According to the latest World Health Organization report, stroke is the second leading cause of death worldwide [
Only clinical randomized controlled trials (RCTs) and case-controlled trials (CCTs) of probiotics for the treatment of stroke were included. We only included literature published in Chinese or English.
The research subjects were patients with general clinical signs and symptoms of stroke (including ischemic stroke and hemorrhagic stroke) and without other diseases [
We defined the treatment group as patients treated with probiotics combined with EN and life support treatment, and the control group as patients who received EN and life support treatment. There was no significant difference in either EN or life support between the two groups (
First, we extracted the national institutes of health stroke scale (NIHSS) scores, average hospitalization time, average bedrest duration, time to reach the target supply of nutrient solution, hemoglobin (Hb), albumin (ALB), serum total protein (TP), physical and chemical properties of prealbumin (PA), tumor necrosis factor-
Papers with inconsistent document types, documents with inconsistent intervention measures, duplicate documents, and documents without the above 24 outcome indicators were excluded.
The databases searched were as follows: PubMed ( #7 Search: ((((stroke) OR (cerebral infarction)) OR (ischemic stroke)) OR (hemorrhagic stroke)) AND (probiotics) #6 Search: (((stroke) OR (cerebral infarction)) OR (ischemic stroke)) OR (hemorrhagic stroke) #5 Search: probiotics #4 Search: hemorrhagic stroke #3 Search: ischemic stroke #2 Search: cerebral infarction #1 Search: stroke
We entered the papers retrieved from each database into CNKI E-study to eliminate duplicates. According to the Patient, Intervention, Comparison and Outcome (PICO) principle, we read the titles and conducted an initial abstract screening. Then, we read the full texts and decided whether to include them in the study. For the RCT quality evaluation, we referred to the Cochrane risk bias assessment tool, which includes the following 7 evaluation items: random allocation method, allocation plan hiding, participant blinding, analyst blinding, resulting data completeness, selective reporting, and other biases. For the CCT quality evaluation, we referred to the Newcastle-Ottawa Scale, which includes the following 8 evaluation items: case definition adequacy, representativeness of the cases, selection of controls, definition of controls, comparability of cases and controls, and exposure ascertainment, which is the same method used to determine case and control exposure factors and nonresponse rate. Two of the authors (Da-yuan Zhong and Lan Li) completed quality evaluation of the studies. In the case of disagreement, the decision was made by a third author (Yi-hui Deng).
We used RevMan5.3 for data analysis. Odds ratios (OR) and relative risk (RR) served as effect indicators for binary variables, and the weighted mean difference (MD) served as effect indicators for continuous variables. We used
A total of 21 RCT articles and 2 CCT articles meeting the criteria were included in the sample [
Document retrieval flowchart.
Basic information on the included literature.
Study | Disease | Course of treatment | Treatment group | Control group | ||||
---|---|---|---|---|---|---|---|---|
Treatment | Male/female | Age | Treatment | Male/female | Age | |||
Bai et al. [ | IS + HS | 1 w | PLBP + ENS | — | — | ENS | — | — |
Ban et al. [ | IS + HS | 14 w | BLTLBT + ENS | 25/10 | 65.8 ± 10.5 | ENS | 26/9 | 66.5 ± 8.3 |
Chen [ | IS + HS | 14d | BLTLBT + F | — | — | F | — | — |
Chen et al. [ | IS | 14 d | PLBP + S | 23/12 | 70.69 ± 11.68 | S | 19/15 | 71.37 ± 12.56 |
Dong [ | IS + HS | 14 d | PLBP + ENS | — | — | ENS | — | — |
Feng [ | IS | 14 d | PLBP + FD | 27/13 | 58.55 ± 8.67 | FD | 25/15 | 54.78 ± 7.74 |
Gao [ | IS + HS | 2 w | PLBP + ENS | 21/19 | 58.2 ± 2.1 | ENS | 15/25 | 51.1 ± 2.3 |
Geng et al. [ | IS | 15 d | PP + ENS | 33/24 | 65.8 ± 2.7 | ENS | 30/27 | 66.4 ± 22.4 |
He et al. [ | IS | 1 m | PLBP + ENS | 36/24 | 70.97 ± 10.86 | ENS | 16/14 | 69.21 ± 1 2.08 |
Huang and Yuan [ | IS | 21 d | BLTLBT + ENS | 16/19 | 54.98 ± 5.10 | ENS | 15/20 | 55.21 ± 5.12 |
Jin [ | IS | 4 w | BTVEC + ENS | — | — | ENS | — | — |
Jin et al. [ | IS + HS | 7–14 d | BLTLBT + ENS | 13/15 | 62.18 ± 11.12 | ENS | 17/11 | 62.07 ± 10.94 |
Liang et al. [ | HS | 60 d | PLBP + FD | — | 60.19 ± 18.65 | FD | — | 62.13 ± 13.97 |
Li et al. [ | IS + HS | 2 w | BQVT + ENS | 29/11 | — | ENS | 28/12 | — |
Li [ | HS | 21 d | LCC + F | 13/10 | 60.9 ± 8.7 | F | 14/8 | 59.5 ± 8.9 |
Li et al. [ | HS | 2 w | BLTLBT + ENE | 24/19 | 60.90 ± 8.60 | ENE | 27/16 | 61.66 ± 10.64 |
Ma [ | HS | 20 d | BQVT + ENS | 25/22 | 52 ± 6 | ENS | 26/20 | 52 ± 6 |
Pei [ | IS + HS | 4 w | PLBP + ENS | 32/28 | 64 ± 10 | ENS | 35/25 | 62 ± 11 |
Yang [ | IS + HS | 7 w | PLBP + ENS | 15/15 | 59.89 ± 3.46 | ENS | 17/13 | 60.23 ± 4.56 |
Yuan [ | HS | 2 w | BTVEC + HP | 28/12 | 58.4 ± 9.3 | HP | 27/13 | 59.1 ± 8.8 |
Zhang [ | IS | 14 d | PLBP + EP | 30/40 | 64.21 ± 9.27 | EP | 32/38 | 63.49 ± 10.64 |
Zhang et al. [ | HS | 21 d | LCC + F | 13/10 | 60.9 ± 8.7 | F | 14/8 | 59.5 ± 8.9 |
Zhang [ | IS + HS | 8 w | LCBC | — | — | — | — | — |
IS: ischemic stroke; HS: hemorrhagic stroke; m: month; w: week; d: day; PLBP: probiotic live bacteria preparation; BLTLBT: bifidobacterium lactobacillus triple live bacteria tablets; BTVEC: bifidobacterium triple viable enteric-coated capsules; PP: probiotic pellets; BQVT: bifidobacterium quadruple viable tablets; LCC: Livzon Changle capsules; LCBC: live clostridium butyricum capsules; EN: enteral nutrients; ST: life support treatment; F: fresubin; S: supportan; FD: fresubin diabetic; EP: ensure powder; HP: homogenate preparation; ENS: enteral nutrient solution or suspension; ENE: enteral nutrient emulsion.
We evaluated a total of 21 RCTs [
Quality evaluation results of the 21 included RCTs.
Study | Random sequence generation | Allocation hiding | Blind researchers and subjects | Blind evaluation of research results | Integrity of result data | Optional reporting of research results | Other biases |
---|---|---|---|---|---|---|---|
Bai et al.[ | L | U | U | U | H | L | U |
Ban et al. [ | L | U | U | U | L | L | U |
Chen et al. [ | L | U | U | U | H | L | U |
Chen et al. [ | L | U | U | U | L | L | U |
Dong [ | U | U | U | U | H | L | U |
Feng [ | U | U | U | U | L | L | U |
Gao [ | U | U | U | U | L | L | U |
Geng et al. [ | U | U | U | U | L | L | U |
He [ | U | U | U | U | L | L | U |
Huang and Yuan [ | L | U | U | U | L | L | U |
Jin et al. [ | L | U | U | U | L | L | U |
Liang et al. [ | U | U | U | U | H | L | U |
Li et al. [ | L | U | U | U | H | L | U |
Li [ | U | U | U | U | H | L | U |
Li et al. [ | U | U | U | U | L | L | U |
Ma [ | L | U | U | U | L | L | U |
Pei [ | L | U | U | U | L | L | U |
Yuan [ | L | U | U | U | L | L | U |
Zhang et al. [ | L | U | U | U | L | L | U |
Zhang et al. [ | U | U | U | U | L | L | U |
Zhang [ | U | U | U | U | H | L | U |
L: low risk; U: unknown risk; H: high risk.
Quality evaluation results of the 2 included CCTs.
Study | Case definition adequacy | Case representativeness | Control selection | Definition of controls | Comparability of cases and controls | Exposure ascertainment | Uses the same method to determine case and control exposure factors | Nonresponse rate | Total |
---|---|---|---|---|---|---|---|---|---|
Jin [ | ★ | ★ | ★★ | ★ | ★ | 6 | |||
Yang [ | ★ | ★ | ★★ | ★ | ★ | 6 |
The effect on stroke was assessed using the three following indicators: NIHSS score, hospital stay duration, and bed rest duration. The meta-analysis results showed that probiotics combined with EN did not significantly reduce the NIHSS scores (MD = -1.11, 95% confidence interval [CI] (−7.92, 5.70),
Meta-analysis for continuous variables.
Effect index | Detail index | Studies included | Heterogeneity test | Model | MD (95% CI) |
---|---|---|---|---|---|
Effect on stroke | NIHSS score | 2 [ | Random effects model | −1.11 (−7.92, 5.70), | |
Hospital stay duration | 5 [ | Random effects model | −8.94 (−11.39, −6.50), | ||
Bedrest duration | 3 [ | Fixed effects model | −10.34 (−11.30, −9.39), | ||
Blood nutrition indicators | HB | 7 [ | Fixed effects model | 8.36 (6.34, 10.38), | |
ALB | 9 [ | Random effects model | 2.91 (2.45, 3.37), | ||
TP | 6 [ | Random effects model | 4.90 (2.43, 7.38), | ||
PA | 4 [ | Random effects model | 15.50 (9.2, 21.79), | ||
Inflammation indicators | TNF- | 3 [ | Random effects model | −3.22 (−5.61, −0.82), | |
IL-6 | 2 [ | Fixed effects model | −16.40 (−21.97, −10.83), | ||
IL-10 | 2 [ | Fixed effects model | −6.63 (−12.55, −0.70), | ||
hs-CRP | 2 [ | Fixed effects model | −2.82 (−10.10, 4.47), | ||
PCT | 2 [ | Fixed effects model | −0.35 (−2.58, 1.89), |
The four blood nutrition indicators were HB, ALB, TP, and PA. The meta-analysis results showed that probiotics combined with EN increased HB (MD = 8.36, 95% CI (6.34, 10.38),
The five inflammation indicators were TNF-
Adverse events were assessed using 8 indicators: vomiting, esophageal reflux, abdominal distension, stress ulcer, constipation, diarrhea, gastric retention, and gastrointestinal bleeding. The meta-analysis results showed that probiotics combined with EN did not reduce the occurrence of vomiting (RR = 0.83, 95% CI (0.46, 1.51),
Binary variable meta-analysis results.
Effect index | Detail index | Study | Heterogeneity test | Model | RR (95% CI) |
---|---|---|---|---|---|
Adverse reactions | Vomiting | 3 [ | Fixed effects model | 0.83 (0.46, 1.51), | |
Stress ulcer | 2 [ | Fixed effects model | 0.47 (0.22, 1.02), | ||
Esophageal reflux | 8 [ | Fixed effects model | 0.43 (0.25, 0.74), | ||
Bloating | 10 [ | Fixed effects model | 0.39 (0.26, 0.58), | ||
Constipation | 12 [ | Fixed effects model | 0.31 (0.21, 0.45), | ||
Diarrhea | 14 [ | Fixed effects model | 0.22 (0.14, 0.34), | ||
Gastric retention | 4 [ | Fixed effects model | 0.34 (0.19, 0.60), | ||
Gastrointestinal bleeding | 10 [ | Fixed effects model | 0.39 (0.28, 0.54), | ||
Complication rate | Lung infection | 12 [ | Random effects model | 0.44 (0.27, 0.72), | |
Gastrointestinal infection | 4 [ | Fixed effects model | 0.40 (0.23, 0.68), | ||
Urinary tract infection | 6 [ | Fixed effects model | 0.27 (0.15, 0.49), | ||
Poor prognostic indicators | Mortality rate | 4 [ | Fixed effects model | 0.45 (0.22, 0.93), | |
Bacterial imbalance rate | 6 [ | Fixed effects model | 0.32 (0.21, 0.48), |
Complications were assessed using the three following indicators: the incidence of lung infection, gastrointestinal infection, and urinary tract infection. The meta-analysis results showed that probiotics combined with EN reduced lung infection (RR = 0.44, 95% CI (0.27, 0.72),
The poor prognostic indicators included the two following items: fatality rate and the intestinal flora imbalance rate. The meta-analysis results showed that probiotics combined with EN reduced the mortality rate (RR = 0.45, 95% CI (0.22, 0.93),
We used a funnel chart to evaluate publication bias for adverse reactions, including bloating, constipation, diarrhea, and gastrointestinal bleeding. The funnel plots for bloating, constipation, and diarrhea had good symmetry, which suggests that there was a low possibility of publication bias in comparing these indicators. However, the diarrhea funnel chart had poor symmetry, which suggests that there may have been publication bias. The results of Egger’s test were consistent with the results of the funnel chart, as shown in Figure
Funnel plots of publication bias analysis results.
Stroke is an acute cerebrovascular disease that manifests primarily as blood vessel blockage in the brain. Stroke usually manifests as either ischemic or hemorrhagic and mostly occurs in men over 40 years old. The most common clinical treatments for stroke are drug therapy and thrombolytic therapy. Both of these treatments have a high risk of complications and so do not improve stroke prognosis [
However, the relationship between probiotics and stroke is more complex. Huang found that changes in intestinal flora can affect ischemic brain injury symptoms in mice [
Meta-analysis is a statistical method that combines the results of several studies into a quantitative indicator. The combination of data from multiple studies can increase the sample size and improve a test’s reliability. In this meta-analysis, we combined 23 RCTs of probiotics combined with EN in the treatment of stroke. Our findings showed that probiotics combined with EN did not reduce NIHSS scores (
In terms of improving inflammation, we found that probiotics combined with EN reduced hs-CRP and PCT, but this difference was not significant (
The present systematic review has some limitations that should be noted. First, the methodological quality of the RCTs included was generally poor, and only 11 included studies [
In summary, additional rigorous randomized double-blind trials are needed to verify the safety and effectiveness of probiotics combined with EN in stroke treatment. However, this was a comprehensive meta-analysis of all published studies on the use of probiotics combined with EN in treating stroke that meet the standards. We objectively evaluated the clinical efficacy of probiotics combined with EN in treating stroke. Therefore, the results of the study still have significance for clinical guidelines.
Requests for additional data may be granted upon reasonable request by contacting the first author (Da-yuan Zhong,
The authors declare no conflicts of interest pertaining to this work.
Yi-Hui Deng determined the direction of the topic. Ruo-Meng Ma and Lan Li carried out literature search. Da-yuan Zhong and Lan Li made the evaluation of research quality and contributed equally to this study. Da-Yuan Zhong did data curation and carried out drafting of the manuscript. Da-Yuan Zhong and Yi-hui Deng did article revision and did article translation.
This study was funded by the National Natural Science Foundation of China’s General Project (81874416).