[ Original Research Signs and Symptoms of Chest Disease ]

Use of Management Pathways or Algorithms

in Children With Chronic Cough
Systematic Reviews
Anne B. Chang, MBBS, PhD; John J. Oppenheimer, MD; Miles Weinberger, MD; Kelly Weir, PhD; Bruce K. Rubin, MD;
and Richard S. Irwin, MD, Master FCCP

BACKGROUND: Use of appropriate cough pathways or algorithms may reduce the morbidity of
chronic cough, lead to earlier diagnosis of chronic underlying illness, and reduce unnecessary
costs and medications. We undertook three systematic reviews to examine three related key
questions (KQ): In children aged #14 years with chronic cough (> 4 weeks’ duration), KQ1,
do cough management protocols (or algorithms) improve clinical outcomes? KQ2, should the
cough management or testing algorithm differ depending on the duration and/or severity?
KQ3, should the cough management or testing algorithm differ depending on the associated
characteristics of the cough and clinical history?
METHODS: We used the CHEST expert cough panel’s protocol. Two authors screened
searches and selected and extracted data. Only systematic reviews, randomized controlled
trials (RCTs), and cohort studies published in English were included.
RESULTS: Data were presented in Preferred Reporting Items for Systematic Reviews and
Meta-analyses flowcharts and summary tabulated. Nine studies were included in KQ1
(RCT ¼ 1; cohort studies ¼ 7) and eight in KQ3 (RCT ¼ 2; cohort ¼ 6), but none in KQ2.
CONCLUSIONS: There is high-quality evidence that in children aged #14 years with chronic
cough (> 4 weeks’ duration), the use of cough management protocols (or algorithms) im-
proves clinical outcomes and cough management or the testing algorithm should differ
depending on the associated characteristics of the cough and clinical history. It remains
uncertain whether the management or testing algorithm should depend on the duration or
severity of chronic cough. Pending new data, chronic cough in children should be defined as
> 4 weeks’ duration and children should be systematically evaluated with treatment targeted
to the underlying cause irrespective of the cough severity. CHEST 2016; 149(1):106-119

KEY WORDS: children; cough; guidelines; systematic review

ABBREVIATIONS: BTS = British Thoracic Society; CHEST = American
College of Chest Physicians; KQ = key question; LR = likelihood ratio;
PV = predictive value; QoL = quality of life; RCT = randomized controlled
trial; TSANZ = Thoracic Society of Australia and New Zealand
AFFILIATIONS: From the Menzies School of Health Research; and Res-
piratory Deptartment (Dr Chang), Lady Cilento Children’s Hospital,
Queensland University of Technology, Queensland, Australia; Speech
Pathology Deptartment (Dr Weir), Lady Cilento Children’s Hospital,
Brisbane, Australia; Division of Allergy and Immunology (Dr Oppen-
heimer), UMDNJ-New Jersey Medical School, Cedar Knolls, NJ; Division
of Pulmonary, Allergy & Critical Care Medicine (Dr Irwin), UMass Me-
morial Medical Center, Worcester, MA; the Children’s Hospital of Rich-
mond at Virginia Commonwealth University (Dr Rubin), Richmond, VA;
and the Pediatric Allergy, Immunology, and Pulmonology Division
(Dr Weinberger), University of Iowa Children’s Hospital, Iowa City, IA.
FUNDING/SUPPORT: Dr Chang is supported by a National Health and
Medical Research Council (NHMRC) practitioner fellowship (grant
1058213) and holds multiple grants awarded from the NHMRC related
to diseases associated with pediatric cough. The views expressed in this
publication are those of the authors and do not reflect the views of the
NHMRC.
CORRESPONDENCE TO: Anne B. Chang, MBBS, PhD, Queensland
Children’s Respiratory Centre, Royal Children’s Hospital and Menzies
School of Health Research, Charles Darwin University, Herston,
Queensland, Brisbane, Queensland 4029, Australia; e-mail: annechang@
ausdoctors.net
Copyright Ó 2016 American College of Chest Physicians. Published by
Elsevier Inc. All rights reserved.
DOI: http://dx.doi.org/10.1378/chest.15-1403

106 Original Research [ 149#1 CHEST JANUARY 2016 ]

Cough is the most common symptom presenting to
primary care providers in many countries where data are
available, such as in the United States1 and Australia.2
Whereas most of these consultations are likely for acute
cough, a substantial proportion would be for chronic
cough. Indeed, chronic cough is one of the most
common presenting symptoms to respiratory specialty
physicians.
The burden of the symptom is significant in terms of
personal cost, with impaired quality of life (QoL)3,4 and
at a societal level at which physician visits, medication
expenses, absenteeism, and the tendency to work while
sick to avoid the stigma of being absent are substantial.
In an attempt to ameliorate cough, various prescription
and nonprescription medications are widely used.
Furthermore, the presence of chronic cough may reflect
an underlying serious disorder.4,5 Delayed diagnosis
(eg, foreign body) may cause chronic respiratory
morbidity6; early diagnosis of chronic diseases leading
to appropriate management and subsequent resolution
of cough and improved QoL4 is important. Thus, in the
evaluation of children with chronic cough, determining
which children require further investigations and/or
treatment is a key management strategy.
Materials and Methods
11
We undertook the systematic reviews based on the protocol
established by selected members of the CHEST expert cough panel.
We used the Preferred Reporting Items for Systematic Reviews and
Meta-analyses statement for reporting.
Study Identification and Eligibility Criteria
Librarians from the University of Massachusetts Medical School
undertook searches for all three questions between February 28 and
March 11, 2015, using the search strategies outlined in e-Table 1.
For the CHEST cough guidelines, it was determined a priori that the
age cutoff for pediatric and adult components was to be 14 years.
However, to ensure that all relevant studies were captured, the
search filter included studies up to age 18 years. We included only
studies published in English. The librarians identified and removed
duplicates between Scopus and PubMed searches before sending the
abstracts to the two authors (A. B. C. and J. J. O.) who reviewed the
abstracts.
Results
Preferred Reporting Items for Systematic Reviews and
Meta-analyses diagrams for all KQs are presented in
Figures 1-3. The risk of bias for RCTs included for KQ1
and KQ3 was combined into a single figure (Fig 4).
journal.publications.chestnet.org
Use of appropriate cough pathways or algorithms is one
such strategy that has the potential to reduce the morbidity
of chronic cough, lead to earlier diagnosis of chronic
underlying illness, and reduce the unnecessary costs and
adverse events from medications used. Indeed, the use of
guidelines, recommendations, and clinical pathways is
usually considered an important factor for improving
the quality of care and outcomes in the current era of
evidence-based medicine.7,8 Successful development of
clinical guidelines requires many strategies, including
endorsement from experts as well as determination of
the quality of the evidence.9,10
In this article, our overall aim was to evaluate the use of
management pathways or algorithms in children with
chronic cough. We undertook three systematic reviews
to examine three related key questions (KQs). In
children aged #14 years with chronic cough (> 4 weeks’
duration): (1) do cough management protocols (or
algorithms) improve clinical outcomes? (2) should
the cough management or testing algorithm differ
depending on the duration and/or severity? and
(3) should the cough management or testing algorithm
differ depending on the associated characteristics of
the cough and clinical history?
Data Extraction and Quality Assessment
The two reviewers independently reviewed all abstracts and fully
agreed on which full-text articles to retrieve to assess for potentially
eligible studies. It was planned that disagreements that could not be
resolved by consensus would be adjudicated by a third reviewer.
For randomized controlled trials (RCTs), the reviewers independently
assessed the risk of bias criteria using criteria in Cochrane Reviews. Criteria
used were random sequence generation (selection bias), allocation
concealment (selection bias), blinding of participants and personnel
(performance bias) to the study protocol, blinding of outcome assessment
(detection bias), incomplete outcome data (attrition bias), and selective
reporting (reporting bias). For cohort studies, data were extracted by a
single author (A. B. C.) and checked by a second (J. J. O.). In cohort studies,
we reported on the study’s setting, number enrolled and completing the
study, inclusion and exclusion criteria, and other factors (Tables 1-3)12-34
that we considered important for interpreting studies on chronic cough
specific to the KQs. These factors included an a priori definition for
diagnoses, how cough was measured and resolution defined, and whether
the period effect was considered. Reasons for these factors, considered
quality factors for pediatric cough studies, are published elsewhere.35
Key Question 1
Nine studies were included in the systematic review for
KQ1 (Fig 1, Table 1). A single systematic review17 was
eligible for inclusion, but as the review consisted of a sole
study,12 we described the study instead.
107
 ] Prospective Studies That Used Cough Management Protocols or Algorithms in Children (Key Question 1)
108 Original Research

TABLE 1

Diagnosis
No. Enrolled, Defined a
First Author/ Guideline Inclusion Criteria; No. Completed, Primary Priori; FU Period Effect Main Findings (Top 3
Year Country Setting Base Exclusion (Age, y) Outcome Length Considereda Diagnoses)
Randomized
controlled
trial or
systematic
review
Chang Australia Multicenter Modified Aged <18 y, > 4 270, 253 At 6 wk: Yes; FU: Yes, 2 wk Absolute risk reduction
et al12/ resp OPD CHEST wk cough, newly (mean, 4.5; Cough max 12 treatment with early use of
2013 200613 referred; excl: SD, 3.7) resolution mo for algorithm, 24.7%;
and chronic by cough Dx and 6 95% CI, 13-35.
TSANZ14 respiratory diary,15 mo post Mean difference
illness PC-QOL16,c Dx between groups in
PC-QOL, 0.6 (PBB,
asthma, resolved
without specific Dx)
McCallum et Australia Cochrane Data same as above
al17/2014 Review study14 as it was
sole study in review
Cohort studies
Asilsoy Turkey Single CHEST > 4 wk cough 108, 108 Cough, Partial; Not specified 94.5% success
et al18/ center, 200613 excl: none (mean, 8.4; unspecified FU: NR (asthma, PBB,
2008 pediatric reported range, UACS)
OPD 6-14)
Chang et al4/ Australia Multicenter, Modified Aged <18 y, > 4 346, 346 Cough Yes; FU: Yes, 2 wk Diagnoses differed in
2012b resp OPD CHEST wk cough, newly (mean, 4.5; resolution max 12 treatment age groups (PBB,
200613 referred; excl: SD, 3) by cough mo for asthma, resolved
[

and chronic diary,15 Dx and 6 without specific Dx)

149#1 CHEST JANUARY 2016

TSANZ14 respiratory PC-QOL16 mo post

illness Dx
Karabel Turkey Single CHEST > 4 wks cough 270, 270 Not Partial; Not specified Standardized
et al19/ center, 200613 excl: NM, cardiac, (mean, 6.5; described FU: 12 approach “increases
2014 resp OPD syndromes, RTI range, 7 mo mo possibility for fast
in past 4 wk to 17 y) and accurate
diagnoses .”(p152)
(asthma, asthma-
like, GERD)

(Continued)
]
journal.publications.chestnet.org

TABLE 1 ] (Continued)
Diagnosis
No. Enrolled, Defined a
First Author/ Guideline Inclusion Criteria; No. Completed, Primary Priori; FU Period Effect Main Findings (Top 3
Year Country Setting Base Exclusion (Age, y) Outcome Length Considereda Diagnoses)
Marchant Australia Single Modified > 3 wk cough 108, 103 Cough Yes; FU: Yes, 2 wk 91% success (PBB,
et al20/ centre, CHEST Aged <18 y (median, diary38 max 12 treatment, natural resolution,
2006 resp OPD 199821 Newly referred; 2.6 yr; IQR mo for bronchiectasis)
excl: NR 1.2-6.9) Dx, post
Dx NR
Rehman Pakistan Single Locally Aged 6-59 mo, 172, 161 Parents NR; FU: No Underlying cause
et al22/ center, designed > 4 wk cough; (summary reporting, until identified in 92% of
2009 pediatric algorithm excl: use of ACE not unspecified cough children (asthma,
OPD with inhibitors reported) resolved postviral,
Mantoux (max 18 tuberculosis)
test mo)
Spelman23/ Ireland 18 general Asthma > 4 wk; excl: 106, final, 81 Parents NR; FU: Unclear, By visit 1, 71% cough
1991 practices protocolf wheeze, (range, reporting, 20 wk visits 4 wk gone or much better
previously used 0.5-10) unspecified then 2 y apart (73% diagnosed as
bronchodilators asthma)
Usta et al24/ Turkey Single British Inclusion: NR; 156, 156 Cough, Partial; Not specified Diagnoses changed
2014 center, Thoracic excl:d (mean, 8.4; unspecified FU: max from initial to final
pediatric Society SD, 2.6) 18 mo assessment
allergy for Dx, (postnasal drip þ
OPD NR post asthma, postnasal
Dx drip, asthma, PBBe)

ACE ¼ angiotensin-converting enzyme; CHEST ¼ American College of Chest Physicians; Dx ¼ diagnosis; excl ¼ exclusion criteria; FU ¼ follow-up period; GERD ¼ gastroesophageal reflux disease; IQR ¼ interquartile
range; max ¼ maximum; NM ¼ neuromuscular; NR ¼ not reported; OPD ¼ respiratory outpatient; PBB ¼ protracted bacterial bronchitis; PC-QOL ¼ Parent Cough-Specific Quality of Life; resp ¼ respiratory;
RTI ¼ respiratory tract infection; TSANZ ¼ Thoracic Society of Australia and New Zealand; UACS ¼ upper airway cough syndrome.
a
Period effect considered (ie, use of time to response): temporal relationship between use of medication and outcome was defined a priori.
b
Children in this study were also in the randomized controlled trial.12
c
Improvement of $75% or total resolution according to parental reports and cough diary data for $3 days.
d
“Premature birth, neuromotor development retardation, development-growth retardation, chest wall deformity, a smoking habit, clubbing, cardiac disease, any known chronic disease and/or a pulmonary disease,
and those who could not cooperate in pulmonary function test.”24(p406)
e
Equal number of children with “asthma” and PBB.
f
First visit: Aged <2 years, given oral orciprenaline; aged >3 years, given oral salbutamol. Second visit: If “cough gone or much better” medications continued; if not, children given theophylline. Third visit: If “cough
gone or much better” medications ceased; if not, children given oral prednisolone. Fourth visit: If “cough gone or much better” medications ceased; if not, children given inhaled sodium cromoglycate. Fifth visit: If
“cough gone or much better” medications ceased; if not, children given oral prednisolone. For those receiving sodium cromoglycate, medications were continued.
109
TABLE 2 ] Pediatric Chronic Cough Guidelines Relating to Whether the Cough or Testing Algorithm Should Differ
Depending on Duration and/or Severity
Definition of
Chronic Cough, Severity Assessment Mentioned? Severity Used to
First Author/Year Country Society wk If So, How Determine Pathway?
Chang et al14/ Australia Thoracic Society of >4 Yes, nonspecific effects on No
2006 Australia and New child and parents,
Zealand expectations and burden
Chang and USA American College of >4 Yes, nonspecific effects on No
Glomb13/ Chest Physicians child and parents,
2006 expectations and burden
Gibson et al25/ Australia Australia Lung > 4 (pediatric Yes, nonspecific effects on No
2008 Foundation component) child and parents
Kohno et al26/ Japan Japanese Respiratory > 4 (pediatric No No
2006 Society component)
Shields et al27/ England British Thoracic Society >8 Yes, nonspecific effects on No
2008 child and parents
Leconte et al28/ Belgium Primary care >4 Yes, quality of life, No
2008 expectations,
nonspecific effects
Lu29/2014 China Multiple societies >4 No (based on translated No
article)
Zacharasiewicz Austria Austrian Society of >4 Yes, parental No
et al30/2014 Pediatrics, Austrian understanding
Society of
Pneumology

The studies were published between 1991 and 2014. All
but one23 were published within the past 10 years. There
was one RCT12; the rest were cohort studies. Three
studies4,12,23 were multicenter studies, and the rest (n ¼ 5)
were unicenter studies. Five4,12,18-20 were based on
previous American College of Chest Physicians (CHEST)
algorithms, one24 was based on the British Thoracic
Society (BTS) algorithm, one22 was performed in
Pakistan, one23 used a locally developed algorithm, and
one23 employed an asthma algorithm that used an oral
b2-agonist. Four studies were undertaken in respiratory
specialist clinics,4,12,19,20 two in general pediatrics,18,22 one
in an allergy clinic,24 and a single study23 performed in
general practice. All but one study22 were undertaken in
an upper middle- to high-income country.
A total of 1,266 children were enrolled in the studies
(children in the RCT12 who were also in the cohort
study4 were counted once), and data were available for
1,225 children. The mean or median age of children in
the studies was 4.5-8.4 years (range, 17 years), and study
cohort size ranged from 108 to 346.
Only three studies4,12,20 defined cough resolution a
priori and reported on how cough was measured. The
follow-up phase after diagnosis was not reported for
four studies, whereas the other four studies reported a
follow-up from 6 months4,12 to 2 years23 after resolution
of cough (Table 1).
All studies found that the use of cough algorithms was
beneficial, although the measurement of benefit used to
assess the studies varied. Only the RCT12 had a
comparator group. That study found that the
intervention group (early use of cough algorithm) had
significantly better QoL, a higher proportion of children
who were cough-free, and a shorter duration of cough
(after randomization) compared with the control group
at the study’s end point. Consistent with the RCT and
systematic review, all of the cohort studies reported high
success in achieving a diagnosis irrespective of the
algorithm used. The three most common diagnoses
varied among studies (Table 1), and only one study4
evaluated the effect of age on the most common
diagnoses.
Key Question 2
No eligible studies examined whether the cough-testing
algorithm should differ depending on the duration
and/or severity of cough (Fig 2). In the absence of
appropriate studies, we summarized the major current
guidelines in the past decade relating to this question
(Table 2). We restricted guidelines to the past decade

110 Original Research [ 149#1 CHEST JANUARY 2016 ]

journal.publications.chestnet.org

TABLE 3 ] Prospective Studies That Evaluated the Use of Characteristics of Cough and/or Clinical History to Determine Cough Management or Testing Algorithm
(Key Question 3)
Main No. Enrolled, Period Effect
Inclusion Criteria; Characteristic No. Completed Primary Considered; FU Main Findings Related to
First Author/Year Country Setting Guideline Base Exclusion Used (Age, y) Outcome Perioda Key Question
Randomized
controlled
trial or
systematic
review
Chang Australia Multicenter, Modified Aged <18 y, Nonspecific 270, 253 At 6 wk: Yes, 2 wk Nonspecific cough:
et al12/ resp OPD CHEST > 4 wk cough, and (mean, Cough treatment; “self-resolved”
2013 200613 and newly specific 4.5; SD, resolution FU: max was most common
TSANZ14 referred; excl: cough 3.7) by cough 12 mo for Specific cough: PBB,
chronic pointerse diary,15 Dx, 6 mo asthma were most
respiratory PC-QOL16,c post Dx common
illness
Marchant Australia Single Not Aged 0.5-18 y, Wet cough 50, 47 Cough Yes; FU: 2 Amoxicillin-
et al31/ center, applicable wet cough (mean, resolution wk clavulanate
2012 pediatric > 3 w, doctor- 1.8-2.8; by cough effective for wet
and resp observed wet IQR, diary38,c cough, BAL
OPD cough 0.9-5.3) findings in
subgroup indicate
PBB
Cohort studies
Asilsoy Turkey Single CHEST > 4 wk cough; Nonspecific 108, 108 Cough, Not “Watch and wait
et al18/ center, 200613 excl: none and (mean, unspecified specified; approach” as part
2008 pediatric reported specific 8.4; range, FU: NR of nonspecific
OPD cough 6-14) cough particularly
pointerse beneficial.
Chang Australiab Multicenter, Modified Aged < 18 y, Nonspecific 346, 326 Cough Yes, 2 wk Children with chronic
et al32/ resp OPD CHEST > 4 wk cough, and (mean, resolution treatment; dry cough without
2015 200613 and newly specific 4.5; SD, 3) by cough FU: max any cough
TSANZ14 referred; excl: cough diary,38 12 mo for pointers can be
chronic pointerse PC-QOL39 Dx, 6 mo safely treated
respiratory post Dx using watchful
illness waiting approach.
High positive LRs
of cough pointers

(Continued)
111
112 Original Research

TABLE 3 ] (Continued)
Main No. Enrolled, Period Effect
Inclusion Criteria; Characteristic No. Completed Primary Considered; FU Main Findings Related to
First Author/Year Country Setting Guideline Base Exclusion Used (Age, y) Outcome Perioda Key Question
Karabel Turkey Single CHEST > 4 wk cough; Nonspecific 270, 270 Not Not Use of specific cough
et al19/ center, 200613 excl: cardiac, and (mean, described specified; in this study less
2014 resp OPD neuromuscular specific 6.5; range, FU: 12 mo successful as 22
syndromes, cough 0.6-17) children in cohort
RTI in past pointerse had bronchiectasis
4 wk and were identified
through abnormal
chest radiograph
pathway
Marchant Australia Single Not > 3 wk cough, Nonspecific 100, 100 Cough Yes; FU: Most useful clinical
et al33/ center, applicable age < 18 y, and (median, resolution max 12 marker in
2006 resp OPD newly specific 2.8; IQR, by cough mo for Dx, predicting specific
referred; excl: cough 1.0-6.5) diaryc NR post Dx cough is presence
NR pointerse of daily moist
cough
Rehman Pakistan Single Locally Aged 6-59 mo, Signs of 172, 161 Parents No; FU: until Country-specific
et al22/ center, designed > 4 wk cough; heart (summary reporting, cough flowchart
2009 pediatric algorithm, excl: Use of disease age not unspecified resolved
OPD includes ACE inhibitors reported) (max 18
Mantoux mo)
test
Usta et al24/ Turkey Single British Inclusion: NR; Nonspecific 156, 156 Cough, Not Presence of rhinitis
2014 center, Thoracic excl:d and (mean, unspecified specified; and allergic salute
pediatric Society specific 8.4 y; FU: max considered as
allergy cough SD, 2.6) 18 mo for specific cough
[

OPD pointerse Dx, NR pointer. Asthma

149#1 CHEST JANUARY 2016

post Dx described as most

common cause of
nonspecific cough

LR ¼ likelihood ratio. See Table 1 legend for expansion of other abbreviations.

a
Period effect considered: temporal relationship between use of medication and outcome was defined a priori.
b
Children in this study were from the same study4; hence, only one article was presented.
c
Improvement of $75% or total resolution according to parental reports and cough diary data for $3 days.
d
“Premature birth, neuromotor development retardation, development-growth retardation, chest wall deformity, a smoking habit, clubbing, cardiac disease, any known chronic disease and/or a pulmonary disease,
and those who could not cooperate in pulmonary function test.”24(p331)
e
Differentiation of specific from nonspecific cough is the presence of cough pointers, first described in a review article.34
]
 Figure 1 – Selection of studies that
351 abstracts identified by searchersa addresses key question 1: In children
aged # 14 years with chronic cough
(> 4 weeks’ duration), do cough man-
agement protocols (or algorithms)
improve clinical outcomes? aSearches
331 studies excluded (as discussed in e-Table 1) were under-
taken by Nancy Harger, MLS, and Judy
Nordberg, MLS, Education and Clinical
Services Librarians working in the Uni-
20 full text retrieved for further versity of Massachusetts Medical School
evaluation of eligibility Library in Worcester, Massachusetts.

13 studies excluded. Reasons:

• Subjects aged ≥ 14 years (n = 6)
• Old Cochrane review (n = 1)
• Study protocol only (n = 1)
• Guidelines with no patient data (n = 3)
• Algorithm for cough not used (n = 2)

2 additional study identified

7 studies met inclusion criteria
from authors’ database

9 studies included in current systematic review (Table 1)

as there have been substantial changes in guideline
development36,37 and advances in treatment of chronic
cough during this period.
All but one guideline27 defined the duration of
chronic cough as > 4 weeks (Table 2). Most existing
nationally based cough guidelines suggest that cough
severity should be assessed, along with burden and
expectations. However, these recommendations were
generic, with little detail regarding how to assess cough
severity.
Key Question 3
Eight studies were included for KQ3 (Fig 3). Two were
RCTs12,31 and six were cohort studies (Table 3). Six
studies were also included in the KQ1 systematic review.
The studies were published between 2006 and 2015. Two
studies12,32 were multicenter studies and the rest (n ¼ 6)
were unicenter studies. Four studies were undertaken in
respiratory specialty clinics,12,19,32,33 one was in general
pediatrics,22 one was in an allergy clinic,24 and a single
study31 enrolled children from both general pediatric
and respiratory specialist clinics. All but one study22
were undertaken in an upper-middle to high-income
country.
A total of 1,182 children were enrolled in the studies
(children in the RCT12 who were also in the cohort
study32 were counted once), and data were available
for 1,168 children. Mean or median age of children
in the studies was 1.8-8.4 years (range, 17 years), and
cohort sizes ranged from 50 to 346. Four studies12,31-33
defined cough resolution a priori and reported on
how cough was measured. The follow-up phase after
diagnosis was not reported for many studies. Other
studies reported follow-up periods of 2 weeks,31
6 months,12,32 and 12 months19 after resolution of
cough (Table 3).
Collectively, all studies demonstrated that cough
management should differ depending on clinical history,
including cough characteristics. None of the studies used
an empirical approach. At a divergent point, all studies
used the concept of wet or productive cough but the
point of when this divergence occurred differed between
the algorithms. In studies based on the CHEST and BTS
algorithms, this point occurred early on, whereas in the
case of the study from Pakistan22 it occurred late, after
considering factors such as cardiac failure, exposure to
tuberculosis, Mantoux test results, and developmental
disorders.
One study31 used the presence of chronic wet cough
as the inclusion criteria in an RCT in which cough
resolution rates were significantly (P ¼ .016) higher in
children who received amoxicillin-clavulanate compared

journal.publications.chestnet.org 113
Figure 2 – Selection of studies that
addresses key question 2: In children 253 abstracts identified by searchersa
aged # 14 years with chronic cough
(> 4 weeks’ duration), should the cough
or testing algorithm differ depending on
the duration and/or severity? aSearches
(as discussed in e-Table 1) were under- 242 studies excluded
taken by Nancy Harger, MLS, and Judy
Nordberg, MLS, Education and Clinical
Services Librarians working in the Uni-
versity of Massachusetts Medical School 11 full text retrieved for further
Library in Worcester, Massachusetts. evaluation of eligibility

11 studies excluded. Reasons:

• Study did not study effect of duration
or severity on algorithm (n = 8)
• Guidelines with no patient data (n = 2)
• Algorithm for cough not used (n = 1)

0 studies met inclusion criteria

No study included in current

systematic review

with those who received placebo (difference between
proportions, 0.32; 95% CI, 0.08-0.56). A subgroup had
bronchoscopy performed before treatment and the
BAL data were consistent with protracted bacterial
bronchitis.31
Five studies were considered eligible as they used
algorithms based on those of the CHEST12,18-21 or BTS24
that included the use of clinical cough characteristics
and clinical history. In these algorithms, the presence of
cough pointers or indicators was used to define specific
and nonspecific cough. These cough pointers listed in
the CHEST algorithm include symptoms such as wet
cough, exertional dyspnea, and recurrent pneumonia
(Table 4).33 However, only two studies32,38 examined the
statistical measures of the performance of one or more
cough pointers (ie, specificity, sensitivity, predictive
value [PV], likelihood ratio [LR]). Summary data from

251 abstracts identified by searchersa

231 studies excluded

20 full text retrieved for further

Figure 3 – Selection of studies that evaluation of eligibility
addresses key question 3: In children
aged # 14 years with chronic cough
(> 4 weeks’ duration), should the cough
or testing algorithm differ depending on 12 studies excluded. Reasons:
the associated characteristics of the cough • Reviews (n = 7)
and clinical history? aSearches (as dis- • Association study (n = 4)
cussed in e-Table 1) were undertaken by • Case series (n = 1)
Nancy Harger, MLS and Judy Nordberg,
MLS, Education and Clinical Services
Librarians working in the University of 8 studies met inclusion criteria
Massachusetts Medical School Library in (Table 3)
Worcester, Massachusetts.

114 Original Research [ 149#1 CHEST JANUARY 2016 ]

 that the presence of cough pointers depends on the

Blinding of participants and personnel (performance bias)

recognition of symptoms and signs.
In light of these findings, we also reviewed chronic

Blinding of outcome assessment (detection bias)

cough guidelines published by national societies (Table 5)

Random sequence generation (selection bias)

in the past decade. All but one26 suggested a benefit to

Incomplete outcome data (attrition bias)

using cough pointers to distinguish specific from

Allocation concealment (selection bias)

nonspecific cough. However, while the principles were the

Selective reporting (reporting bias)

same, the exact steps differed.

Discussion
We undertook three related systematic reviews
involving using algorithms or pathways in the
management of chronic cough in children. We found
high-quality evidence that in children (aged #14
years) with chronic cough, the use of cough
management protocols improves clinical outcomes
(KQ1). We found no studies that addressed the
Chang 2013 + + ? + + +
question of whether the cough management or testing
Marchant 2012 + + + + + + algorithm should differ depending on the duration,
severity, or both in children with chronic cough
(KQ2). A review of published national guidelines
Figure 4 – Risk of bias summary: review authors’ (A. B. C. and J. J. O.’s)
judgments about each risk of bias item for the randomized controlled found that all studies (Table 2) mentioned assessing
trials (Chang et al12 and Marchant et al31) included (key questions 1 cough severity and all but one defined chronic cough
and 3). The figure was generated using Review Manager software
(RevMan, version 5.3; The Nordic Cochrane Centre, Cochrane Collab- in children as duration > 4 weeks. For KQ3, we
oration, Copenhagen, 2014). found high-quality evidence that in children
aged #14 years with chronic cough, the cough
management or testing algorithm should differ
the multicenter study32 showed that the presence of a depending on the associated characteristics of
specific cough pointer indicating a cause of chronic the cough and clinical history.
cough (as opposed to resolution without specific
treatment) had a sensitivity of 1.0 (95% CI, 0.98-1.0), Overall Completeness and Applicability of Evidence
specificity of 0.95 (95% CI, 0.82-0.99), positive PV These systematic reviews were limited by the few
of 0.99 (95% CI, 0.97-1.0), negative PV of 1.0 RCTs for KQ1 and KQ3 and no studies for KQ2.
(95% CI, 0.89-1.0), positive LR of 20 (95% CI, 5.18-77.21), When data were available, we found that findings
and negative LR of 0 (95% CI, 0-0.03). Summary data from RCTs were consistent with those from cohort
of individual cough pointers were reported,32 and the studies. The algorithms varied, but common to all
two most common symptoms were wet or productive (except for the asthma-specific algorithm23) was a
cough and wheeze, with ORs of 73.7 (95% CI, 10-544.2) systematic evaluation of the child and appropriate
and 21.2 (95% CI, 1.3-349.8), respectively, of finding treatment based on the underlying cause rather
a treatable cause. In the smaller unicenter study,33(p696) than an empirical approach based on treatment
the “OR for the presence of any historical pointers in of gastroesophageal reflux, upper airway cough
predicting specific cough was 5.28 (95% CI 1.23-22.73), syndrome, or asthma.
whereas the OR for doctor-observed moist cough was
The strongest evidence for using a chronic cough
9.34 (95% CI 3.49-25.03).”
pathway is that from CHEST,13 as there was an RCT
All studies used one or more cough pointers to and several cohort studies, whereas evidence for other
determine subsequent management steps. However, the pathways is restricted to single-cohort studies (Table 1).
reported occurrence of diagnoses of a serious underlying The variations in algorithms raise the question of
illness (eg, bronchiectasis, tuberculosis, cystic fibrosis) in whether setting specific algorithms should be used. The
one study19 of children who were categorized as having Pakistan algorithm22 focused on tuberculosis, signs of
nonspecific cough in the initial assessment suggested severe respiratory distress, and cardiac failure early in

journal.publications.chestnet.org 115
TABLE 4 ] Extended List of Cough Pointers (Modified From Previous Articles13,14,34)
Systemic Pulmonary
Cardiac abnormalities Chest pain
Digital clubbing Daily moist or productive cough
Failure to thrive Hemoptysis
Medications or drugs associated with chronic cough Abnormal cough characteristics (brassy, plastic bronchitis,
(angiotensin-converting enzyme inhibitors, illicit drug paroxysmal with/without posttussive vomiting, staccato,
use) cough from birth)
Neurodevelopmental abnormality Recurrent pneumonia
Fever Hypoxia/cyanosis
Immune deficiency (primary or secondary) Previous lung disease or predisposing causes (eg, neonatal lung
disease, foreign body aspiration)
Feeding difficulties Exertional dyspnea
History of contacts (eg, tuberculosis) Dyspnea at rest or tachypnea
Chest wall deformity
Auscultatory findings (eg, stridor, wheeze, crackles)
Chest radiograph abnormalities
Pulmonary function abnormalities

the stepwise approach. However, only four patients had was caused by tuberculosis, the child would receive an
cardiac failure (2.5%) and 14 had tuberculosis (8.7%). accurate diagnosis when the specific cough pathway was
The time to reach these diagnoses was not provided.22 If followed, but this would depend on accurate clinical
the CHEST algorithm was used in a child whose cough assessment.

TABLE 5 ] Pediatric Chronic Cough Guidelines Advocating Different Management Based on Associated
Characteristics of Cough and Clinical History
Mx Based on Cough
Characteristics? If So, Mx Based on Other Clinical History?
First Author/Year Country Society What? If So, What?
Chang et al14/ Australia Thoracic Society of Australia Yes, cough Yes, nonspecific and specific
2006 and New Zealand constructs and cough
pointers
Chang and United American College of Chest Yes, cough Yes, nonspecific and specific
Glomb13/ States Physicians constructs and cough
2006 pointers
Gibson et al25/ Australia Australian Lung Foundation Yes, cough Yes, nonspecific and specific
2008 constructs and cough
pointers
Kohno et al26/ Japan Japanese Respiratory Society Partial, algorithm Yes, age, generic medical
2006 not presented history
Leconte et al28/ Belgium Primary care Yes, cough Yes, nonspecific and specific
2008 constructs and cough
pointers
Lu29/2014 China Multiple societies Yes, cough Yes (based on translated
constructs article), nonspecific and
specific cough
Shields et al27/ England British Thoracic Society Yes, cough Yes, nonspecific and specific
2008 constructs and cough
pointers
Zacharasiewicz Austrian Austrian Society of Pediatrics, Yes Yes, unspecific and specific
et al30/2014 Austrian Society of cough
Pneumology

Mx ¼ management.

116 Original Research [ 149#1 CHEST JANUARY 2016 ]

All but one study23 employed the concept of looking for
associated features in the history, cough pointers, or both.
The use of cough pointers to differentiate specific from
nonspecific cough was first described in a review article.34
This was formally analyzed statistically in one study32
and the high positive LRs of cough pointers support the
use of individual cough pointers to identify the cause of
cough. However, recognition of cough pointers depends
on the ability to identify them (ie, expertise of physicians
and the caregiver’s history). For example, recognition
of wet cough depends on the setting; Brisbane-based
parents were accurate in determining the type of cough
(compared with pulmonologists [kappa ¼ 0.75; 95% CI,
0.58-0.93] and bronchoscopic findings),39 whereas
Indigenous caregivers were inaccurate.40 It is also likely
age-dependent, as it is easier to detect wet cough in young
children, and older children may have productive cough
but may have a dry cough when asked to cough (“cough
on request”).
Thus, the use of cough algorithms needs to be
accompanied by training. In primary care, abnormal
chest signs such as wheeze and crepitations are often
unreliable in young children (kappa for agreement
among primary care doctors, 0.39; 95% CI, 0.26-0.53).41
Therefore, the performance of the cough algorithms
likely depends on the setting (eg, study population,
expertise of the clinicians, study setting). The studies
included in the systematic reviews were performed
predominantly in specialty clinics, and so evidence of
the applicability of cough algorithms in primary care
remains limited.
Furthermore, all but one study included in the
systematic reviews were undertaken in upper middle- to
high-income countries (defined by the World Bank).
The frequency of diseases associated with chronic
cough in children, such as tuberculosis42 and parasitic
pulmonary diseases,43 depends at least partially on
the country’s wealth. For example, an Iranian study
suggested that chronic cough may be attributable to
eosinophilia related to Toxocara infection in 25% of
children.43 Also, the availability of the depth of tests
to determine an underlying disorder (such as a
videofluoroscopic swallowing study to determine
recurrent small-volume aspiration) depends on the
health system and/or setting.
Other possibilities for why all studies reported successful
cough algorithms include the period effect (natural
resolution of cough with time44), the bias in cohort studies
(compared with RCTs), and the lack of standardization
journal.publications.chestnet.org
when defining cough outcomes. Studies on cough can limit
the risk of bias of placebo and period effects by limiting the
time frame in which response to treatment is considered
(time to response).35 Of the studies included in the
systematic reviews (Tables 1, 3), fewer than half defined a
priori how cough was measured or what constituted cough
resolution and/or considered the period effect. Thus,
further high-quality studies using validated tools45 are
needed, particularly in primary care settings.
We used a broader search filter and included studies
that included children up to age 18 years, although the
cutoff for pediatric components of the CHEST cough
guidelines was age 14 years. Based on the mean and
median ages of the cohorts of children in the included
studies for all KQs (Tables 1, 3) that range from 1.8 to
8.4 years, it is highly unlikely that the broader search we
used affected the data.
Potential Biases in the Review Process
One of the authors was involved in the RCTs12,31 and
other cohort studies4,20,32,33 included in these systematic
reviews. However, we took steps to reduce potential
bias by having another author reviewing the extracted
data; the risk of bias assessments was undertaken
independently.
Implications for Practice
There is high-quality evidence that in children aged #14
years with chronic cough (> 4 weeks’ duration), the use
of cough management protocols or algorithms improves
clinical outcomes and cough management or testing
algorithms should differ depending on the associated
characteristics of the cough and clinical history.
There are no studies (and hence there is no evidence)
regarding whether cough management or testing
algorithms should depend on the duration or severity in
children aged #14 years with chronic cough. Pending
new data, chronic cough in children should be defined
as > 4 weeks’ duration, and children should be
systematically evaluated with treatment targeted to the
underlying cause irrespective of the severity of the cough.
Implications for Research
Multicenter parallel RCTs addressing all KQs are
needed, particularly in primary care. The use of
validated cough outcome measures45 is essential, and the
diagnostic criteria and definition of cough resolution
ascribed should be defined a priori. Ideally, an objective
cough outcome such as cough counts should be
included.
117
Acknowledgments
Author contributions: A. B. C. drafted the
manuscript and had full access to the data
and takes responsibility for the integrity
of all of the data and the accuracy of the
data analysis. A. B. C., J. J. O., and R. S. I.
drafted the key questions; A. B. C. and J. J. O.
selected the studies, extracted data, and
undertook the bias assessment; J. J. O. and
R. S. I. contributed to the data analyses
and interpretation and critically reviewed the
manuscript; and K. W., B. K. R., and
M. W. critically reviewed the manuscript.
Financial/nonfinancial disclosures: A. B. C.
has an intellectual conflict of interest because
she is the author of several of the articles
included in this review (as discussed in the
bias section of the article). J. J. O. is on the
Board of Directors of the American Board of
Allergy and Immunology; is associate editor
of Annals of Allergy and Allergy Watch; is
a reviewer for Up to Date; has performed
clinical research for BI, AZ, Glaxo,
MedImmune, and Novartis; is on the
adjudication committee for AZ and Novartis,
DSMB-Ohio State University; and is a
consultant for Glaxo, Myelin, Church and
Dwight, and Meda. R. S. I. reports that he has
no financial or intellectual conflicts of interest
regarding the content of the manuscript.
Moreover, while R. S. I. is the editor in chief
of CHEST, the review and all editorial
decisions regarding the manuscript were
independently made by others. All other
authors have reported to CHEST that no
potential conflicts of interest exist with any
companies/organizations whose products
or services may be discussed in this article.
Other contributions: We are grateful to
Nancy Harger, MLS, and Judy Nordberg,
MLS, education and clinical services
librarians working in the University of
Massachusetts Medical School Library in
Worcester, MA, who undertook all of the
searches for these systematic reviews.
Additional information: The e-Table can be
found in the Supplemental Materials section
of the online article.
References
1. Irwin RS. Introduction to the diagnosis
and management of cough: ACCP
evidence-based clinical practice
guidelines. Chest. 2006;129(1 suppl):
25S-27S.
2. Britt H, Miller GC, Charles J, et al.
General Practice Activity in Australia
1999-00 to 2008-09: 10 Year Data Tables.
General Practice Series No. 26. Cat. No.
GEP 26. Canberra, Australia: AIHW;
2010.
3. Marchant JM, Newcombe PA, Juniper EF,
et al. What is the burden of chronic
cough for families? Chest. 2008;134(2):
303-309.
4. Chang AB, Robertson CF, van Asperen PP,
et al. A multi-centre study on chronic
cough in children: burden and etiologies
118 Original Research
based on a standardized management
pathway. Chest. 2012;142:943-950.
5. Bush A. Paediatric problems of cough.
Pulm Pharmacol Ther. 2002;15(3):309-315.
6. Karakoc F, Karadag B, Akbenlioglu C,
et al. Foreign body aspiration: what is the
outcome? Pediatr Pulmonol. 2002;34(1):
30-36.
7. Wall RJ, Dittus RS, Ely EW. Protocol-
driven care in the intensive care unit: a
tool for quality. Crit Care. 2001;5(6):
283-285.
8. Preiser JC, Ledoux D. The use of protocols
for nutritional support is definitely needed
in the intensive care unit. Crit Care Med.
2004;32(11):2354-2355.
9. Wilson EJ, Nasrin D, Banwell C, et al.
Realities of practice: engaging parents
and GPs in developing clinical practice
guidelines. Aust Fam Physician.
2000;29(5):498-503.
10. National Health and Medical Research
Council. A Guide to the Development,
Implementation and Evaluation of Clinical
Practice Guidelines. Commonwealth of
Australia: National Health and Medical
Research Council; 1999.
11. Lewis SZ, Diekemper RL, French CT, et al.
Methodologies for the development of the
management of cough: CHEST guideline
and expert panel report. Chest.
2014;146(5):1395-1402.
12. Chang AB, Robertson CF, van Asperen PP,
et al. A cough algorithm for chronic cough
in children: a multicentre, randomized
controlled study. Pediatrics. 2013;131(5):
e1576-e1583.
13. Chang AB, Glomb WB. Guidelines for
evaluating chronic cough in pediatrics:
ACCP evidence-based clinical practice
guidelines. Chest. 2006;129(1 suppl):
260S-283S.
14. Chang AB, Landau LI, van Asperen PP,
et al. The Thoracic Society of Australia
and New Zealand. Cough in children:
definitions and clinical evaluation.
Med J Aust. 2006;184(8):398-403.
15. Chang AB, Newman RG, Carlin J, et al.
Subjective scoring of cough in children:
parent-completed vs child-completed
diary cards vs an objective method.
Eur Respir J. 1998;11(2):462-466.
16. Newcombe PA, Sheffield JK, Juniper EF,
et al. Development of a parent-proxy
quality-of-life chronic cough-specific
questionnaire: clinical impact
vs psychometric evaluations. Chest.
2008;133(2):386-395.
17. McCallum GB, Bailey EJ, Morris PS, et al.
Clinical pathways for chronic cough in
children. Cochrane Database Syst Rev.
2014;(9):CD006595.
18. Asilsoy S, Bayram E, Agin H, et al.
Evaluation of chronic cough in children.
Chest. 2008;134(6):1122-1128.
19. Karabel M, Kelekci S, Karabel D, et al. The
evaluation of children with prolonged
cough accompanied by American College
of Chest Physicians guidelines. Clin Respir J.
2014;8(2):152-159.
20. Marchant JM, Masters IB, Taylor SM,
et al. Evaluation and outcome of young
children with chronic cough. Chest.
2006;129(5):1132-1141.
21. Irwin RS, Boulet LP, Cloutier MM, et al.
Managing cough as a defense mechanism
and as a symptom: a consensus panel
report of the American College of Chest
Physicians. Chest. 1998;114(2 suppl
Managing):133S-181S.
22. Rehman A, Waraich MM, Ullah I.
Algorithm for the diagnosis of chronic
cough in children 6 to 59-months.
Pak Paed J. 2009;33(1):30-38.
23. Spelman R. Two-year follow up of the
management of chronic or recurrent
cough in children according to an asthma
protocol. Br J Gen Pract. 1991;41(351):
406-409.
24. Usta GB, Asilsoy S, Durmaz C. The
assessment and management of chronic
cough in children according to the British
Thoracic Society guidelines: descriptive,
prospective, clinical trial. Clin Respir J.
2014;8(3):330-337.
25. Gibson PG, Chang AB, Glasgow NJ, et al.
CICADA: Cough in Children and Adults:
Diagnosis and Assessment. Australian
Cough Guidelines summary statement.
Med J Aust. 2010;192(5):265-271.
26. Kohno S, Ishida T, Uchida Y, et al. The
Japanese Respiratory Society guidelines
for management of cough. Respirology.
2006;11(suppl 4):S135-S186.27.
27. Shields MD, Bush A, Everard ML, et al.
British Thoracic Society Guidelines:
recommendations for the assessment and
management of cough in children.
Thorax. 2008;63(suppl 3):iii1-iii15.
28. Leconte S, Paulus D, Degryse J. Prolonged
cough in children: a summary of the
Belgian primary care clinical guideline.
Prim Care Respir J. 2008;17(4):206-211.
29. Lu Q. Guideline for diagnosis and
treatment of chronic cough in Chinese
children [in Chinese]. Zhonghua Er Ke Za
Zhi. 2014;52(3):184-188.
30. Zacharasiewicz A, Eber E, Riedler J, et al.
Consensus statement on the evaluation
and therapy of chronic cough in children
[in German]. Wien Klin Wochenschr.
2014;126(13-14):439-450.
31. Marchant JM, Masters IB, Champion A,
et al. Randomised controlled trial of
amoxycillin-clavulanate in children with
chronic wet cough. Thorax. 2012;67(8):
689-693.
32. Chang AB, van Asperen PP, Glasgow N,
et al. Children with chronic cough: when
is watchful waiting appropriate?
Development of likelihood ratios for
assessing children with chronic cough.
Chest. 2015;147(3):745-753.
33. Marchant JM, Masters IB, Taylor SM,
et al. Utility of signs and symptoms of
chronic cough in predicting specific cause
in children. Thorax. 2006;61(8):694-698.
34. Chang AB. State of the art: cough, cough
receptors, and asthma in children. Pediatr
Pulmonol. 1999;28(1):59-70.
[ 149#1 CHEST JANUARY 2016 ]
35. Chang AB. Therapy for cough: where does
it fall short? Expert Rev Respir Med.
2011;5(4):503-513.
36. Lenzer J, Hoffman JR, Furberg CD, et al.
Ensuring the integrity of clinical practice
guidelines: a tool for protecting patients
[published online ahead of print
September 17, 2013]. BMJ. http://dx.doi.
org/10.1136/bmj.f5535.
37. Woolf S, Schunemann HJ, Eccles MP,
et al. Developing clinical practice
guidelines: types of evidence and
outcomes; values and economics,
synthesis, grading, and presentation and
deriving recommendations [published
online ahead of print July 4, 2012].
Implement Sci. http://dx.doi.org/10.1186/
1748-5908-7-61.
journal.publications.chestnet.org
38. Mazumder DN, Steinmaus C, 42. Alsan MM, Westerhaus M, Herce M, et al.
Bhattacharya P, et al. Bronchiectasis in Poverty, global health and infectious disease:
persons with skin lesions resulting from lessons from Haiti and Rwanda. Infect Dis
arsenic in drinking water. Epidemiology. Clin North Am. 2011;25(3):611-622.
2005;16(6):760-765.
43. Alavi SM, Sefidgaran G. Frequency of anti
39. Chang AB, Eastburn MM, Gaffney J, et al. Toxocara antibodies in schoolchildren with
Cough quality in children: a comparison chronic cough and eosinophilia in Ahwaz,
of subjective vs. bronchoscopic findings. Iran, 2006. Pak J Med Sci. 2008;24:360-363.
Respir Res. 2005;6:3.
44. Evald T, Munch EP, Kok-Jensen A.
40. Morey MJ, Cheng AC, McCallum GB, Chronic non-asthmatic cough is not
et al. Accuracy of cough reporting affected by inhaled beclomethasone
by carers of Indigenous children. dipropionate: a controlled double blind
J Paediatr Child Health. 2013;49(3): clinical trial. Allergy. 1989;44(7):510-514.
E199-E203.
45. Boulet L, Coeytaux RR, McCrory DC,
41. Hay AD, Wilson A, Fahey T, et al. The et al. Tools for assessing outcomes in
inter-observer agreement of examining pre- studies of chronic cough: CHEST
school children with acute cough: a nested guideline and expert panel report. Chest.
study. BMC Fam Pract. 2004;5(1):4. 2015;147(3):804-814.
119