Karakike 

et al. Critical Care (2022) 26:183

https://doi.org/10.1186/s13054-022-04055-4

RESEARCH Open Access

Effect of intravenous clarithromycin

in patients with sepsis, respiratory and multiple
organ dysfunction syndrome: a randomized
clinical trial
Eleni Karakike1, Brendon P. Scicluna2,3,4, Maria Roumpoutsou1†, Ioannis Mitrou1†, Niki Karampela5,
Athanasios Karageorgos1, Konstantinos Psaroulis6, Eleni Massa7, Achillefs Pitsoulis8, Panagiotis Chaloulis9,
Evanthia Pappa10, Irene T. Schrijver11, Frantzeska Frantzeskaki12, Malvina Lada13, Nicolas Dauby14,15,
David De Bels16^, Ioannis Floros10, Souzana Anisoglou9, Eleni Antoniadou8, Maria Patrani5,
Glykeria Vlachogianni6, Eleni Mouloudi7, Anastasia Antoniadou1, David Grimaldi17, Thierry Roger11,
W. Joost Wiersinga2, Iraklis Tsangaris12 and Evangelos J. Giamarellos‑Bourboulis1* 

Abstract 
Background:  Clarithromycin may act as immune-regulating treatment in sepsis and acute respiratory dysfunction
syndrome. However, clinical evidence remains inconclusive. We aimed to evaluate whether clarithromycin improves
28-day mortality among patients with sepsis, respiratory and multiple organ dysfunction syndrome.
Methods:  We conducted a multicenter, randomized, clinical trial in patients with sepsis. Participants with ratio of
partial oxygen pressure to fraction of inspired oxygen less than 200 and more than 3 SOFA points from systems other
than the respiratory function were enrolled between December 2017 and September 2019. Patients were randomized
to receive 1 gr of clarithromycin or placebo intravenously once daily for 4 consecutive days. The primary endpoint
was 28-day all-cause mortality. Secondary outcomes were 90-day mortality; sepsis response (defined as at least 25%
decrease in SOFA score by day 7); sepsis recurrence; and differences in peripheral blood cell populations and leuko‑
cyte transcriptomics.
Results:  Fifty-five patients were allocated to each arm. By day 28, 27 (49.1%) patients in the clarithromycin and
25 (45.5%) in the placebo group died (risk difference 3.6% [95% confidence interval (CI) − 15.7 to 22.7]; P = 0.703,
adjusted OR 1.03 [95%CI 0.35–3.06]; P = 0.959). There were no statistical differences in 90-day mortality and sep‑
sis response. Clarithromycin was associated with lower incidence of sepsis recurrence (OR 0.21 [95%CI 0.06–0.68];
P = 0.012); significant increase in monocyte HLA-DR expression; expansion of non-classical monocytes; and

†
Maria Roumpoutsou and Ioannis Mitrou have contributed equally to this
work.
^David De Bels This author passed away.
*Correspondence: [email protected]
1
4th Department of Internal Medicine, National and Kapodistrian University
of Athens, 1 Rimini Street, 124 62 Athens, Greece
Full list of author information is available at the end of the article

© The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the
original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or
other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this
licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​
mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Karakike et al. Critical Care (2022) 26:183 Page 2 of 11

upregulation of genes involved in cholesterol homeostasis. Serious and non-serious adverse events were equally
distributed.
Conclusions:  Clarithromycin did not reduce mortality among patients with sepsis with respiratory and multiple
organ dysfunction. Clarithromycin was associated with lower sepsis recurrence, possibly through a mechanism of
immune restoration.
Clinical trial registration clinicaltrials.gov identifier NCT03​345992 registered 17 November 2017; EudraCT
2017-001056-55.
Keywords:  Macrolides, Clarithromycin, Sepsis, Multiple organ dysfunction, Recurrence, Cholesterol

Introduction Patients and methods

Sepsis mortality remains unacceptably high, reaching Ethics
26.7% for in-hospital cases and 41.9% for patients hos- INCLASS was a phase 3, multi-center, randomized, pla-
pitalized in the intensive care unit (ICU) [1]. It accounts cebo-controlled, double blind, clinical trial, conducted
for 19.7% of global deaths [2], and it is accompanied in thirteen study sites (11 multidisciplinary ICUs and 2
by considerable long-term morbidity [3]. Despite general Internal Medicine wards) in tertiary, teaching
early recognition, timely antimicrobial administration hospitals in Greece and Belgium (additional information
and organ support, further adjunctive therapies are in  Additional files 1 and 2). The protocol and informed
required [4]. However, the majority of potential inter- consent form were approved in Greece [National Organi-
ventions targeting the host immune response yielded zation for Medicines (51239/01-06-2017), National Eth-
conflicting results [5–11], possibly due to incomplete ics Committee (52086/2017)] and Belgium [Federal
understanding of underlying pathophysiological mech- Agency of Medicines and Health Products (1078386/16-
anisms [12]. 04-2018), Central Ethics Committee, Erasme University
Macrolides may exert immune-modulating effects Hospital (P2018/376, 19-10-2018)]. Study registration
[13], as demonstrated for the exacerbation of chronic was with EudraCT (2017-001056-55) and Clinicaltri-
obstructive pulmonary disease [14] and for the man- als.gov (NCT03345992). Written informed consent was
agement of community-acquired pneumonia (CAP) provided by patients or legal representatives, prior to
[15, 16]. Combination therapy with β-lactams leads to inclusion. The complete study protocol and history of
decreased mortality and is currently recommended as amended versions are provided in Additional files 1 and
first-line treatment in CAP [16]. Our group investigated 2.
the immune-modulating properties of clarithromycin
in severe infections caused by bacteria outside the mac- Participants
rolide antimicrobial spectrum, i.e., ventilator-associ- Adults with sepsis and multiple organ dysfunction syn-
ated pneumonia (VAP) and Gram-negative infections. drome (MODS) were eligible to participate. Inclusion
Results showed that adjunctive clarithromycin, com- criteria were sepsis associated with hospital-acquired
pared to placebo, improved survival among the most (HAP), healthcare-associated pneumonia (HCAP), VAP,
severely ill patients, particularly those with acute res- primary Gram-negative bacteremia or intra-abdominal
piratory dysfunction syndrome (ARDS) [17, 18]. This infection; ­PaO2/ ­FiO2 < 200; and total Sequential Organ
benefit has been associated with improving signs of Failure Assessment (SOFA) score for non-respiratory
sepsis-induced immunosuppression, a hallmark of the organ dysfunctions more than 3. Main exclusion criteria
protracted course of healthcare-associated infections were pregnancy or lactation, neutropenia (< 1000 neu-
[19]. This was characterized by higher production of trophils/ ­mm3), recent high-dose corticosteroid intake,
interleukin (IL)-6 from circulating monocytes, decrease macrolide allergy and macrolide intake for the current
in the ratio of circulating IL-10 to TNFα (tumor necro- infection. A complete list of exclusion criteria and defini-
sis factor-alpha) and increase in CD86 expression on tions is provided in Additional file 1.
circulating monocytes.
Based on those observations, we initiated the INtra- Randomization and intervention
venous CLArithromycin in Sepsis and multiple organ Patients were assigned to blind treatment with clarithro-
dysfunction Syndrome (INCLASS) study with the aim to mycin or placebo, following a random allocation
assess the adjunctive role clarithromycin on top of stand- sequence, with a 1:1 design and by block sizes of 10,
ard-of-care treatment for high-risk patients with health- stratified per study site. The allocation sequence was gen-
care-associated infections and sepsis. erated by an independent statistician prior to the study
 Karakike et al. Critical Care (2022) 26:183 Page 3 of 11

commencement and delivered to each study site within time to new sepsis episode; and differences in cell popu-
sealed individual envelopes, labeled as per study par- lations and gene expression.
ticipant code. The envelope was unsealed by the study
pharmacist, to prepare the study drug. All other parties Statistical analysis
involved (investigators, patients, healthcare providers, The sample size was calculated for the primary end-
data collectors) were blinded to the study arm. point, anticipating a 55% 28-day mortality, among con-
Patients were randomized to intravenous clarithromy- trol patients with sepsis and respiratory dysfunction, and
cin (1gr dissolved into 20 ml water for injection and then 25% reduction with clarithromycin [17, 18]. To detect
diluted to a final volume of 250 ml 5% dextrose in water) this difference, with 80% power at 10% significance level,
or placebo (equal volume of water for injection diluted to 55 patients were required in each study arm. No interim
a final volume of 250  ml 5% dextrose in water), infused analysis was planned.
once daily within 1 h, for four consecutive days. The final Clinical outcomes were assessed on an intention-to-
preparations were visually similar. Other therapies were treat principle, according to the pre-defined statistical
left at the discretion of the attending physicians. There analysis plan (Additional file  2). The Fisher’s two-sided
was no specific time window from enrollment to study exact test, confirmed by logistic regression analysis,
drug administration, which was required to be as soon as was used to assess the primary outcome. Kaplan–Meier
logistically possible. curves were used to assess survival by arm, and the
arm effect on 28-day mortality was evaluated with Cox-
regression model. Univariate and stepwise multivariable
Procedures logistic regression analyses were done using pre-specified
Patients were followed-up daily until day 28 or hospital co-variables (age, sex, SOFA, Acute Pathophysiology
discharge (whichever came first). In case of earlier dis- and Chronic Health Evaluation—APACHE II, Charlson
charge, visit of day 28 was performed by phone call, to comorbidity index—CCI, and adequacy of empirical anti-
assess the primary outcome and safety. Another phone microbial treatment) associated with 28-day mortality.
call was performed at day 90 with the patient or their car- The treatment effect on secondary outcomes was com-
egiver, to assess survival. Data were captured in one Case pared with the Pearson Chi-square test, or the two-sided
Report Form; source data verification was performed by Fisher’s exact test, if categorical, whereas quantitative
trained clinical research associates. Adverse events were variables were assessed using Student’s t test, or Mann–
captured daily until day 28 or discharge, and as patient Whitney U test, as appropriate. Subgroup analyses were
or provider-reported events up to day 90. National Can- pre-planned and are shown in Additional file 1.
cer Institute Common Terminology Criteria for Adverse A number of post hoc multivariable regression analyses
Events, version 5.0 (2017), were used for classification. were performed with regard to the primary outcome, to
Blood cell populations and monocyte human leuko- explore whether a simplified model, adapted for the study
cyte antigen (mHLA)-DR expression were measured on sample size (including SOFA, CCI, appropriateness of
days 1, 5 and 10, using flow cytometry. Total ribonucleic empirical antimicrobial treatment), early versus late start
acid (RNA) was isolated from PAXgene blood RNA tubes of treatment and co-administered antimicrobials would
(Qiagen), collected on days 1 and 5, using PaxGene Blood impact on 28-day mortality. The impact of enrollment
miRNA kits according to the manufacturer’s instructions site was assessed by the Breslow–Day test, and hetero-
(Qiagen). RNA-sequencing libraries were prepared using geneity was further evaluated by the I2 statistic. Post hoc
KAPA RNA Hyperprep with RiboErase (Roche) kits. Cox- and Poisson regression analyses were performed to
Libraries were sequenced using the Illumina HiSeq4000 assess the cumulative incidence of new sepsis and further
instrument (Illumina). All day 1 sampling was performed explore study outcomes (IBM SPSS statistics, version
before study drug administration. Detailed methodology 24.0). Any two-tailed p < 0.05 was considered significant.
and bioinformatics are shown in Additional file 1. No adjustment was performed for multiple compari-
sons, and secondary endpoints should be interpreted as
Outcomes exploratory.
The primary outcome was 28-day all-cause mortal- RNA-sequencing data analysis, as well as the I2 statistic
ity. Secondary outcomes were 90-day mortality; 28-day for study site heterogeneity, was performed in R (version
mortality in septic shock; early sepsis response (≥ 25% 3.51, R Core Team 2014). Additional information is pro-
decrease in day 3 SOFA score from baseline); sepsis vided in Additional file 1.
response (≥ 25% decrease in day 7 SOFA score from base-
line); new sepsis among patients with sepsis response; the
Karakike et al. Critical Care (2022) 26:183 Page 4 of 11

Results December 20, 2017, and the last visit of the last par-
Baseline characteristics ticipant was on September 22, 2019. Main reasons for
From December 2017 to June 2019, 241 patients were exclusion of patients were respiratory ratio ≥ 200, non-
assessed for eligibility and 110 were randomized to eligible infections and immunosuppression (Fig.  1).
blind treatment. The first patient was enrolled on Fifty-five patients were included in each study arm

Fig. 1  CONSORT Flow Diagram in INCLASS trial. SOFA, Sequential Organ Failure Assessment
 Karakike et al. Critical Care (2022) 26:183 Page 5 of 11

within median 2 (1–4) days after meeting inclusion cri- common infection (Table 1; Additional file 1: Tables S1
teria and 4 (1–6) days after sepsis onset. One patient and S2). Enrollment by participating site is shown in
(1.8%) in the clarithromycin arm did not receive any Additional file 1: Table S3.
dose of study drug due to early death, and another
patient (1.8%) in the clarithromycin arm did not receive Primary outcome
the fourth dose of the study drug due to protocol devi- By day 28, 27 (49.1%) patients in the clarithromycin
ation judged to be a  study team error. There were no group and 25 (45.5%) in the placebo group had died,
losses to follow-up, nor consent withdrawals, and all yielding an absolute difference in mortality risk of 3.6%
110 patients were included in the analysis of the pri- (95%CI − 15.7 to 22.7; P = 0.703); the unadjusted odds
mary endpoint and the 90-day outcomes. ratio (OR) for clarithromycin relative to placebo was 1.16
Overall, baseline demographic and clinical character- (95% CI 0.55–2.45; P = 0.849) (Table  2). This was con-
istics were similar between study arms; patients were firmed by survival analysis (Fig. 2A). After adjustment for
predominantly male, with high comorbidity burden and covariates, the OR was 1.03 (95%CI 0.35–3.06; P = 0.604)
high SOFA and APACHE II scores. VAP was the most (Additional file  1: Table  S4), and the model was able to

Table 1  Baseline characteristics of patients enrolled in the INCLASS study

Clarithromycin (n = 55) Placebo (n = 55) Total (n= 110) P value

Age (years), median (Q1–Q3) 74 (67–80) 73 (60–79) 74 (62–80) 0.354

Sex, n (%) 0.316
 Male 39 (70.9) 33 (60.0) 72 (65.5)
 Female 16 (29.1) 22 (40.0) 38 (34.5)
Charlson comorbidity index, mean (SD) 5.3 (2.8) 5.6 (2.8) 5.4 (2.8) 0.496
APACHE II score, mean (SD) 20.0 (6.4) 22.0 (7.1) 21.0 (6.8) 0.193
SOFA score, median (Q1-Q3) 10 (9–12) 11 (9–13) 10 (9–12) 0.364
Presence of ARDS, n (%) 23 (41.8) 25 (45.5) 48 (43.6) 0.848
Presence of septic shock, n (%) 30 (54.5) 39 (70.9) 69 (62.7) 0.114
Laboratory values, median (Q1-Q3)
 Lactate, mmol/l 1.8 (1.4–2.6) 2.1 (1.4–3.0) 2.0 (1.4–2.8) 0.459
 ­PaO2/FiO2 142 (111–171) 149 (116–167) 144 (114–165) 0.886
 White blood cell count, × ­103/mm3 15.9 (10.8–20.8) 16.7 (10.9–21.4) 16.2 (10.9–20.9) 0.832
 Platelet count, × ­103/mm3 130 (196–264) 197 (138–278) 197 (137–268) 0.907
 Creatinine, mg/dl 1.80 (1.05–2.41) 1.40 (0.80–2.40) 1.60 (0.87–2.4) 0.140
 CRP, mg/l 169.0 (91.5–259.4) 158.0 (66.2–233.8) 162.9 (81.7–254.8) 0.543
 PCT, ng/ml 1.42 (0.48–9.54) 1.36 (0.55–5.43) 1.42 (0.55–5.71) 0.900
Organ support at enrollment
 Mechanical Ventilation, n (%) 45 (81.8) 44 (80.0) 89 (80.9) 1.00
 ­PEEPa, median (Q1-Q3) 8 (6–10) 8 (7–10) 8 (6–10) 0.524
  Tidal ­Volumea, median (Q1-Q3) 502 (460–550) 500 (450–550) 500 (455–550) 0.638
 Noradrenaline use, n (%) 54 (98.2) 55 (100.0) 99 (90.0) 1.00
 Renal Replacement Therapy, n (%) 8 (14.5) 9 (16.4) 17 (15.5) 1.00
Source of sepsis, n (%)
 Lower respiratory tract infection 34 (61.8) 41 (74.5) 75 (68.2) 0.219
 Healthcare-associated 12 (21.8) 11 (20.0) 23 (20.9) 1.00
 Hospital-acquired 6 (10.9) 15 (27.3) 21 (19.1) 0.051
 Ventilator-associated 16 (29.1) 15 (27.3) 31 (28.2) 1.00
Intra-abdominal infection 15 (27.3) 12 (21.8) 27 (24.5) 0.658
Primary Gram-negative bacteremia 6 (10.9) 2 (3.6) 8 (7.3) 0.271
9 9
SI conversion factors: To convert white blood cell count to ­10 /L, multiply by 1; platelet count to ­10 /L, multiply by 1; and (serum) creatinine to μmol/L, multiply by 88.4
APACHE, Acute Pathophysiology and Chronic Health Evaluation; ARDS, acute respiratory distress syndrome; CRP, C-reactive protein; PCT, procalcitonin; PEEP, positive
end-expiratory pressure; SD, standard deviation; and SOFA, Sequential Organ Failure Assessment
a
Refers to patients under mechanical ventilation
Karakike et al. Critical Care (2022) 26:183 Page 6 of 11

Table 2  Study clinical outcomes

Clarithromycin (n = 55) Placebo (n = 55) OR for clarithromycin P value

Primary outcome, n (%, 95% CI)

 Mortality at 28 days 27 (49.1, 35.5–62.8) 25 (45.5, 32.2–59.3) 1.157 (0.547–2.448) 0.849
Secondary outcomes
 Mortality at 90 days, n (%, 95% CI) 41 (74.5, 60.7–84.9) 38 (69.1, 55.0–80.5) 1.310 (0.569–3.016) 0.672
 Early sepsis response on day 3, n (%, 95% CI) 18 (32.7, 21.1–46.8) 23 (41.8, 28.9–55.9) 0.677 (0.311–1.473) 0.430
 Sepsis response on day 7, n (%, 95% CI) 23 (41.8, 28.9–55.9) 28 (50.9, 37.2–64.5) 0.693 (0.327–1.471) 0.445
 New sepsis episode until day ­28a, n (%, 95% CI) 7 (30.4, 14.1–53.0) 19 (67.9, 47.6–83.4) 0.207 (0.063–0.682) 0.012
 Days to new sepsis episode up to day 28, mean 18 (7) 15 (6) – 0.368
(SD)a n (%, 95% CI)
CI, confidence intervals; OR, odds ratio; and SD, standard deviation
a
Among 51 patients experiencing sepsis response, defined by SOFA decrease of ≥ 25% on day 7

(See figure on next page.)

Fig. 2  Twenty-eight-day survival among trial participants (A) 28-day survival analysis by Kaplan–Meier curves among patients with sepsis and
multiple organ dysfunction syndrome, treated with clarithromycin or placebo. Hazard ratio is provided by Cox-regression analysis. (B) Risk of death
within 28-days in pre-specified subgroups among patients treated with clarithromycin or placebo. P values for interactions between treatment
arm and subgroup are provided by the Breslow–Day test. *Calculated using the Firth correction. ARDS, acute respiratory distress syndrome; CI,
confidence intervals; ICU, intensive care unit; SOFA, Sequential Organ Failure Assessment

explain 30.3% of 28-day mortality variance. No heteroge- 0.21 [95%CI 0.06–0.68]; P = 0.012) (Table  2). In Poisson
neity related to study site was detected (Additional file 1: regression model, the incident rate of new septic epi-
Figure S1). Study enrollment within the first 48  h from sodes in the clarithromycin arm was lower than in the
sepsis onset did not impact final outcome (Additional placebo arm (incident rate ratio [IRR] 0.44 [95% CI 0.19–
file 1: Table S5). In post hoc regression analysis, the study 0.99]; P = 0.048), and in Cox-regression analysis, survival
drug treatment effect remained unaltered by the type of free from new sepsis episode was significantly prolonged,
antimicrobial treatment (Additional file  1: Tables S6 to following clarithromycin treatment (Additional file  1:
S10). Figure S3).
No differences regarding the primary outcome were The type and pathogen of the first recurrent septic epi-
detected in any of the 8 pre-planned subgroup analyses. sode among patients with sepsis response by day 7 are
However, treatment impacted differently patients with shown in Additional file  1: Table  S11 and did not differ
extreme disease severity (within the highest quartile of between the two study arms. The distribution of recur-
SOFA) compared to those with lower SOFA scores, as rent septic sites, compared to the original infections, is
well as patients enrolled in the general ward, compared to presented in Additional file 1: Table S12 and mainly con-
those in the ICU (Fig. 2B). cerned cases of pneumonia.

Secondary outcomes
There were no significant differences between treatment Impact of clarithromycin treatment on the host immune
arms in 90-day mortality (OR 1.31 [95%CI 0.57–3.05]; response
P = 0.672) (Table  2). Subgroup analysis suggested dif- Treatment with clarithromycin was associated with an
ferential treatment response among patients with and increase in non-classical monocytes on day 10 (Addi-
without extremely severe disease, with clarithromycin tional file 1: Figure S4). Counts of classical monocytes, T
favoring better outcomes among patients with SOFA cells, B cells and NK cells remained unaltered. A signifi-
score 12 or more (Additional file 1: Figure S2). cant increase in mHLA-DR expression was also observed
Early sepsis response by day 3 (OR 0.68 [95%CI 0.31– among clarithromycin-treated patients on day 10, com-
1.47]; P = 0.430) and sepsis response by day 7 did not dif- pared to placebo (Additional file 1: Figure S5).
fer (OR 0.69 [95%CI 0.33–1.47]; P = 0.445). Twenty-three Similarly, among those with sepsis response on day 7,
patients in the clarithromycin group and 28 patients in patients in the clarithromycin arm presented higher lev-
the placebo group had sepsis response by day 7; in seven els of HLA-DR on day 10 (Additional file 1: Figure S6).
(30.4%) and 19 (67.9%), respectively, sepsis recurred (OR
 Karakike et al. Critical Care (2022) 26:183 Page 7 of 11

Fig. 2  (See legend on previous page.)

Karakike et al. Critical Care (2022) 26:183 Page 8 of 11

Gene set enrichment analysis identified a cholesterol days after the stop of the study drug. The investigators
homeostasis gene set upregulated in the clarithromy- associated all eight cases either with sepsis progression
cin-treated group, relative to placebo (Additional file  1: or with the administration of other drugs (Additional
Results and Figures S7–9). file 1: Table S13). The two groups of treatment did not
differ in the incidence of non-serious adverse events
Safety (Additional file  1: Table  S14). In nil patient, the study
At least one serious adverse event (SAE) was reported drug was discontinued [20].
for most of the participants. SAEs were equally distrib-
uted among the groups of treatment (Table  3). Eight Discussion
cases of acute kidney injury were reported as SAEs: In this randomized clinical trial, adjunctive clarithro-
seven in the clarithromycin arm and one in the placebo mycin treatment did not affect 28-day survival among
arm. Nil case was reported as associated with the study patients with sepsis and MODS. Clarithromycin treat-
drug. With the exception of one case in each group, the ment was associated with fewer sepsis recurrences
other cases of acute kidney injury presented several among patients with original sepsis response and findings

Table 3  Serious treatment-emergent adverse events (sTEAEs)

sTEAE, n (%) Clarithromycin (n = 55) Placebo (n = 55) P value

At least one sTEAE 50 (90.9) 50 (90.9) 1.000

 Infections and infestations 31 (56.4) 33 (60.0) 0.847
 Acute kidney injury 7 (12.7) 1 (1.8) 0.060
 Disseminated intravascular coagulation 1 (1.8) 0 (0.0) 1.00
 Arterial ischemia 7 (12.7) 5 (9.1) 0.761
 Cardiac disorders 7 (12.7) 5 (9.1) 0.776
  Non-ST elevation myocardial infarct 2 (3.6) 1 (1.8) 0.999
  Ventricular tachycardia 2 (3.6) 2 (3.6) 1.00
  Pulmonary edema 3 (5.5) 2 (3.6) 1.00
 Vascular disorders 8 (14.5) 6 (10.9) 1.00
  Arterial ischemia 7 (12.7) 5 (9.1) 0.761
  Venous thromboembolism 0 (0.0) 1 (1.8) 1.00
  Hemoptysis 1 (1.8) 0 (0.0) 1.00
 Hemorrhagic complications 3 (5.5) 0 (0.0) 0.243
  Surgical site bleeding with hemorrhagic shock 2 (3.6) 0 (0.0) 0.495
 Μetabolic and nutrition disorders 2 (3.6) 0 (0.0) 0.495
  Myxedema coma 1 (1.8) 0 (0.0) 1.00
  Hypoglycemia 1 (1.8) 0 (0.0) 1.00
 Thoracic, pulmonary or mediastinal disorders 3 (5.5) 2 (3.6) 1.00
  Airway obstruction (tracheotomy plugs) 3 (5.5) 2 (3.6) 1.00
  Pneumothorax 1 (1.8) 1 (1.8) 1.00
 Neurological disorders 1 (1.8) 3 (5.5) 0.618
  Brain edema 0 (0.0) 1 (1.8) 1.00
  Hypoxic encephalopathy 0 (0.0) 1 (1.8) 1.00
  Seizures 1 (1.8) 1 (1.8) 1.00
 Surgery complications 4 (7.3) 1 (1.8) 0.363
  Retinal fissure detachment 0 (0.0) 1 (1.8) 1.00
  Nephrostomy 1 (1.8) 0 (0.0) 1.00
  Intra-abdominal retention of surgical compress 1 (1.8) 0 (0.0) 1.00
  Ileal perforation 1 (1.8) 0 (0.0) 1.00
  Surgical wound dehiscence 1 (1.8) 0 (0.0) 1.00
  Mediastinal abscess drainage 1 (1.8) 0 (0.0) 1.00
Percentages may not add up to 100% since some patients have experienced more than one serious adverse event
 Karakike et al. Critical Care (2022) 26:183 Page 9 of 11

compatible with modulation of the immune response outcome used for power calculation was ambitious, the
toward return to homeostasis. However, due to the lim- authors expected that the prognostic enrichment in
ited study size secondary outcomes were not adjusted for MODS patients (where clarithromycin had shown maxi-
multiple comparisons and subgroup analyses showing mum benefit) would translate into the same treatment
treatment benefit need to be interpreted with caution. effect, shown in both previous RCTs [17, 18]; (c) the high
Overall study mortality was high. This was anticipated infection rate, exceeding 50%, by extremely or pan-drug-
since the baseline SOFA score was high (median 10 in the resistant isolates, particularly Acinetobacter bauman-
placebo group; median 11 in the clarithromycin group). nii, making direct comparisons with other European
The finding on sepsis recurrence is in agreement with ICUs complicated. This may have also prevented survival
the secondary benefit of clarithromycin treatment in benefit from being demonstrated, while lower sepsis
patients with CAP. In this RCT, patients were rand- recurrence in this high-resistance setting may be con-
omized to β-lactam monotherapy or to combination of sidered as a relevant clarithromycin treatment outcome;
β-lactams and clarithromycin. Fewer patients receiving and (d) the delay in study enrollment reaching median of
combination therapy were readmitted to hospital by day 4 days after sepsis onset.
30, mostly due to recurrence of pneumonia [21].
Subgroup analysis suggested 90-day survival benefit Conclusions
among extremely severe patients with SOFA score 12 or Among patients with sepsis and MODS, clarithromycin
more. However, such a benefit was missing from patients did not affect all-cause 28-day mortality; sepsis recur-
with SOFA score less than 12 and these patients may rence among patients with sepsis response by day 7 was
be even harmed. However, the limited size of the study decreased. Reprogramming of genes involved in choles-
population makes this post hoc survival benefit incon- terol biosynthesis pathway in circulatory immune cells
clusive. The incidence of acute kidney was greater in the may underlie this process.
clarithromycin arm, although no association with the
study drug was documented.
The above-described clinical benefit in the clarithromy- Abbreviations
APACHE: Acute Pathophysiology and Chronic Health Evaluation; ARDS: Acute
cin group may be linked with the modulation of the host respiratory distress syndrome; CAP: Community-acquired pneumonia; CCI:
immune response toward recovery from sepsis-induced Charlson comorbidity index; HAP: Hospital-acquired pneumonia; HCAP:
immunosuppression. This is reflected by the increase in Healthcare-associated pneumonia; ICU: Intensive care unit; IRR: Incident rate
ratio; mHLA-DR: Monocyte human leukocyte antigen-DR; MODS: Multiple
mHLA-DR expression and expansion of non-classical organ dysfunction syndrome; OR: Odds ratio; RCT​: Randomized controlled
monocytes. Low mHLA-DR expression is a hallmark of trial; RNA: Ribonucleic acid; SOFA: Sequential Organ Failure Assessment; VAP:
sepsis-induced immune suppression, associated with Ventilator-associated pneumonia.
secondary infections and mortality, while recovery is a
surrogate of improved outcomes [22] and has been used Supplementary Information
to monitor immunotherapy efficacy [23, 24]. This is in The online version contains supplementary material available at https://​doi.​
org/​10.​1186/​s13054-​022-​04055-4.
line with previous observations from our group in VAP;
clarithromycin treatment was associated with better Additional file 1. Contains Supplementary Methods and Results.
ex vivo function of PBMCs, decreased IL-10/ TNFα ratio, Additional file 2. Contains the complete Study Protocol versions 1 and 2,
improved antigen presentation and greater apoptosis of a detailed description of amendments between protocol versions and the
monocytes, consistent with immune restoration [20]. Statistical Analysis Plan.
Although non-referring to the entire study popula-
tion, transcriptomic data indicated that genes involved in Acknowledgements
cholesterol biosynthesis were altered in clarithromycin- The authors acknowledge Miltiades Kyprianou (Hellenic Institute for the Study
of Sepsis), who provided statistical advice and support throughout the clinical
treated patients, relative to placebo, at study day 5. Other trial. They also wish to thank the patients, their families and the clinical, labora‑
data also indicate the presence of an axis between cho- tory and research staff of each site, who contributed to the trial.
lesterol synthesis, antigen presentation and monocyte
Notation of abstract publication/ presentation
reprogramming toward a trained immunity phenotype Some of the results were presented as an abstract at the 32nd European
[25–27]. Whether this axis is modulated by clarithromy- Congress of Clinical Microbiology and Infectious Diseases (ECCMID-Abstract
cin remains to be demonstrated. 03038); Lisbon, Portugal; April 23–26, 2022.
Our study had four main limitations: (a) the limited Author contributions
sample size to detect secondary outcomes with sufficient EJGB conceptualized the study design, participated in data analysis and draft‑
power and to account for multiple testing and differences ing of the manuscript, had full access to all of the study data and takes respon‑
sibility for the integrity of the data and the accuracy of the data analysis. EK
among subgroups; (b) the probability for type II error. participated in data analysis and drafted the manuscript, had full access to all
Although the anticipated 25% difference in the primary of the study data and takes responsibility for the integrity of the data and the
Karakike et al. Critical Care (2022) 26:183 Page 10 of 11

accuracy of the data analysis. EJGB, EK, MR and IM verified the underlying data. institution from MDS, GSK and Swedish Orphan Biovitrum AB. A. Antonia‑
BPS analyzed the RNA-sequencing data, participated in data analysis and criti‑ dou has received honoraria from Gilead, Pfizer, MSD, ViiV, BMS, Astellas and
cally revised the manuscript for important intellectual content. AK performed independent educational grants from Gilead, GSK and Biotest. D. Grimaldi
laboratory measurements and critically revised the manuscript for important received consultation fees from Transgene SA Illkirch-Graffenstaden (France).
intellectual content. MR, IM, NK, KP, EM, AP, PC, EP, ITS, FF, ML, ND, DDB, IF, SA, E. J. Giamarellos-Bourboulis has received honoraria from Abbott CH, bioMé‑
EA, MP, GV, EM, AA, DG and IT enrolled patients, collected clinical data and rieux, GSK, InflaRx GmbH, ThermoFisher Brahms GmbH, Sobi and XBiotech Inc;
critically revised the manuscript for important intellectual content. ITS and independent educational grants from Abbott CH, AxisShield, bioMérieux Inc,
TR analyzed the flow cytometry data and critically revised the manuscript for InflaRx GmbH, Johnson & Johnson, MSD, Sobi and XBiotech Inc.; and funding
important intellectual content. WJW contributed in data interpretation and from the Horizon2020 Marie Skłodowska-Curie International Training Network
critically revised the manuscript for important intellectual content. All authors “the European Sepsis Academy” (granted to the National and Kapodistrian
approved the final version of the manuscript to be submitted for publication. University of Athens); the Horizon 2020 European Grants ImmunoSep and
RISC in COVID (granted to the Hellenic Institute for the Study of Sepsis); and
Funding the Horizon Health grant EPIC-CROWN-2 (granted to the Hellenic Institute for
The study was supported by the Hellenic Institute for the Study of Sepsis and the Study of Sepsis). The other authors do not report any conflict of interest.
by the Horizon2020 Marie Skłodowska-Curie International Training Network
“the European Sepsis Academy” (grant number 676129—granted to the Author details
1
National and Kapodistrian University of Athens). The funders had no role in  4th Department of Internal Medicine, National and Kapodistrian University
the design and conduct of the study; collection, management, analysis and of Athens, 1 Rimini Street, 124 62 Athens, Greece. 2 Division of Infectious
interpretation of the data; preparation, review or approval of the manuscript; Diseases, Center for Experimental Molecular Medicine, Amsterdam University
and decision to submit the manuscript for publication. Medical Centers, Academic Medical Center, University of Amsterdam, Amster‑
dam, The Netherlands. 3 Department of Applied Biomedical Science, Faculty
Availability of data and materials of Health Sciences, Mater Dei Hospital, University of Malta, Msida, Malta. 4 Cen‑
The sequence libraries generated in this study are publicly available through tre for Molecular Medicine and Biobanking, University of Malta, Msida, Malta.
5
the National Center for Biotechnology Information (NCBI) gene expression  Intensive Care Unit, Korgialeneio Benakeio General Hospital, Athens, Greece.
6
omnibus (GEO) under the accession number GSE196117. Requests for de-  Intensive Care Unit, Aghios Dimitrios General Hospital, Thessaloniki, Greece.
7
identified and protected health information (PHI)-stripped patient data can  Intensive Care Unit, Hippokration General Hospital, Athens, Greece. 8 Intensive
be made to the corresponding author with specific data needs, analysis and Care Unit, G. Gennimatas General Hospital, Thessaloniki, Greece. 9 Intensive
dissemination plans. Previous Institutional Review Board (IRB)/Independent Care Unit, Theageneion General Hospital, Thessaloniki, Greece. 10 Intensive
Ethics Committee (IEC) approval will be required, if applicable. Dates will be Care Unit, Laiko General Hospital, Athens, Greece. 11 Infectious Diseases
time-shifted to eliminate PHI, as needed. Requests for supporting documents, Service, Department of Medicine, Lausanne University Hospital and University
such as statistical analysis plan, complete trial protocol and subsequent of Lausanne, Lausanne, Switzerland. 12 2nd Department of Critical Care Medi‑
protocol amendments, can also be addressed to the corresponding author cine, National and Kapodistrian University of Athens, Athens, Greece. 13 2nd
with specific needs, analysis and dissemination plans. All the above requests Department of Internal Medicine, Sismanogleion General Hospital, Athens,
will be reviewed by the study sponsor for release, upon publication. Contact: Greece. 14 Department of Infectious Diseases, Centre Hospitalier Universitaire
[email protected]. Saint‑Pierre, Université Libre de Bruxelles (ULB), Brussels, Belgium. 15 Institute
for Medical Immunology, Université Libre de Bruxelles (ULB), Brussels, Belgium.
16
 Department of Intensive Care, Centre Hospitalier Universitaire Brugmann,
Declarations Brussels, Belgium. 17 Department of Intensive Care, CUB‑Erasme, Université
Libre de Bruxelles (ULB), Brussels, Belgium.
Ethics approval and consent to participate
The study was conducted in accordance with the declaration of Helsinki and Received: 30 March 2022 Accepted: 8 June 2022
national and institutional standards. The protocol and informed consent form
were approved in Greece [National Organization for Medicines (51239/01-06-
2017), National Ethics Committee (52086/2017)] and Belgium [Federal Agency
of Medicines and Health Products (1078386/16-04-2018), Central Ethics Com‑
mittee, Erasme University Hospital (P2018/376, 19-10-2018)]. Study registration
References
was with EudraCT (2017-001056-55) and Clinicaltrials.gov (NCT03345992).
1. Fleischmann-Struzek C, Mellhammar L, Rose N, Cassini A, Rudd KE,
Written informed consent was provided by patients or legal representatives,
Schlattmann P, et al. Incidence and mortality of hospital- and ICU-treated
prior to inclusion.
sepsis: results from an updated and expanded systematic review and
meta-analysis. Intensive Care Med. 2020;46:1552–62.
Consent for publication
2. Rudd KE, Johnson SC, Agesa KM, Shackelford KA, Tsoi D, Kievlan DR, et al.
Not applicable.
Global, regional, and national sepsis incidence and mortality, 1990–2017:
analysis for the global burden of disease study. Lancet. 2020;395:200–11.
Competing interests
3. Prescott HC, Angus DC. Enhancing recovery from sepsis: a review. JAMA.
E. Karakike is supported by the Horizon2020 Marie Skłodowska-Curie Inter‑
2018;319:62–75.
national Training Network “the European Sepsis Academy” (grant number
4. Mouncey PR, Osborn TM, Power GS, Harrison DA, Sadique MZ, Grieve RD,
676129- granted to the National and Kapodistrian University of Athens). I.T.
et al. Trial of early, goal-directed resuscitation for septic shock. N Engl J
Schrijver is supported by the Horizon2020 Marie Skłodowska-Curie Inter‑
Med. 2015;372:1301–11.
national Training Network “the European Sepsis Academy” (grant number
5. WHO Rapid Evidence Appraisal for COVID-19 Therapies (REACT) Working
676129- granted to Lausanne University Hospital) and received a scholarship
Group, Sterne JAC, Murthy S, Diaz JV, Slutsky AS, Villar J, Angus DC, et al.
from the Société Académique Vaudoise (Lausanne, Switzerland). N. Dauby
Association between administration of systemic corticosteroids and mor‑
is a post-doctorate clinical master specialist of the Belgian F.R.S-FNRS and
tality among critically ill patients with COVID-19: a meta-analysis. JAMA.
reports personal fees from Roche and Boehringer Ingelheim, and non-
2020;324:1330–1341.
financial support from Pfizer, Janssen and Merck Sharp & Dohme, all outside
6. Annane D, Bellissant E, Bollaert PE, Briegel J, Keh D, Kupfer Y, et al. Corti‑
the submitted work. T. Roger is funded by the European Union Horizon
costeroids for treating sepsis in children and adults. Cochrane Database
2020 Marie Skłodowska-Curie Action Innovative Training Network European
Syst Rev. 2019;12:CD002243.
Sepsis Academy (ESA-ITN, grant number 676129) and by the European Union
7. Hotchkiss RS, Colston E, Yende S, Angus DC, Moldawer LL, Crouser ED,
Horizon 2020 grant ImmunoSep (847422). W. J. Wiersinga is supported by
et al. Immune checkpoint inhibition in sepsis: a phase 1b randomized,
the Netherlands Organization for Scientific Research (VIDI grant 91716475)
placebo-controlled, single ascending dose study of antiprogrammed cell
and the Horizon2020 Marie Skłodowska-Curie International Training Network
death-ligand 1 antibody (BMS-936559). Crit Care Med. 2019;47:632–42.
“the European Sepsis Academy” and received consulting fee paid to the host
 Karakike et al. Critical Care (2022) 26:183 Page 11 of 11

8. Shakoory B, Carcillo JA, Chatham WW, Amdur RL, Zhao H, Dinarello CA, immunometabolic and epigenetic programs in trained immunity. Cell
et al. Interleukin-1 receptor blockade is associated with reduced mortality Metab. 2016;24:807–19.
in sepsis patients with features of macrophage activation syndrome:
reanalysis of a prior phase III trial. Crit Care Med. 2016;44:275–81.
9. Leentjens J, Kox M, Koch RM, Preijers F, Joosten LA, van der Hoeven JG, Publisher’s Note
et al. Reversal of immunoparalysis in humans in vivo: a double-blind, Springer Nature remains neutral with regard to jurisdictional claims in pub‑
placebo-controlled, randomized pilot study. Am J Respir Crit Care Med. lished maps and institutional affiliations.
2012;186:838–45.
10. Vincent JL, Marshall JC, Dellinger RP, Simonson SG, Guntupalli K, Levy
MM, et al. Talactoferrin in severe sepsis: results from the phase ii/iii oral
talactoferrin in severe sepsis trial. Crit Care Med. 2015;43:1832–8.
11. Francois B, Jeannet R, Daix T, Walton AH, Shotwell MS, Unsinger J, et al.
Interleukin-7 restores lymphocytes in septic shock: the IRIS-7 randomized
clinical trial. JCI Insight. 2018;3:e98960.
12. Van der Poll T, Shankar-Hari M, Wiersinga WJ. The immunology of sepsis.
Immunity. 2021;54:2450–64.
13. Reijnders TDY, Saris A, Schultz MJ, van der Poll T. Immunomodulation
by macrolides: therapeutic potential for critical care. Lancet Respir Med.
2020;8:619–30.
14. Herath SC, Normansell R, Maisey S, Poole P. Prophylactic antibiotic
therapy for chronic obstructive pulmonary disease (COPD). Cochrane
Database Syst Rev. 2018;10:CD009764.
15. Vardakas KZ, Trigkidis KK, Apiranthiti KN, Falagas ME. The dilemma of
monotherapy or combination therapy in community-acquired pneumo‑
nia. Eur J Clin Invest. 2017. https://​doi.​org/​10.​1111/​eci.​12845.
16. Metlay JP, Waterer GW, Long AC, Anzueto A, Brozek J, Crothers K, et al.
Diagnosis and treatment of adults with community-acquired pneu‑
monia—an official clinical practice guideline of the american thoracic
society and infectious diseases society of America. Am J Respir Crit Care
Med. 2019;200:e45–67.
17. Giamarellos-Bourboulis EJ, Mylona V, Antonopoulou A, Tsangaris I,
Koutelidakis I, Marioli A, et al. Effect of clarithromycin in patients with
suspected Gram-negative sepsis: results of a randomized controlled trial.
J Antimicrob Chemother. 2014;69:1111–8.
18. Giamarellos-Bourboulis EJ, Pechère JC, Routsi C, Plachouras D, Kollias S,
Raftogiannis M, et al. Effect of clarithromycin in patients with sepsis and
ventilator-associated pneumonia. Clin Infect Dis. 2008;46:1157–64.
19. Tsaganos T, Raftogiannis M, Pratikaki M, Christodoulou S, Kotanidou A,
Papadomichelakis E, et al. Clarithromycin leads to long-term survival and
cost benefit in ventilator-associated pneumonia and sepsis. Antimicrob
Agents Chemother. 2016;60:3640–6.
20. Spyridaki A, Raftogiannis M, Antonopoulou A, Tsaganos T, Routsi C, Bazi‑
aka F, et al. Effect of clarithromycin in inflammatory markers of patients
with ventilator-associated pneumonia and sepsis caused by Gram-nega‑
tive bacteria: results from a randomized clinical study. Antimicrob Agents
Chemother. 2012;56:3819–25.
21. Garin N, Genné D, Carballo S, Chuard C, Eich G, Hugli O, et al. β-Lactam
monotherapy vs β-lactam-macrolide combination treatment in moder‑
ately severe community-acquired pneumonia: a randomized noninferior‑
ity trial. JAMA Intern Med. 2014;174:1894–901.
22. Leijte GP, Rimmelé T, Kox M, Bruse N, Monard C, Gossez M, et al.
Monocytic HLA-DR expression kinetics in septic shock patients with
different pathogens, sites of infection and adverse outcomes. Crit Care.
2020;24:110.
23. Hotchkiss RS, Colston E, Yende S, Crouser ED, Martin GS, Albertson T, et al.
Immune checkpoint inhibition in sepsis: a Phase 1b randomized study to
evaluate the safety, tolerability, pharmacokinetics, and pharmacodynam‑
Ready to submit your research ? Choose BMC and benefit from:
ics of nivolumab. Intensive Care Med. 2019;45:1360–71.
24. Meisel C, Schefold JC, Pschowski R, Baumann T, Hetzger K, Gregor J, et al.
• fast, convenient online submission
Granulocyte-macrophage colony-stimulating factor to reverse sepsis-
associated immunosuppression: a double-blind, randomized, placebo- • thorough peer review by experienced researchers in your field
controlled multicenter trial. Am J Respir Crit Care Med. 2009;180:640–8. • rapid publication on acceptance
25. Brands X, Haak BW, Klarenbeek AM, Butler J, Uhel F, Qin W, et al. An
• support for research data, including large and complex data types
epigenetic and transcriptomic signature of immune tolerance in human
monocytes through multi-omics integration. Genome Med. 2021;13:131. • gold Open Access which fosters wider collaboration and increased citations
26. Barlage S, Gnewuch C, Liebisch G, Wolf Z, Audebert FX, Glück T, et al. • maximum visibility for your research: over 100M website views per year
Changes in HDL-associated apolipoproteins relate to mortality in human
sepsis and correlate to monocyte and platelet activation. Intensive Care At BMC, research is always in progress.
Med. 2009;35:1877–85.
27. Arts RJ, Novakovic B, Ter Horst R, Carvalho A, Bekkering S, Lachman‑ Learn more biomedcentral.com/submissions
das E, et al. Glutaminolysis and fumarate accumulation integrate