Sepsis is a complication caused by a disorder of host response to infection. Septic shock is a serious disease characterized by circulatory and cellular/metabolic dysfunction which is associated with a higher risk of mortality [
Programmed cell death-1 (PD-1) is expressed on activated T cells, natural killer cells, and B cells [
Although several studies have explored the association between PD-1/PD-L1 expression and survival rate of patients with sepsis, few studies have explored the relationship between PD-1-related blockade and the survival rate of patients with sepsis because of medical ethics guidelines. Thus far, research of the effects of PD-1/PD-L1 blockade on sepsis survival has been limited to animal studies. In addition, although most previous reports focused on the effects of PD-1-related blockade in increasing the survival in animals with sepsis, there was no meta-analysis describing their effects. Thus, we performed a meta-analysis to assess the association between PD-1-related blockade and the survival rate in animals with sepsis. Our study aimed to shed new light on precise effects of the estimated treatment on the human trials, reduce the risk of false results [
To identify relevant studies published up to February 2017, we searched PubMed, EMBASE, and the Cochrane Library. The following combination of terms (“programmed cell death 1 blockade” or “PD-1 blockade” or “anti PD-1” or “programmed cell death-1 ligand blockade” or “anti PD- L1”) AND (“sepsis” or “septic shock” or “pyemia” or “septicemia”) were used.
The inclusion and exclusion criteria were designed based on the Cochrane handbook for systematic review of interventions (Version 5.3). The inclusion criteria were as follows: (1) evaluation of the association between PD-1 related blockade and survival rate in animals with sepsis; (2) independent randomized-controlled studies; (3) availability of full paper or acquiring it from the author; (4) selecting the most recent studies from the same author or institution; (5) manuscripts that were published in English.
We excluded studies if (1) enough data for pooling or additional data by contacting authors twice were not obtained; (2) the manuscripts comprised case reports, reviews, comments, abstracts, and editorials and clinical trials; (3) they were duplicate publications.
(1) Two reviewers (ZQ and QZJ) independently extracted relevant data, including study and animal features and outcomes from the key words, titles, abstracts, and full articles. When necessary, they compared the results, arrived at the same opinions, and solved disagreements by discussion with a third reviewer (LCS).
(2) Data pertaining to the dose and time of drug administration, animal age, country, and year of publication were used, including the number of cases in each group, results of the study, and related outcome, were extracted. If the survival outcomes were presented from the studies on animals at different time points, data for the last time point prior to the end of experiment were extracted
Due to the lack of tools for assessing the quality of randomized controlled trial (RCT) animal experiments, Peters’ [
Study quality was evaluated for each publication using a modified 10-point checklist (Amarasingh et al., 2009) with one point allocated to each reported item: (1) peer reviewed publication, (2) sample size calculation, (3) random allocation to groups, (4) blinded assessment of outcome, (5) compliance with animal welfare regulations, (6) statement of potential conflicts of interest, (7) statement of control of temperature, (8) blinded application of PD-1 related blockade, (9) reported number of animals in whom the xenograft did not grow, and (10) presentation of evidence that acts directly against PD-1 related blockade.
Review Manager 5.3 software was used to analyze data. We pooled data using relative risks (RRs) for continuous outcomes with corresponding 95% confidence intervals (95% CIs) to compare the differences. A fixed-effect model was selected if I2 was ≤ 50%. A random-effects model was selected if I2 > 50%. We performed subgroup analysis to identify the sources of heterogeneity. A sensitivity analysis was performed using Stata version 12.0 to identify influence of an individual study on the pooled RR. Publication bias was assessed using a funnel plot and Egger’s test.
A total of 80 studies were retrieved. Among these studies, 19 were duplicated and 57 of them were found to be unrelated to our study. Therefore, 4 studies [
The flow of article selection.
Characteristics of the included studies are described in Table
Characteristics of the included studies.
Paper/author | Animal/sex | Drug | Country | Administration | Dose | Age | Year | Control |
| ||||||||
Brahmamdam et al. | Mice/male | Anti-PD-1 | USA | 24-48h | 200ug in 200mL saline | 8-10w | 2010 | (1)isotype |
(2)saline | ||||||||
Zhang et al. | Mice/male | Anti-PD-L1 | CHINA | (a) 24h before | 50 | 8-10w- | 2010 | (3)isotype |
(4)saline | ||||||||
(b) 3h after CLP | (5)isotype | |||||||
(6)saline | ||||||||
Chang et al. | Mice/male | (a) Anti-PD-1 | USA | 24-48h | 200ug- | 8-10w | 2015 | (7)isotype |
(b) Anti-PD-L1 | 2015 | (8)isotype | ||||||
Shindo et al. | Mice/male | Anti-PD-L1 | USA | 48h after CLP | - | 6-8w | 2017 | (9)inactive |
The overviews of the quality assessments of the four studies are shown in Table
Quality assessment of included studies.
Study/author | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | score |
---|---|---|---|---|---|---|---|---|---|---|---|
Brahmamdam et al. | + | - | + | - | + | - | - | - | - | - | 3 |
Zhang et al. | + | - | - | - | + | + | - | - | - | - | 3 |
Chang et al. | + | - | - | - | + | - | - | - | - | - | 2 |
Shindo et al. | + | - | - | - | + | + | - | - | - | - | 3 |
All the included studies were used to estimate the effects of PD-1 related blockade on survival of animals with sepsis (Figure
Forest of pooling effects of PD-1 related blockade in mice with sepsis.
A funnel plot (Figure
A funnel plot of effects of PD-1 related blockade in mice with sepsis.
A sensitivity analysis (Figure
A sensitivity analysis of effects of PD-1 related blockade in mice with sepsis.
Subgroup analysis (Table
Subgroup results of PD-1 related blockade on survival in animals with sepsis.
Subgroup | Group | Included | p-value | RR (95%CI) | I2 | Subgroup |
---|---|---|---|---|---|---|
Dose | 200 ug | 4 | 0.94 | 1.97 (1.51-2.56) | 0 | P=0.14 |
50 ug | 4 | 0.43 | 3.26 (1.75-6.08) | 0 | ||
Drug | Anti-PD-1 | 3 | 0.83 | 2.02 (1.43-2.82) | 0 | P=0.56 |
Anti-PD-L1 | 6 | 0.54 | 2.31 (1.70-3.14) | 0 | ||
Administration time | <48h | 8 | 0.73 | 2.23 (1.74-2.81) | 0 | P=0.69 |
>48h | 1 | 0.03 | 1.96 (1.08-3.54) | 0 |
Meta-analysis offers more significant evidence for clinical decision. Thus, this meta-analysis was conducted to assess the effects of PD-1 related blockade on survival in animal models of sepsis. Our major finding was that both anti-PD-1 and anti-PD-L1 have an increasing effect on the survival of mice with sepsis. This finding is based on a comprehensive systematic review, which included studies of over 394 mice. Moreover, because of medical ethics and other therapeutic methods, human studies remain difficult to conduct. Thus our results provide a good basis for further clinical research.
The variability in rats is far lower than that in patients. Furthermore, due to the strictly controlled experimental conditions, the results obtained in animal research are highly homogeneous, and the experimental procedures are highly repeatable. This may be the reason for the lack of heterogeneity in our meta-analysis.
PD-1 and its ligands PD- L1 are expressed not only on activated immune cells but also on several nonimmune tissues. Nishimura et al. [
We could not include studies, such as that of Bergerat [
The preclinical experiment should include enough data to detect a treatment effect if such an effect truly exists. However, studies included in our analysis were of inadequate size. Trials with small sizes often have the risk of overestimation of intervention benefits [
In recent years, increasing studies such as those of Liu et al. [
There were some limitations in our study: (1) our sample size does not have sufficient statistical power; (2) after the initial screening and subsequent application of exclusion criteria, only four articles were included for the meta-analysis, although we searched several databases. This, in turn, might not allow for us to draw reliable conclusions; (3) the methodological quality of the included studies was generally low, which is an inherent limitation. This may result in an overestimation of effect size [
Blockade of PD-1/PD-L1 appears to improve the survival in animals with sepsis. Large randomized clinical trials testing the effect of PD-1/PD-L1 blockade in patients with sepsis are warranted.
All data during the current study are available for consultation and for the public, upon request to the corresponding author.
All the authors declare that there are no actual or potential conflicts of interest, including any financial, personal, or other relationships.