JD extracted data and
rewrote the Characteristics of included studies and Description of studies. All authors contributed to the Results and Discussion. LL, JD,
and HW performed the GRADE assessment of the certainty of evidence for each outcome. LL searched Chinese language databases and
screened the search results. CY and BC also screened the Chinese language search results.
For the 2014 version: Deirdre Beecher (DB) developed the search strategy, searched English language databases, and screened the search
results. ES screened some of the search results. LL developed the search strategy, searched Chinese language databases, and screened
the search results. CY and BC also screened the Chinese language search results. DB, LL, and ES updated the manuscript. All authors
contributed to the updated review.
For the 2010 and 2012 versions: Luying Ryan Li (LL) and Chao You (CY) were responsible for writing the protocol. Bhuwan Chaudhary
(BC) and LL selected the trials from English language databases. LL and CY selected trials from Chinese language databases. CY offered
interpretation of the clinical features of paraquat poisoning and arbitrated on the inclusion of one trial (Afzali 2008). Emma Sydenham (ES),
LL, and BC independently assessed the risk of bias of the included trials and extracted data. LL and ES interpreted the data and wrote the
manuscript. All authors agreed on the final manuscript.

Informed decisions.

DECLARATIONS OF INTEREST
Luying Ryan Li: none known.
Bhuwan Chaudhary: none known.
Chao You: none known.
Jane Dennis: none known. JD was employed by Cochrane Injuries during her involvement in the development of the review.
Helen Wakeford: none known. HW was employed by Cochrane Injuries and the Cochrane Central Executive Team during her involvement
in the development of the review.
SOURCES OF SUPPORT
Internal sources
• No sources of support provided
External sources
• National Institute for Health Research (NIHR), UK
Cochrane Infrastructure Funding to the Cochrane Injuries Group
DIFFERENCES BETWEEN PROTOCOL AND REVIEW
We adjusted the title for the 2021 update from the original 'Glucocorticoid with cyclophosphamide for paraquat-induced lung fibrosis' to
'Glucocorticoid with cyclophosphamide for oral paraquat poisoning' to reflect the scope of the review more accurately.
In the protocol for this review, we specified that we would search the Chinese language databases Chinese Bio-Medical Literature &
Retrieval System (CBM), Chinese Medical Current Contents (CMCC), and Chinese Medical Academic Conference (CMAC). These databases
have now been included in the China National Knowledge Infrastructure database. For the 2014 and 2021 updates, we searched the
databases China National Knowledge Infrastructure (数据库), Wanfang Data (万⽅数据库), and VIP (维普数据库).
The protocol for this review (2009) contained no plans for dealing with issues concerning unit of analysis or subgroup or sensitivity analyses.
These have been added to the 2021 update to meet current Cochrane methodological standards.
For the 2021 version, we added infection within one week after initiation of the treatment as an outcome for side effects of the treatment.
We also mentioned leukopenia, as related to infection risk.
We used the GRADE approach to assess the certainty of evidence for each important outcome.
For the 2021 version, a peer reviewer suggested that we exclude any study not assessing plasma for time-adjusted paraquat concentration.
Whilst we did not agree to the extent of excluding studies as a whole, we did perform sensitivity analysis to exclude these studies for the
most important outcome (mortality).
INDEX TERMS
Medical Subject Headings (MeSH)

Bias; Cause of Death; Cyclophosphamide [*therapeutic use]; Drug Therapy, Combination [methods]; Glucocorticoids [*therapeutic
use]; Herbicides [*poisoning]; Immunosuppressive Agents [*therapeutic use]; Paraquat [*poisoning]; Poisoning [drug therapy]
[mortality]; Pulmonary Fibrosis [chemically induced] [*drug therapy] [immunology] [mortality]; Randomized Controlled Trials as
Topic; Time Factors
MeSH check words

Adult; Female; Humans; Male


Li Et Al-2021-Cochrane Database of Systematic Reviews PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Li Et Al-2021-Cochrane Database of Systematic Reviews PDF

Uploaded by

Copyright:

Available Formats

Li LR, Chaudhary B, You C, Dennis JA, Wakeford H

Li LR, Chaudhary B, You C, Dennis JA, Wakeford H.

Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) i

Glucocorticoid with cyclophosphamide for oral paraquat poisoning

Contact: Luying Ryan Li, lirang58@hotmail.com.

Editorial group: Cochrane Injuries Group.

Data collection and analysis

No studies assessed the outcome of mortality at 30 days following ingestion of paraquat.

- We are uncertain whether these medicines increase the risk of infection.

What happens in people with paraquat poisoning?

What did we want to find out?

Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) 2

What did we do?

What did we find?

All participants were given either:

- usual care only, or

Death while in hospital

Death 3 months after hospital discharge

What are the limitations of the evidence?

How up to date is this evidence?

Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) 3

Patient or population: people with moderate to severe oral paraquat poisoning

Mortality at 30 days fol- No included studies reported this outcome.

Cochrane Database of Systematic Reviews

CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio

GRADE Working Group grades of evidence

Cochrane Database of Systematic Reviews

How the intervention might work cyclophosphamide by themselves) or other combinations of

METHODS • the Cochrane Central Register of Controlled Trials (CENTRAL)

• China National Knowledge Infrastructure (CNKI 数据库)

Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) 7

Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) 8

Data extraction and management Assessment of risk of bias in included studies

Random sequence generation (selection bias)

0% 25% 50% 75% 100%

Low risk of bias Unclear risk of bias High risk of bias

Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) 9

Blinding (performance bias and detection bias): All outcomes

Selective reporting (reporting bias)

Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) 10

Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) 11

Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) 13

Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) 14

Glucoc & cyclophosphamide Standard care Risk Ratio Risk Ratio

1.1.1 Mortality at hospital discharge

1.1.2 Mortality at 3 months

Glucoc & cyclophosphamide Standard care Risk Ratio Risk Ratio

1.2.1 Mortality at hospital discharge

Test for subgroup differences: Not applicable 0.2 0.5 1 2 5

Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) 15

Glucoc & cyclophosphamide Standard care Risk Ratio Risk Ratio

1.3.1 Mortality at hospital discharge

Test for subgroup differences: Not applicable 0.2 0.5 1 2 5

Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) 16

Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) 17

Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) 18

Lin 2006 {published data only} Dinis-Oliveira 2006

Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) 19

Ginsberg 2005 Lee 1984

Higgins 2021 Smith 1975

Glucocorticoid with cyclophosphamide for oral paraquat poisoning (Review) 20

Xu 2019 Available at cccrg.cochrane.org/author-resources (accessed

Characteristics of included studies [ordered by study ID]