European Journal of Cancer 103 (2018) 7e16

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.ejcancer.com

Review

Haematological toxicities with immunotherapy in

patients with cancer: a systematic review and
meta-analysis

Fausto Petrelli a,*, Raffaele Ardito b, Karen Borgonovo a,

Veronica Lonati a, Mary Cabiddu a, Mara Ghilardi a, Sandro Barni a

a
Oncology Unit, Medical Department, ASST Bergamo Ovest, Piazzale Ospedale 1, 24047, Treviglio BG, Italy
b
Oncological Day Hospital, IRCCS Centro di Riferimento Oncologico Della Basilicata (CROB), Via Padre Pio 1, 85028
Rionero in Vulture PZ, Italy

Received 20 January 2018; received in revised form 20 June 2018; accepted 24 July 2018

KEYWORDS Abstract Introduction: Programmed cell death-1 or ligand 1 (PD-(L)1) inhibitors are associ-
Anti-PD-(L)1; ated with immune-related adverse events. Conversely, little is known about the incidence of
Cancer; haematological toxicities across published trials. We have performed a systematic review
Toxicity; and meta-analysis to evaluate the incidence of immunotherapy-related anaemia, neutropenia
Anaemia; and thrombocytopenia among different tumour types, trials phases and anti-PD-(L)1 agents.
Neutropenia; Material and methods: A PubMed, Embase and Cochrane library search on 23rd December
Thrombocytopenia; 2017 and a review of references from relevant articles were done. Studies regarding haemato-
Meta-analysis logical diseases were excluded. The pooled incidence rates weighted for the individual sample
sizes were calculated according to fixed or random effect models. Incidence of all-grade and
grade (G) III or higher anaemia were the primary end-points. Neutropenia, febrile neutrope-
nia and thrombocytopenia were secondary end-points.
Results: Forty-seven studies of PD-(L)1 inhibitors for a total of 9324 evaluable patients were
included in the meta-analysis. The overall incidence of anaemia during PD-(L)1 inhibitor was
9.8% (95% confidence interval [CI], 6e13.6%) for all-grade and 5% (95% CI, 3.3e6.7%) for
G3-5 anaemia. The incidence was higher in diseases different from genitourinary, lung and
melanoma, with avelumab and in phase II studies. In randomised trials, relative risk of all-
grade anaemia for patients receiving anti-PD-(L)1 agents compared with control arms was
0.25 (95% CI, 0.16e0.39; p < 0.001). Incidence of all grades and G3-5 neutropenia and throm-
bocytopenia were 0.94%, 1.07%, 2.8% and 1.8%, respectively. Febrile neutropenia was 0.45%.

* Corresponding author: Fax: þ39 0363424380.

E-mail address: faupe@libero.it (F. Petrelli).

https://doi.org/10.1016/j.ejca.2018.07.129
0959-8049/ª 2018 Elsevier Ltd. All rights reserved.
8 F. Petrelli et al. / European Journal of Cancer 103 (2018) 7e16

Conclusions: The incidence of PD-(L)1 inhibitor-related anaemia was not negligible. Severe
neutropenia, thrombocytopenia and febrile neutropenia were rare. These findings are useful
for clinicians and suggest that blood cell count should be checked before every cycle and sup-
port should be given when severe toxicity appears.
ª 2018 Elsevier Ltd. All rights reserved.

1. Introduction EMBASE and Cochrane Library search. Also, references

Several anti-programmed cell death (ligand) protein 1
(anti-PD-[L]1) monoclonal antibodies have been recently
approved for the treatment of advanced cancers such as
non-small-cell lung cancer, genitourinary tumours and
melanoma. They are generally more tolerated than
chemotherapy, and their toxicities are manageable with
appropriate supportive therapies [1]. The use of agents
that block co-inhibitory immune checkpoint molecules,
such as PD-1, unleashes the immune system to control
malignancy [2]. However, this can lead to imbalances in
immunologic tolerance that result in an unexpected im-
mune response against self-tissues. This may clinically
manifest with autoimmune-like side-effects, such as
dermatological, gastrointestinal, hepatic, pulmonary and
endocrine toxicities [3,4]. Such adverse events, named
‘immune-related adverse events,’ are principally linked to
a dysregulated T-cell effect [5].
Although most clinical trials have reported haema-
tological toxicities (e.g. anaemia and neutropenia) of
immunotherapy, no systematic review or meta-analysis
reports the incidence of PD-(L)1 inhibitor-related hae-
matological toxicity across different solid tumours.
Because this is a relatively rare adverse event, aggre-
gated data from several prospective studies are note-
worthy for clinical practice. Given the increasing number
of published trials of PD-(L)1 inhibitors, such informa-
tion may provide valuable knowledge of these rare but
clinically significant toxicities. Anaemia, in fact, can
worsen fatigue, a very commonly reported adverse event
of immunotherapy, and similarly to neutropenia and
thrombocytopenia, has been described as a result of bone
marrow hypoplasia due to an autoimmune process [6].
We conducted a systematic review and meta-analysis of
trials of PD-(L)1 inhibitors in patients with cancer and
calculated the incidence of haematological toxicities
among different tumours types, agents and study phases.
2. Material and methods
2.1. Search methods and study selection
of published trials, review articles and editorials were
examined for other relevant articles. For the PubMed
search, the following terms were used: (PD-1[All Fields]
OR PD-L1[All Fields]) AND ((‘anaemia’[All Fields]
OR ‘anemia’[MeSH Terms] OR ‘anemia’[All Fields])
OR (‘thrombocytopaenia’[All Fields] OR ‘thrombocyto-
penia’[MeSH Terms] OR ‘thrombocytopenia’[All
Fields]) OR (‘neutropenia’[All Fields] OR ‘neu-
tropenia’[MeSH Terms] OR ‘neutropenia’[All Fields])
OR (‘haemoglobin’[All Fields] OR ‘hemoglobins’[MeSH
Terms] OR ‘hemoglobins’[All Fields] OR ‘hemoglobi-
n’[All Fields]) OR (‘blood platelets’[MeSH Terms] OR
(‘blood’[All Fields] AND ‘platelets’[All Fields]) OR
‘blood platelets’[All Fields] OR ‘platelets’[All Fields])
OR (‘febrile neutropenia’[MeSH Terms] OR (‘febri-
le’[All Fields] AND ‘neutropenia’[All Fields]) OR
(‘febrile neutropenia’[All Fields])). The databases were
searched for articles published from inception to 23rd
December 2017. Meeting abstracts without published
full-text original articles other than letters/case reports/
commentaries and articles not written in the English
language were not eligible for this study.
2.2. Data extraction
The total number of patients treated with PD-(L)1 in-
hibitors, the number of patients with anaemia for all
grades and for grade III or higher (G3-5), severe neu-
tropenia (G3-5), febrile neutropenia and severe throm-
bocytopenia (G3-5) were collected from the eligible
articles. Cases listed as both anaemia and haemoglobin
decrease were included in the number of anaemia events.
The trial phase, disease type and stage and types of
specific agents, doses and frequency of drug adminis-
tration were recorded. Treatment regimen was classified
as PD-(L)1 inhibitor monotherapy or combination
therapy. The data extraction were performed by one
primary reviewer (F.P.) and then independently
reviewed by two secondary reviewers (R.A. and S.B.)
following Preferred Reporting Items for Systematic
Reviews and Meta-Analyses guidelines.

Original articles that have published the results of clinical 2.3. Statistical analysis
trials of PD-(L)1 inhibitors (either monotherapy or
combination immunotherapy) including at least 20 pa- We extracted the number of patients experiencing
tients with solid tumours were identified by a PubMed, anaemia (primary end-point), neutropenia, febrile
 F. Petrelli et al. / European Journal of Cancer 103 (2018) 7e16
neutropenia and thrombocytopenia (secondary end-
points), and the total number of patients being treated
with the study drug and evaluable for toxicity, to
calculate incidence. Weights for each study in the anal-
ysis were based on the individual sample sizes. To assess
the stability of the results, a sensitivity analysis was
carried out by sequential omission of individual studies.
Between-study heterogeneity was estimated using the
c2-based Q test and I2 statistic. The random- or fixed-
effect models were used in the presence or absence of
heterogeneity. We considered an I2 value higher than
50% to be indicative of substantial heterogeneity [7]. We
also conducted the following pre-specified subgroup
analyses for all grade anaemia incidence: different PD-
(L)1 agents, type of treatment (monotherapy vs. com-
bination), phase of trial, study design (randomised vs.
not randomised studies) and tumour types (genitouri-
nary, lung, melanoma or other disease types). For the
randomised clinical trials, the relative risk (RR) of all
grade anaemia was calculated by comparing anaemia
rates in the experimental (immunotherapy) groups with
the anaemia rate in the control groups. Publication bias
was assessed using the Begg and Egger tests with funnel
plots [8,9].
All statistical analyses were performed using Meta-
Essential software, version 1.1.
Fig. 1. Literature search flow diagram.
9
3. Results
The search and the review of the reference lists identified
a total of 486 records for screening (Fig. 1). After
screening and eligibility assessment, a total of 386 du-
plicates or not pertinent paper (abstract, reviews, letter
or commentaries) were excluded. Other 46 publications
that referred to haematological conditions or not im-
mune checkpoint inhibitors drugs were further deleted.
Finally, among the 54 remaining publications, nine were
finally excluded because they had been recently updated
by newer versions (n Z 2) or did not report data relating
to haematological toxicities (n Z 7). Finally, n Z 47
studies (n Z 9324 patients) were finally included in a
systematic review of the incidence of haematological
toxicity (Table 1) [7e56]. Tumour types tested in these
studies included melanoma (n Z 7), lung cancer
(n Z 13) and other advanced solid tumours, including
mixed histology series, Merkel cell carcinomas, sar-
comas, head and neck and gastrointestinal cancers
(n Z 13) and genitourinary, including gynaecological
cancers (n Z 14). Among these, n Z 19 were phase I
studies, n Z 14 were phase II studies, n Z 12 were phase
III studies, one was a phase IeII trial and one was a
phase IIeIII trial. Drugs used in the 47 studies were as
follows: atezolizumab (n Z 6 studies), avelumab (n Z 2
haematological
randomised
10 F. Petrelli et al. / European Journal of Cancer 103 (2018) 7e16

Table 1
Characteristics of included trials.
Author/year Phase No of Stage Tumour type Drug Dose Schedule
pts
Alley/2017 1b 25 Locally advanced Mesothelioma Pembrolizumab 10 mg/kg Every 2 weeks
or metastatic
Antonia/2017 3 475 III NSCLC Durvalumab 10 mg/kg Every 2 weeks
Balar/2017 2 119 Locally advanced Urothelial Atezolizumab 1200 mg Every 3 weeks
or metastatic
Balar/2017 2 370 Locally advanced Urothelial Pembrolizumab 200 mg Every 3 weeks
or metastatic
Bauml/2017 2 171 Recurrent/metastatic Head & neck Pembrolizumab 200 mg Every 3 weeks
Bellmunt/2017 3 266 Advanced Urothelial Pembrolizumab 200 mg Every 3 weeks
Borghaei/2015 3 287 Advanced NSCLC Nivolumab 3 mg/kg Every 2 weeks
Brahmer/2012 1 207 Advanced Solid tumours BMS-936559 0.3e10 mg/kg Every 2 weeks
Brahmer/2015 3 131 Advanced NSCLC Nivolumab 3 mg/kg Every 2 weeks
Carbone/2017 3 267 Advanced NSCLC Nivolumab 3 mg/kg Every 2 weeks
El-Khoueiry/2017 1e2 48 Advanced HCC Nivolumab 0-1-10 / 3 mg/kg* Every 2 weeks
Fehrenbacher/2016 2 142 Advanced NSCLC Atezolizumab 1200 mg Every 3 weeks
Ferris/2016 3 236 Recurrent Head & neck Nivolumab 3 mg/kg Every 2 weeks
Frenel/2017 1b 24 Advanced Cervical Pembrolizumab 10 mg/kg Every 2 weeks
Garon/2015 1 495 Advanced NSCLC Pembrolizumab 2 & 10 mg/kg or Every 3 or 2 weeks
10 mg/kg
Gettinger/2015 1 129 Advanced NSCLC Nivolumab 1,3 or 10 mg/kg Every 2 weeks
Gulley/2017 1b 184 Metastatic or NSCLC Avelumab 10 mg/kg Every 2 weeks
recurrent
Hamanishi/2015 2 20 Platinum-resistant Ovarian Nivolumab 1 or 3 mg/kg Every 2 weeks
(90% III-IV)
Hamid/2013 1 135 Advanced Melanoma Lambrolizumab 2 & 10 or 10 mg/kg Every 3 or 2 weeks
Hamid/2017 2 357 Advanced Melanoma Pembrolizumab 2 & 10 mg/kg Every 2 weeks
Hellmann/2017 1 77 Advanced NSCLC Nivolumab þ 3 þ 1 & 3 þ 2 mg/kg Every 2/12 &
ipilimumab 2/6 weeks
Herbst/2016 2e3 682 Advanced NSCLC Pembrolizumab 2 & 10 mg/kg Every 3 weeks
Hodi/2016 2 94 Advanced Melanoma Nivolumab þ 1 & 3 mg/kg Every 3 weeks
ipilimumab
Hsu/2017 1b 27 Recurrent/metastatic Nasopharyngeal Pembrolizumab 10 mg/kg Every 2 weeks
carcinoma
Kaufman/2016 2 88 Metastatic Merkel cell carcinoma Avelumab 10 mg/kg Every 2 weeks
Le/2015 2 41 Metastatic Solid tumours Pembrolizumab 10 mg/kg Every 2 weeks
McDermott/2015 1 34 Advanced RCC Nivolumab 10 mg/kg Every 2 weeks
McDermott/2016 1a 70 Metastatic RCC Atezolizumab 10 or 15 or 20 mg/kg Every 3 weeks
or 1200 mg
Morris/2017 2 37 Metastatic Anal Nivolumab 3 mg/kg Every 2 weeks
Motzer/2015 3 410 Advanced RCC Nivolumab 3 mg/kg Every 2 weeks
Nanda/2016 1b 32 Advanced TNBC Pembrolizumab 10 mg/kg Every 2 weeks
Nghiem/2016 2 26 Advanced MCC Pembrolizumab 2 mg/kg Every 2 weeks
Ott/2017 1b 24 Advanced Endometrial Cancer Pembrolizumab 10 mg/kg Every 2 weeks
Ott/2017 1b 24 Extensive SCLC Pembrolizumab 10 mg/kg Every 2 weeks
Powles/2014 1 68 Metastatic UC Atezolizumab 15 mg/kg Every 2 weeks
Powles/2017 3 459 Metastatic UC Atezolizumab 1200 mg Every 3 weeks
Powles/2017 1/2 191 Advanced UC Durvalumab 10 mg/kg Every 2 weeks
Reck/2016 3 154 Advanced NSCLC Pembrolizumab 200 mg Every 3 weeks
Rizvi/2015 2 117 Advanced NSCLC Nivolumab 3 mg/kg Every 3 weeks
Robert/2014 1 173 Advanced Melanoma Pembrolizumab 2 or 10 mg/kg Every 3 weeks
Robert/2015 3 834 Advanced Melanoma Pembrolizumab 10 mg/kg Every 2 or 3 weeks
Ipilimumab 3 mg/kg Every 3 weeks
Rosenberg/2016 2 310 Advanced UC Atezolizumab 1200 mg Every 3 weeks
Sharma/2016 1/2 86 Metastatic UC Nivolumab 3 mg/kg Every 2 weeks
Tawbi/2017 2 84 Advanced Sarcomas Pembrolizumab 200 mg Every 3 weeks
Topalian/2012 1 296 Advanced Melanoma, RCC, Nivolumab 0.1 or 0.3 or 1 or 3
NSCLC, CRPC, CRC or 10 mg/kg
Weber/2015 3 272 Advanced Melanoma Nivolumab 3 mg/kg Every 2 weeks
Yamazaki/2017 1b 42 Advanced Melanoma Pembrolizumab 2 mg/kg Every 2 weeks
*, In phase I and II study; no of pts, number of patients; HCC, hepatocellular carcinoma; NSCLC, non-small-cell lung cancer;  , only
nivolumab þ ipilimumab arm; RCC, renal cell carcinoma; TNBC, triple negative breast cancer; MCC, merkel cell carcinoma; SCLC, small cell lung
cancer; UC, urothelial carcinoma; CRC, colorectal cancer.
 F. Petrelli et al. / European Journal of Cancer 103 (2018) 7e16
studies), durvalumab (n Z 2 studies), nivolumab
(n Z 14 studies), pembrolizumab (n Z 19 studies),
nivolumab þ ipilimumab (n Z 2 studies) and lam-
brolizumab and BMS 936559 (n Z 1 study each).
3.1. Incidence of anaemia
Forty-five studies were available for inclusion. The
incidence of anaemia ranged from 0.3% to 70.3% for all-
grade anaemia and from 0.1% to 17.1% for G3-5
anaemia. The pooled incidence rates for all-grade and
G3-5 anaemia were 9.8% (9% confidence interval [CI],
6e13.6%) (Fig. 2) and 5% (95% CI, 3.3e6.7%) (Fig. 3)
according to the random and fixed effect models,
respectively.
The estimated incidences of all grades of anaemia was
5.6% (95% CI, 3.1e8.2%) for atezolizumab, 14.7% (95%
CI, 14.6e14.8%) for avelumab, 5.8% (95% CI, 0-
1e37%) for durvalumab, 12.1% (95% CI, 2.8e21.4%)
for nivolumab and 9.1% (95% CI, 3.8e14.3%) for
pembrolizumab (P for difference <0.001). By the uni-
variate meta-regression model, including tumour type
alone as a predictor, the incidence of all grades of
anaemia in other sites (17.25%) was significantly higher
compared with melanoma (3.3%), lung cancer (7.2%) or
genitourinary cancers (7.8%) (P Z 0.001). The incidence
Fig. 2. Forest plots of the incidence of all-grade anaemia in cancer patients assigned anti-PD-(L1) agents.
11
of all grades of anaemia was also significantly different
in phase I (9.5%), phase II (16.8%) and phase III studies
(6%) (P Z 0.015). Randomised trials were associated
with less risk compared with non-randomised trials (6%
vs 11.6%; P Z 0.05). Only one study included the
combination of two agents (nivolumab þ ipilimumab),
and the rate of all grades of anaemia was similar to the
overall effect size (7.8%). However, after adjusting for
correlated incidence data and controlling for agents,
disease types, study design and trial phases, no covariate
was significantly associated with risk of anaemia. To
assess the relative rate of anaemia in PD-(L)1 compared
with the control arms, we calculated the RR of devel-
oping anaemia (all grades) in randomised clinical trials
in which immunotherapy was administered in the
experimental arm. Results for all-grade anaemia were
lower with anti-PD-(L)1 compared with conventional
agents, with a RR of 0.25 (95% CI, 0.16e0.39;
P < 0.001, I2 Z 84%).
3.2. Incidence of other haematological toxicities
Data on neutropenia (all grades, G3-5 and febrile neu-
tropenia) were available in n Z 17, n Z 16 and n Z 6
studies, respectively. Overall, incidence were 0.94 (95%
12 F. Petrelli et al. / European Journal of Cancer 103 (2018) 7e16
Fig. 3. Forest plots of the incidence of high-grade anaemia in cancer patients assigned anti-PD-(L1) agents.
CI, 0.43e1.46%), 1.07 (95% CI, 0.43e1.7%) and 0.45%
(95% CI, 0.84e1.74), respectively.
The incidence of thrombocytopenia was calculated
from n Z 13 trials. Overall, all grades and G3-5
thrombocytopenia were 2.84% (95% CI, 2.4e3.2%)
and 1.83% (95% CI, 1.18e2.48%).
3.3. Publication bias
Visualisation of the funnel plot for anaemia incidence
(all grades) (Fig. 4) shows minimal evidence of publi-
cation bias. Begg test was significant (P Z 0.049).
Instead, the Egger test is not significant (P
Egger Z 0.166). For G3-5 anaemia incidence, both the
Begg and the Egger tests are significant (P Begg Z 0.008
and P Egger Z 0.006) (Fig. 5).
4. Discussion
Haematological toxicities are commonly observed with
chemotherapy and targeted therapies such as small-
molecule tyrosine kinase inhibitors used in cancer pa-
tients [57]. These toxicities are usually the results of a
cytotoxic or cytostatic effect on haematopoietic stem
cells found in bone marrow, deputed to grow and
differentiate to produce the cellular elements of the
blood, including platelets, red blood cells and white
blood cells. Compared with cytotoxic or targeted agents,
immunotherapies have a distinct toxicity profile. They
can, in fact, induce infiltration of immune cells into
normal tissues, which leads to autoimmune-like toxic-
ities, different from the toxicities of traditional chemo-
therapy or targeted therapies. Compared with the
control agents in the randomised trials, they are asso-
ciated with an increased risk of colitis, rash, pneumo-
nitis, hepatitis and hypothyroidism [58,59].
Haematological toxicity such as anaemia or neu-
tropenia is poorly evaluated in clinical trials (focused
mainly on autoimmune events) but are a possible
complication of anti-PD-(L)1 agents in patients with
solid tumours and metastatic disease. A pooled analysis
for calculating the incidence of haematological adverse
events is a robust method to aggregate such data. This
is, to our knowledge, the first meta-analysis that evalu-
ates the rates of anaemia, neutropenia, thrombocyto-
penia and febrile neutropenia with immunotherapies.
We have selected only solid tumours and not haemato-
logical cancers treated with anti-PD(L)1 agents alone or
in combination with other immunotherapies (e.g. anti-
CTLA-4 drugs), thereby excluding the possible inter-
ference of chemotherapy agents in bone marrow re-
serves. Our study pooled data from 47 clinical trials and
 F. Petrelli et al. / European Journal of Cancer 103 (2018) 7e16 13

Fig. 4. Funnel plots for anaemia of all grades. CES, combined effect size.

Fig. 5. Funnel plots for anaemia of grade IIIeV. CES, combined effect size.

calculated an incidence of all-grade and high-grade
anaemia of 9.8% and 5%, respectively. We found
negligible risk of neutropenia (incidence of all grades
and G3-5 of 1%), febrile neutropenia (0.45%) and
thrombocytopenia (incidence of 2.8% and 1.8% for G1-5
and G3-5, respectively). In particular, the risk of all
grades of anaemia was not similar for various agents,
with the largest effects associated with avelumab
(14.7%), nivolumab (12%) and pembrolizumab (9%),
and the lowest effects associated with atezolizumab and
durvalumab (5.6% and 5%, respectively). Also, studies
of solid tumours other than genitourinary, lung and
melanoma (e.g. sarcomas, head and neck or gastroin-
testinal cancers) were those associated with a signifi-
cantly higher risk (17% vs 7.2%, 7.8% and 3% for
genitourinary, lung, and melanoma). Finally, phase II
studies were those with a higher incidence of anaemia
(16.8%) compared with phases I and III (9.5% and 3%,
respectively). Only one study included was a combina-
tion of two agents (nivolumab þ ipilimumab), but the
incidence of all-grade anaemia was similar (7.8%) to the
overall population. We found a high degree of hetero-
geneity among results, and this could be explained with
the different sets of studies (higher rates in phase IeII
trials), histology (cancers different from melanoma and
lung), control arm (type and combination of therapies)
and previous treatments (any radiotherapy or cytotoxic
agents). In particular, in phase III or other randomised
studies, patients with PS 2 and high disease burden, are
generally excluded from enrolment, and so a more
favourable haematological profile at baseline could be
expected.
If the results of neutropenia and thrombocytopenia
are expected to be very low, the incidence of anaemia is
14 F. Petrelli et al. / European Journal of Cancer 103 (2018) 7e16
moderately high, with 5% of patients suffering from
severe anaemia and possibly requiring transfusions. The
mechanisms underlying the risk of immunotherapy-
related anaemia are currently unknown, and multiple
distinct mechanisms could be involved. First, several
studies included patients with lung cancers and genito-
urinary cancers (with some studies enrolling gynaeco-
logical cancer patients). All these subjects could have
received a previous platinum-based chemotherapy, and
this could be the main reason for the higher incidence in
these patients compared with melanoma subjects. In the
phase III first-line trials of Reck et al. and Carbone et al.
[40,41], for example, anaemia was relatively low (3e5%)
with rare cases of severe anaemia (0e2%). Second, all
patients had locally advanced or recurrent/metastatic
disease, and this condition is associated with a higher
risk of anaemia. Active cancer, in fact, stimulates the
synthesis of many pro-inflammatory cytokines, inducing
an excess of hepcidin production with a consequent
status of chronic disease anaemia. Furthermore, phase
III and randomised studies enrolled patients in a better
general condition than those in phase IeII trials, and
this explains the increased risk compared with later-
phase trials. Third, most patients are previously heavily
pretreated with antineoplastic agents and possibly
radiotherapy for curative or palliative purposes. Most of
the studies (32 out of 47) were phase IeII trials, and
enrolled patients had been exposed to at least 2e3 or
more lines of previous anticancer therapies. This could
have reduced the bone marrow reserves in the red blood
cell compartment. Fourth, inclusion criteria usually
permit enrolment of patients with haemoglobin (Hb)
levels of at least 9e10 g/dL, so it is conceivable that the
mild levels of anaemia were preexisting and not neces-
sarily due to experimental drugs. Finally, a possible
autoimmune effect of immunotherapies cannot be
excluded. The pathogenesis of autoimmune phenomena
associated with this class of drugs is likely to be related
to direct activation of autoreactive T and B-cells, as well
as suppression of T-regulatory cells. In cases of
anaemia, activation of B-cells clones with a haemolytic
effect towards red blood cells seems to be possible, and
some case reports of autoimmunity are reported in the
literature with these drugs [60,61].
Haematological toxicities (in particular anaemia) are
an emerging complication of many targeted agents
and are not infrequently associated with
immunotherapy too. Recently, specific guidelines from
ASCO and Gustave Russy have been released with the
aim of help in the management and treatment of auto-
immune toxicities [62,63]. It is also crucial to identify
patients at higher risk and manage them appropriately.
A blood cell count before every cycle, evaluation of any
sudden fever or bleeding, iron parameters, vitamin
B12 and a folate check, as well as identification of other
causes of anaemia (e.g. bleeding, poor/reduced nutrition
status), should be clinically evaluated. Proper
monitoring (before and during treatment) and immedi-
ate management are crucial to maintain adequate Hb,
neutrophils and platelet levels and to avoid treatment
delay or transfusions because colony stimulating agents
are not labelled with these drugs.
Our study has some limitations. First, this is a study-
level meta-analysis and not based on individual patients’
data; therefore, a more comprehensive analysis, such as
adjusting for baseline factors including performance
status, age, previous therapies exposure and other im-
balances that exist between included trials, is not
possible. Second, the clinical trials included in our
analysis enrolled solid tumours with different risks of
haematological toxicities, even if, in this regard, we
attempted a subgroup analysis. Third, these studies were
mostly conducted at cancer centres or major academic
institutions, possibly excluding patients with poor hae-
matological function, so the incidence and severity of
anaemia and other toxicities analysed could be even
higher in daily clinical practice.
In conclusion, despite a previous attempt to aggre-
gate haematological toxicities (from n Z 7 studies) have
been done by Nishijima et al., in 2017 [64], our meta-
analysis is the first to have systematically estimated the
incidence of haematologic toxicities associated with
immunotherapy in cancer patients through a more
comprehensive search analysis. Anti-PD-(L)1 agents are
associated with a moderate risk of anaemia (about 10%)
and a low risk of neutropenia and thrombocytopenia
(0.9% and 2.8%), with negligible risk of febrile neu-
tropenia (0.45%). A periodical check for such toxicities
and an appropriate management is strongly needed for
the safe use of these drugs.
Further research is, however, necessary to clarify the
pathogenesis and risk factors of anti-PD(L)1-related
haematological toxicities.
Conflict of interest statement
None declared.
Acknowledgements
No funding to declare.
References
[1] Man J, Lee CK, Ritchie G, Links M, Lord S. Treatment-related
toxicities of immune checkpoint inhibitors in advanced cancers : a
meta-analysis. 2018;(November 2017). p. 141e52. https:
//doi.org/10.1111/ajco.12838.
[2] Dunn GP, Bruce AT, Ikeda H, Old LJ, Schreiber RD. Cancer
immunoediting: from immunosurveillance to tumor escape. Nat
Immunol 2002;3(11):991e8. https://doi.org/10.1038/ni1102-991.
[3] Eigentler TK, Hassel JC, Berking C, et al. Diagnosis, monitoring
and management of immune-related adverse drug reactions of
anti-PD-1 antibody therapy. Cancer Treat Rev 2016;45:7e18.
https://doi.org/10.1016/j.ctrv.2016.02.003.
 F. Petrelli et al. / European Journal of Cancer 103 (2018) 7e16
[4] Naidoo J, Page DB, Li BT, et al. Toxicities of the anti-PD-1 and
anti-PD-L1 immune checkpoint antibodies. Ann Oncol 2015;
26(12):2375e91. https://doi.org/10.1093/annonc/mdv383.
[5] Michot JM, Bigenwald C, Champiat S, Collins M. ScienceDirect
Immune-related adverse events with immune checkpoint
blockade : a comprehensive review. Eur J Cancer 2016;54:139e48.
https://doi.org/10.1016/j.ejca.2015.11.016.
[6] Akhtari M, Waller EK, Jaye DL, et al. Neutropenia in a patient
treated with ipilimumab (anti-CTLA-4 Antibody). J Immunother
2009;32(3):322e4. https://doi.org/10.1097/CJI.0b013e31819aa40b.
[7] Higgins JPT, Thompson SG. Quantifying heterogeneity in a
meta-analysis. Stat Med 2002;21(11):1539e58. https:
//doi.org/10.1002/sim.1186.
[8] Begg CB, Mazumdar M. Operating Characteristics of a rank
correlation test for publication bias. Author ( s ): Colin B . Begg
and Madhuchhanda Mazumdar Published by : International
Biometric Society Stable URL : Biometrics 1994;50(4):1088e101.
http://www.jstor.org/stable/2533446.
[9] Egger M, Smith GD, Schneider M, Minder C. Bias in meta -
analysis detected by a simple , graphical test. BMJ 2015;14:1e16.
https://doi.org/10.1136/bmj.315.7109.629.
[10] Herbst RS, Baas P, Kim DW, et al. Pembrolizumab versus
docetaxel for previously treated, PD-L1-positive, advanced non-
small-cell lung cancer (KEYNOTE-010): a randomised
controlled trial. Lancet 2016;387(10027):1540e50. https:
//doi.org/10.1016/S0140-6736(15)01281-7.
[11] Hellmann MD, Rizvi NA, Goldman JW, et al. Nivolumab plus
ipilimumab as first-line treatment for advanced non-small-cell
lung cancer (CheckMate 012): results of an open-label, phase 1,
multicohort study. Lancet Oncol 2017;18(1):31e41. https:
//doi.org/10.1016/S1470-2045(16)30624-6.
[12] Hamid O, Puzanov I, Dummer R, et al. Final analysis of a
randomised trial comparing pembrolizumab versus investigator-
choice chemotherapy for ipilimumab-refractory advanced mela-
noma. Eur J Cancer 2017;86:37e45. https://doi.org/10.1016/j.ejca.
2017.07.022.
[13] Balar AV, Galsky MD, Rosenberg JE, et al. HHS Public Access
2017;389(10064):67e76. https://doi.org/10.1016/S0140-6736(16)
32455-2. Atezolizumab.
[14] Antonia SJ, Villegas A, Daniel D, et al. Durvalumab after che-
moradiotherapy in stage III nonesmall-cell lung cancer. N Engl J
Med 2017. https://doi.org/10.1056/NEJMoa1709937. NEJMoa
1709937.
[15] Borghaei H, Paz-Ares L, Horn L, et al. Nivolumab versus doce-
taxel in advanced nonsquamous nonesmall-cell lung cancer. N
Engl J Med 2015;373(17):1627e39. https://doi.org/10.1056/
NEJMoa1507643.
[16] Long M, Tao S, Vega D, Jiang T, Wen Q, Sophia L. HHS Public
Access 2016;8(5):444e54. https://doi.org/10.1158/1940-
6207.CAPR-14-0359. Nrf2-dependent.
[17] Sharma P, Callahan MK, Bono P, et al. Nivolumab monotherapy
in recurrent metastatic urothelial carcinoma (CheckMate 032): a
multicentre, open-label, two-stage, multi-arm, phase 1/2 trial.
Lancet Oncol 2016;17(11):1590e8. https://doi.org/10.1016/S1470-
2045(16)30496-X.
[18] Bellmunt J, de Wit R, Vaughn DJ, et al. Pembrolizumab as
second-line therapy for advanced urothelial carcinoma. N Engl J
Med 2017;376(11):1015e26. https://doi.org/10.1056/NEJMoa
1613683.
[19] Robert C, Schachter J, Long GV, et al. Pembrolizumab versus
ipilimumab in advanced melanoma. N Engl J Med 2015;372(26):
2521e32. https://doi.org/10.1056/NEJMoa1503093.
[20] Bauml J, Seiwert TY, Pfister DG, et al. Pembrolizumab for
platinum- and cetuximab-refractory head and neck cancer: results
from a single-arm, phase II study. J Clin Oncol 2017;35(14):
1542e9. https://doi.org/10.1200/JCO.2016.70.1524.
[21] Robert C, Ribas A, Wolchok JD, et al. Anti-programmed-death-
receptor-1 treatment with pembrolizumab in ipilimumab-
15
refractory advanced melanoma: a randomised dose-comparison
cohort of a phase 1 trial. Lancet 2014;384(9948):1109e17. https:
//doi.org/10.1016/S0140-6736(14)60958-2.
[22] Balar AV, Castellano D, O’Donnell PH, et al. First-line pem-
brolizumab in cisplatin-ineligible patients with locally advanced
and unresectable or metastatic urothelial cancer (KEYNOTE-
052): a multicentre, single-arm, phase 2 study. Lancet Oncol 2017;
18(11):1483e92. https://doi.org/10.1016/S1470-2045(17)30616-2.
[23] Yamazaki N, Takenouchi T, Fujimoto M, et al. Phase 1b study of
pembrolizumab (MK-3475; anti-PD-1 monoclonal antibody) in
Japanese patients with advanced melanoma (KEYNOTE-041).
Cancer Chemother Pharmacol 2017;79(4):651e60. https:
//doi.org/10.1007/s00280-016-3237-x.
[24] Weber JS, D’Angelo SP, Minor D, et al. Nivolumab versus
chemotherapy in patients with advanced melanoma who pro-
gressed after anti-CTLA-4 treatment (CheckMate 037): a rando-
mised, controlled, open-label, phase 3 trial. Lancet Oncol 2015;
16(4):375e84. https://doi.org/10.1016/S1470-2045(15)70076-8.
[25] Tawbi HA, Burgess M, Bolejack V, et al. Pembrolizumab in
advanced soft-tissue sarcoma and bone sarcoma (SARC028): a
multicentre, two-cohort, single-arm, open-label, phase 2 trial.
Lancet Oncol 2017;18(11):1493e501. https://doi.org/10.1016/
S1470-2045(17)30624-1.
[26] Kaufman HL, Russell J, Hamid O, et al. Avelumab in patients
with chemotherapy-refractory metastatic Merkel cell carcinoma: a
multicentre, single-group, open-label, phase 2 trial. Lancet Oncol
2016;17(10):1374e85. https://doi.org/10.1016/S1470-2045(16)
30364-3.
[27] McDermott DF, Drake CG, Sznol M, et al. Survival, durable
response, and long-term safety in patients with previously treated
advanced renal cell carcinoma receiving nivolumab. J Clin Oncol
2015;33(18):2013e20. https://doi.org/10.1200/JCO.2014.58.1041.
[28] Le DT, Uram JN, Wang H, et al. PD-1 blockade in tumors with
mismatch-repair deficiency. N Engl J Med 2015;372(26):2509e20.
https://doi.org/10.1056/NEJMoa1500596.
[29] Morris VK, Salem ME, Nimeiri H, et al. Nivolumab for previ-
ously treated unresectable metastatic anal cancer (NCI9673): a
multicentre, single-arm, phase 2 study. Lancet Oncol 2017;18(4):
446e53. https://doi.org/10.1016/S1470-2045(17)30104-3.
[30] Motzer RJ, Escudier B, McDermott DF, et al. Nivolumab versus
Everolimus in advanced renal-cell carcinoma. N Engl J Med 2015;
373(19):1803e13. https://doi.org/10.1056/NEJMoa1510665.
[31] Nanda R, Chow LQM, Dees EC, et al. Pembrolizumab in pa-
tients with advanced triple-negative breast cancer: phase Ib
keynote-012 study. J Clin Oncol 2016;34(21):2460e7. https:
//doi.org/10.1200/JCO.2015.64.8931.
[32] Hsu C, Lee S-H, Ejadi S, et al. Safety and antitumor activity of
pembrolizumab in patients with programmed death-ligand
1epositive nasopharyngeal carcinoma: results of the
KEYNOTE-028 study. J Clin Oncol 2017;35(36). https:
//doi.org/10.1200/JCO.2017.73.3675. JCO.2017.73.367.
[33] Frenel J, Tourneau C Le, Neil BO, et al. Safety and ef fi cacy of
pembrolizumab in advanced , programmed death ligand 1 e
positive cervical Cancer : results from the phase Ib KEYNOTE-
028. Trial 2017;35(22):1e9. https://doi.org/10.1200/JCO.2017.
74.5471.
[34] Powles T, Eder JP, Fine GD, et al. MPDL3280A (anti-PD-L1)
treatment leads to clinical activity in metastatic bladder cancer.
Nature 2014;515(7528):558e62. https://doi.org/10.1038/nature
13904.
[35] McDermott DF, Sosman JA, Sznol M, et al. Atezolizumab, an
anti-programmed death-ligand 1 antibody, in metastatic renal cell
carcinoma: long-term safety, clinical activity, and immune corre-
lates from a phase Ia study. J Clin Oncol 2016;34(8):833e42.
https://doi.org/10.1200/JCO.2015.63.7421.
[36] Ott PA, Elez E, Hiret S, et al. Pembrolizumab in patients with
extensive-stage small-cell lung cancer: results from the phase Ib
16 F. Petrelli et al. / European Journal of Cancer 103 (2018) 7e16
KEYNOTE-028 study. J Clin Oncol 2017;35(34):3823e9. https:
//doi.org/10.1200/JCO.2017.
[37] Rizvi NA, Mazières J, Planchard D, et al. Activity and safety of
nivolumab, an anti-PD-1 immune checkpoint inhibitor, for patients
with advanced, refractory squamous non-small-cell lung cancer
(CheckMate 063): a phase 2, single-arm trial. Lancet Oncol 2015;
16(3):257e65. https://doi.org/10.1016/S1470-2045(15)70054-9.
[38] Nghiem PT, Bhatia S, Lipson EJ, et al. PD-1 blockade with
pembrolizumab in advanced merkel-cell carcinoma. N Engl J
Med 2016;374(26):2542e52. https://doi.org/10.1056/NEJMoa
1603702.
[39] Powles T, O’Donnell PH, Massard C, et al. Efficacy and safety of
durvalumab in locally advanced or metastatic urothelial carci-
noma. JAMA Oncol 2017;3(9):e172411. https://doi.org/10.1001/
jamaoncol.2017.2411.
[40] Reck M, Rodrı́guez-Abreu D, Robinson AG, et al. Pem-
brolizumab versus chemotherapy for PD-l1epositive nonesmall-
cell lung cancer. N Engl J Med 2016;375(19):1823e33. https:
//doi.org/10.1056/NEJMoa1606774.
[41] Carbone DP, Reck M, Paz-Ares L, et al. First-line nivolumab in
stage IV or recurrent nonesmall-cell lung cancer. N Engl J Med
2017;376(25):2415e26. https://doi.org/10.1056/NEJMoa1613493.
[42] El-Khoueiry AB, Sangro B, Yau T, et al. Nivolumab in patients
with advanced hepatocellular carcinoma (CheckMate 040): an
open-label, non-comparative, phase 1/2 dose escalation and
expansion trial. Lancet 2017;389(10088):2492e502. https:
//doi.org/10.1016/S0140-6736(17)31046-2.
[43] Powles T, Durán I, van der Heijden MS, et al. Atezolizumab
versus chemotherapy in patients with platinum-treated locally
advanced or metastatic urothelial carcinoma (IMvigor211): a
multicentre, open-label, phase 3 randomised controlled trial.
Lancet 2017;6736(17):1e11. https://doi.org/10.1016/S0140-
6736(17)33297-X.
[44] Brahmer J, Reckamp KL, Baas P, et al. Nivolumab versus
docetaxel in advanced squamous-cell nonesmall-cell lung cancer.
N Engl J Med 2015;373(2):123e35. https://doi.org/10.1056/
NEJMoa1504627.
[45] Jr B, Ss T, Lqm C. Supplementary Appendix. 2012. https:
//doi.org/10.1056/NEJMoa1200694. Supplementary.
[46] Garon EB, Rizvi NA, Hui R, et al. Pembrolizumab for the
treatment of nonesmall-cell lung cancer. N Engl J Med 2015;
372(21):2018e28. https://doi.org/10.1056/NEJMoa1501824.
[47] Hamid O, Robert C, Daud A, et al. Safety and tumor responses
with lambrolizumab (AntiePD-1) in melanoma. N Engl J Med
2013;369(2):134e44. https://doi.org/10.1056/NEJMoa1305133.
[48] Ferris RL, Blumenschein G, Fayette J, et al. Nivolumab for
recurrent squamous-cell carcinoma of the head and neck. N Engl
J Med 2016;375(19):1856e67. https://doi.org/10.1056/NEJMoa
1602252.
[49] Frenel J-S, Le Tourneau C, O’Neil B, et al. Safety and efficacy of
pembrolizumab in advanced, programmed death ligand
1epositive cervical cancer: results from the phase Ib KEYNOTE-
028 trial. J Clin Oncol 2017;35(36). https://doi.org/10.1200/
JCO.2017.74.5471. JCO.2017.74.547.
[50] Fehrenbacher L, Spira A, Ballinger M, et al. Atezolizumab versus
docetaxel for patients with previously treated non-small-cell lung
cancer (POPLAR): a multicentre, open-label, phase 2 randomised
controlled trial. Lancet 2016;387(10030):1837e46. https:
//doi.org/10.1016/S0140-6736(16)00587-0.
[51] Hamanishi J, Mandai M, Ikeda T, et al. Safety and antitumor
activity of Anti-PD-1 antibody, nivolumab, in patients with
platinum-resistant ovarian cancer. J Clin Oncol 2015;33(34):
4015e22. https://doi.org/10.1200/JCO.2015.62.3397.
[52] Hodi FS, Chesney J, Pavlick AC, et al. Combined nivolumab and
ipilimumab versus ipilimumab alone in patients with advanced
melanoma: 2-year overall survival outcomes in a multicentre,
randomised, controlled, phase 2 trial. Lancet Oncol 2016;17(11):
1558e68. https://doi.org/10.1016/S1470-2045(16)30366-7.
[53] Gulley JL, Rajan A, Spigel DR, et al. Avelumab for patients with
previously treated metastatic or recurrent non-small-cell lung
cancer (JAVELIN Solid Tumor): dose-expansion cohort of a
multicentre, open-label, phase 1b trial. Lancet Oncol 2017;18(5):
599e610. https://doi.org/10.1016/S1470-2045(17)30240-1.
[54] Gettinger SN, Horn L, Gandhi L, et al. Data supplement. J Clin
Oncol 2015;33(18):2004e12. https://doi.org/10.1200/JCO.
2014.58.3708.
[55] Alley EW, Lopez J, Santoro A, et al. Clinical safety and activity of
pembrolizumab in patients with malignant pleural mesothelioma
(KEYNOTE-028): preliminary results from a non-randomised,
open-label, phase 1b trial. Lancet Oncol 2017;18(5):623e30.
https://doi.org/10.1016/S1470-2045(17)30169-9.
[56] Topalian SL, Hodi FS, Brahmer JR, et al. Safety, activity, and
immune correlates of antiePD-1 antibody in cancer. N Engl J
Med 2012;366(26):2443e54. https://doi.org/10.1056/NEJMoa
1200690.
[57] Barni S, Cabiddu M, Guarneri P, Lonati V, Petrelli F. The risk
for anemia with targeted therapies for solid tumors. Oncologist
2012;17(5). https://doi.org/10.1634/theoncologist.2012-0024.
[58] Chen TW, Razak AR, Bedard PL, Siu LL, Hansen AR. A sys-
tematic review of immune-related adverse event reporting in
clinical trials of immune checkpoint inhibitors. Ann Oncol 2015;
26(9):1824e9. https://doi.org/10.1093/annonc/mdv182.
[59] De Velasco G, Je Y, Bossé D, et al. Comprehensive meta-analysis
of key immune-related adverse events from CTLA-4 and PD-
1/PD-l1 inhibitors in cancer patients. Cancer Immunol Res
2017;5(4):312e8. https://doi.org/10.1158/2326-6066.CIR-16-0237.
[60] Bagley SJ, Kosteva JA, Evans TL, Langer CJ. Immune throm-
bocytopenia exacerbated by nivolumab in a patient with non-
small-cell lung cancer. Cancer Treat Commun 2016;6:20e3.
https://doi.org/10.1016/j.ctrc.2016.02.009.
[61] Kong BY, Micklethwaite KP, Swaminathan S, Kefford RF,
Carlino MS. Autoimmune hemolytic anemia induced by anti-PD-
1 therapy in metastatic melanoma. Melanoma Res 2016;26(2):
202e4. https://doi.org/10.1097/CMR.0000000000000232.
[62] Champiat S, Lambotte O, Barreau E, et al. Management of im-
mune checkpoint blockade dysimmune toxicities: a collaborative
position paper. Ann Oncol 2016;27(4):559e74. https:
//doi.org/10.1093/annonc/mdv623.
[63] Brahmer JR, Lacchetti C, Schneider BJ, et al. Management of
immune-related adverse events in patients treated with immune
checkpoint inhibitor therapy: American Society of Clinical
Oncology clinical practice guideline. J Clin Oncol 2018;77:638.
https://doi.org/10.1200/JCO.2017.77.6385. JCO.2017.
[64] Management S, Care S. Safety and tolerability of PD-1/PD-L1
inhibitors compared with chemotherapy in patients with
advanced cancer: a meta-analysis. 2017. p. 470e9.