The efficacy and safety of prokinetics in critically ill adults receiving gastric feeding tubes: A systematic review and meta-analysis

Background Intolerance to gastric feeding tubes is common among critically ill adults and may increase morbidity. Administration of prokinetics in the ICU is common. However, the efficacy and safety of prokinetics are unclear in critically ill adults with gastric feeding tubes. We conducted a systematic review to determine the efficacy and safety of prokinetics for improving gastric feeding tube tolerance in critically ill adults. Methods Randomized controlled trials (RCTs) were identified by systematically searching the Medline, Cochrane and Embase databases. Two independent reviewers extracted the relevant data and assessed the quality of the studies. We calculated pooled relative risks (RRs) for dichotomous outcomes and the mean differences (MDs) for continuous outcomes with the corresponding 95% confidence intervals (CIs). We assessed the risk of bias using the Cochrane risk-of-bias tool and used the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) methodology to rate the quality of the evidence. Results Fifteen RCTs met the inclusion criteria. A total of 10 RCTs involving 846 participants were eligible for the quantitative analysis. Most studies (10 of 13, 76.92%) showed that prokinetics had beneficial effects on feeding intolerance in critically ill adults. In critically ill adults receiving gastric feeding, prokinetic agents may reduce the ICU length of stay (MD -2.03, 95% CI -3.96, -0.10; P = 0.04; low certainty) and the hospital length of stay (MD -3.21, 95% CI -5.35, -1.06; P = 0.003; low certainty). However, prokinetics failed to improve the outcomes of reported adverse events and all-cause mortality. Conclusion As a class of drugs, prokinetics may improve tolerance to gastric feeding to some extent in critically ill adults. However, the certainty of the evidence suggesting that prokinetics reduce the ICU or hospital length of stay is low. Prokinetics did not significantly decrease the risks of reported adverse events or all-cause mortality among critically ill adults.


Introduction
Critical illness is usually associated with catabolic stress and increases the incidence of infection and multiple organ dysfunction, resulting in a high mortality rate. A systematic review found a strikingly high prevalence of malnutrition in intensive care unit (ICU) patients (ranging from 38% to 78%) [1]. Owing to the benefits of nutrition support with regard to reducing disease severity and favorably impacting patient outcomes, early nutrition support therapy, primarily by the enteral route, is seen as a proactive therapeutic strategy [2]. In addition, if oral intake is not possible, tube feeding through gastric access has been recommended as the standard approach to initiating enteral nutrition in critically ill adult patients [3].
However, enteral tube feeding intolerance is common in critically ill patients, especially those receiving gastric feeding [3,4]. Blaser et al. reported that the pooled proportion of feeding intolerance was 38.3% (95% confidence interval (CI) 30.7-46.2%) [4]; besides, a meta-analysis by the European Society for Clinical Nutrition and Metabolism showed that gastric feeding intolerance was more prevalent than postpyloric feeding intolerance (25.7% vs. 3.5%, p = 0.0005) [3]. In addition, feeding intolerance is associated with elevated mortality, and seven-day feeding intolerance is an independent predictor of 60-day mortality [5]. Given the risk associated with gastric feeding intolerance, it should be treated aggressively.
The administration of prokinetics is the method most commonly used to treat gastric feeding intolerance. Among recipients of gastric feeding, 13% had been prescribed prokinetics preemptively before they developed intolerance. Approximately one-third of patients who developed feeding intolerance were treated with a prokinetic agent during their stay in the ICU [6]. Although the use of prokinetics in the ICU is common, guidelines or recommendations are little agreement on how to use prokinetics for gastric feeding intolerance in critically ill patients [2,3,7,8]. One of the reasons for the different recommendations may be that the definition of feeding intolerance has changed over time, especially regarding a high gastric residual volume (GRV). A GRV of 500 mL is the recommended threshold for a diagnosis of enteral feeding intolerance in US and European critical care and nutrition society guidelines [2,3,9]. Although the updated European Society for Clinical Nutrition and Metabolism (ESPEN) guidelines [3], published in 2019, provide the latest information on enteral nutrition (EN) and parenteral nutrition (PN) in critically ill adult patients, we find that some aspects of the efficacy and safety of prokinetics in critically ill patients are still unclear [10], and it is necessary to find new evidence to address these uncertainties.
On this topic, a previous meta-analysis by Lewis, K. et al. [11] examined the effects of prokinetics on feeding intolerance or high GRV and clinical outcomes. However, Lewis, K. et al. [11] defined feeding intolerance as GRV �150 mL, vomiting, or abdominal distention resulting in feeding interruption. This definition may be considered obsolete [12]. Some new evidence has emerged on this topic; therefore, we conducted this systematic review to determine the efficacy and safety of prokinetics for the treatment of gastric feeding intolerance in critically ill adult patients.

Methods
This systematic review and meta-analysis was conducted according to the Cochrane Handbook for Systematic Reviews of Interventions (version 5.1.0) [13], and the reporting of our study was based on the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement [14]. The review protocol is available on PROSPERO, registration number CRD42020157446.
Neither patients who received gastric feeding in the ICU nor their families were involved in defining the research question or the outcome measures, but they were involved in the design, providing our team with substantial useful advice regarding design ideas.

Search strategy
We searched the Medline and Embase databases as well as the Cochrane Central Register of Controlled Trials (CENTRAL) from their inception dates to November 22, 2019. We combined Medical Subject Headings (MeSH) and free text terms to identify relevant articles. An informatics expert developed our search strategies.
We also searched clinicaltrials.gov (https://clinicaltrials.gov/) and the WHO International Clinical Trials Registry Platform (ICTRP) (http://apps.who.int/trialsearch/) for additional information, using the terms "critically ill patients", and limited our search to studies labeled "completed" AND "interventional studies (clinical trials)" in which summary results were available to identify additional eligible studies. There were no language restrictions. Additionally, we used a manual search strategy to retrieve the relevant articles cited by the retrieved publications (the search strategies are reported in S1 Table).

Inclusion criteria
Trials were selected based on the following inclusion criteria: (1) the study was designed as a randomized controlled trial (RCT) comparing prokinetic treatment with a control group; (2) the population included critically ill adult patients aged �18 years who were admitted to the ICU and received gastric feeding tubes regardless of whether they had pre-existing feeding intolerance; (3) the intervention group received metoclopramide, erythromycin, or other prokinetic agents, such as herbal medicines or natural medicines intended to enhance gastric motility, regardless of the dose, frequency, duration or combination of prokinetics; (4) the control group received no intervention or a placebo; (5) if the gastric feeding patients with feeding intolerance had a GRV �500 mL and/or symptoms of nausea, vomiting, abdominal distention, regurgitation, deterioration in hemodynamics or other symptoms resulting in feeding interruption and failed to respond to interventions, regardless of whether they were in the control group or the prokinetics group, they were switched to postpyloric feeding or had gastric feeding withheld for 4-6 h [2,3]; and (6) the outcomes included any of the following: allcause mortality; Acute Physiology and Chronic Health Evaluation II (APACHE II) or Simplified Acute Physiology Score II; sepsis; use of an artificial airway; pneumonia; hospital or ICU length of stay; patient nutritional status (malnutrition); gastrointestinal symptoms; GRV; feeding intolerance; or side effects of the prokinetics, such as cardiovascular disorders, bronchospasm, extrapyramidal symptoms, abdominal cramps, allergic reactions and pancreas disorders. The exclusion criteria were as follows: (1) the studies had no control group; (2) the studies had no prokinetic treatment group; (3) patients were considered to have feeding intolerance if tube feeding was electively not prescribed or was stopped/interrupted for procedural reasons; (4) the studies discontinued or interrupted gastric feeding prematurely when the GRV was less than 500 mL or the patients did not have any signs of intolerance, such as nausea, vomiting, abdominal pain, abdominal distension, or deterioration in hemodynamics or overall status. For our purposes, gastric feeding intolerance was defined as a "large" GRV (�500 mL), the presence of gastrointestinal symptoms (vomiting, diarrhea, gastrointestinal bleeding, the presence of enterocutaneous fistulas), or inadequate delivery of EN (the energy provided by EN was less than 20 kcal/kg BW/day after 72 h of feeding attempts or less than 60% of the EN target on the fifth day) in critically ill adults receiving nutrition via gastric feeding tubes. Preventive usage of prokinetics meant that prokinetics were prescribed preemptively on the day EN was initiated and before patients presented a GRV >150 mL or symptoms of feeding intolerance. Preventive usage of prokinetics for risk meant that prokinetics were used in patients with GRVs between 150 and 500 mL before the development of intolerance. Therapeutic usage of prokinetics meant that the prokinetics were administered to patients who had developed feeding intolerance.
A reported adverse event was defined as any untoward medical occurrence or unfavorable and unintended sign, including an abnormal laboratory finding, symptom, or disease (new or exacerbated), temporally associated with the use of the study medication. The reported adverse events included abnormal laboratory test results (hematology, clinical chemistry, or urinalysis) or other safety assessments (e.g., ECGs, radiological scans, or measurements of vital signs), including those that worsened from baseline and were deemed clinically significant in the medical and scientific judgment of the investigator; exacerbation of a chronic or intermittent preexisting condition, including an increase in the frequency and/or intensity of the condition; new conditions detected or diagnosed after the administration of study medication even if they may have been present prior to the start of the study; and/or signs, symptoms, or clinical sequelae of a suspected interaction, such as diarrhea, nosocomial pneumonia, severe sepsis, brain herniation, cardiac arrest, or changes in the electrocardiographic QTc interval.

Risk-of-bias assessments
The methodological quality of the included RCTs was assessed independently by 2 researchers (RP, HLL) based on the Cochrane risk-of-bias criteria [13]. The seven items used to evaluate bias in each trial included randomization sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other bias. We defined other bias as being present in the trials in which the baseline characteristics were not similar between different intervention groups. The included trials were graded as low, unclear, or high risk of bias based on the following criteria: (1) trials were considered high risk of bias if either randomization or allocation concealment was assessed as having a high risk of bias, regardless of the risk of other items; (2) trials were considered low risk of bias when both randomization and allocation concealment were assessed as having a low risk of bias and all other items were assessed as having a low or unclear risk of bias; (3) trials were considered to have unclear risk of bias if they did not meet the criteria for high or low risk of bias.

Data extraction
Two researchers (RP, HLL) independently extracted the following information from each eligible RCT: (1) general study characteristics: author name, year of publication, numbers of treatment groups and patients, trial registry number, methods for measuring gastric emptying or GRV, and the definition of feeding intolerance; (2) patient characteristics: sex, age, baseline patient information (presence or absence of pre-existing feeding intolerance, APACHE II score and nutritional status, if reported); (3) primary diseases (the medical, surgical, or neurosurgical conditions of the critically ill patients); (4) interventions: details of the prokinetic treatment group and control group (e.g., dose, frequency, duration and combination of prokinetics for treatment); and (5) outcomes: gastrointestinal symptoms, feeding tolerance, the number of participants with all-cause death, the ICU length of stay, the hospital length of stay, and the number of reported adverse events.
If the trials had more than 2 groups or used factorial designs and could be analyzed using multiple comparisons, we extracted only the information and data of interest reported in the original articles. If a trial had multiple reports, we collated all data into one study. If a trial had reports in both ClinicalTrials.gov and journal publications, we carefully checked data from these two sources for consistency. If outcome data were reported at multiple follow-up points, we used data from the longest follow-up.

Statistical analysis
The effect of prokinetics on gastrointestinal symptoms and feeding tolerance, main clinical outcomes of all-cause mortality, ICU length of stay, hospital length of stay, and reported adverse events were analyzed. We recorded data on the number of participants with each outcome event by allocated group and recorded the number of participants with compliance and the participant, who was later thought to be eligible or otherwise excluded from treatment or follow-up. Intention-to-treat (ITT) analysis was conducted. ITT analysis is a comparison of the treatment groups that include all patients as originally allocated after randomization regardless of whether treatment was initiated or completed [15]. The CONSORT (Consolidated Standards of Reporting Trials) recommends ITT analysis as standard practice [16].
We performed a meta-analysis to calculate the relative risks (RRs) or absolute risk differences (ARDs) for the dichotomous data and mean differences (MDs) for the continuous data, 95% CIs using the Mantel-Haenszel method and the inverse variance statistical method, respectively. If sufficient data were not available in the published reports or conference abstract, we contacted the authors of the paper. If the raw data were not the mean and standard deviation, the sample mean and standard deviation were estimated from the sample size, median, range and/or interquartile range [17,18].
We tested for heterogeneity between trials using a standard Chi 2 test, and statistical heterogeneity between summary data was evaluated using the I 2 statistic. Sensitivity analysis was performed by excluding low-quality studies, trials recruiting participants with particular conditions, or trials with different characteristics from the others. When an inconsistency was detected between the RR and ARD for the same outcome, we explained the results based on the RR because the RR model is more consistent than the ARD, particularly for an intervention aimed at preventing an undesirable event [13,19].
In our meta-analysis, a random-effects model was used. The defining feature of the random-effects model is that there is a distribution of true effect sizes, and there are two sources of variance, within-study error variance and between-study variance [20]. However, if the number of studies is very small, the statistical power will have poor precision due to the variance between studies. Although the random-effects model is still the appropriate model, the information to apply it correctly is not available. In this case, we will add the separate effects. If heterogeneity was identified (I 2 >40% [13]) and sufficient trials were included in the review, we investigated heterogeneity in the specified subgroups based on types of prokinetics (erythromycin, metoclopramide or other prokinetics), the use of a combination of prokinetics (yes or no), and feeding intolerance history (participants with or without pre-existing feeding intolerance before the start of the trial). Analysis was performed to assess whether the difference between the subgroups was statistically significant.
We assessed publication bias by examining funnel plots when the number of trials reporting the primary outcomes was 10 or more. However, if the number of included studies was less than 10 for a given main outcome, the funnel plot may not reliably detect evidence of departure bias. A prototypical situation that should elicit the suspicion of publication bias is when evidence is derived from a small number of studies or small sample sizes and all outcomes favor the intervention [21]. All meta-analyses were performed using RevMan version 5.3 (Cochrane Collaboration). All tests were 2-tailed, and P <0.05 was considered statistically significant.
We used the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) methodology to rate the certainty of evidence as high, moderate, low, or very low. RCTs begin as high-certainty evidence but can be downgraded because of the risk of bias, imprecision, inconsistency, indirectness, or publication bias. If the limitation on the evidence was considered serious, the evidence was downgraded by one level; if the limitation was considered very serious, the evidence was downgraded by two levels [22].
The 15 eligible studies reported the use of 10 prokinetics, including metoclopramide, erythromycin, cisapride, GSK962040, mosapride and herbal medicines or natural medicines intended to enhance gastric motility (Chenxia Sijunzi decoction, ginger, fenugreek seed powder, gastrolit (Zataria multiflora), rikkunshito), respectively. Based on the outcomes measured, the studies were subdivided into those investigating effects on gastrointestinal symptoms, feeding tolerance studies, and clinical outcome studies that investigated the hospital length of stay, ICU length of stay, reported adverse events, and all-cause mortality. The details of the eligible studies are presented in Table 1.

Publication bias
We checked the funnel plots of the main outcomes for asymmetry; however, we included fewer than 10 RCTs in each main outcome; therefore, that the funnel plots could not be used to reliably detect departure bias.

Main outcomes
Effect on gastrointestinal symptoms and feeding tolerance. Thirteen studies evaluated the effect of prokinetics on gastrointestinal symptoms and/or feeding tolerance in adult critically ill patients receiving gastric feeding [23-27, 29-31, 33-37]. The following main results were assessed: gastric emptying, GRV, diarrhea, constipation, feeding complications and feeding intolerance. Gastric emptying was measured by the drug model of acetaminophen absorption or the 13C-octanoic acid breath test with calculation of the gastric emptying time, gastric emptying coefficient or area under the plasma concentration-time curve. The various outcome definitions, especially for gastric tube tolerance, precluded quantitative synthesis of the data.
As a class of drugs, prokinetic agents appear to have positive effects on gastrointestinal function and feeding tolerance. Ten of the thirteen studies reported positive effects on gastric emptying and/or feeding intolerance in critically ill patients who used of prokinetic agents. However, two studies suggested that metoclopramide had no effect on decreasing gastrointestinal complications in adult neurocritical patients or critical traumatic brain injury patients. One study reported that rikkunshito did not improve the achievement of enteral calorie targets in critically ill adults (Table 2).
Effect on hospital or ICU length of stay. The effect of prokinetics on hospital length of stay was examined by five studies [23,[25][26][27]29]. These five studies, which enrolled a total of 250 patients, showed a significant difference in the hospital length of stay between the prokinetic agent-treated group and the control group (MD -3.21, 95% CI -5.35, -1.06; P = 0.003; I 2 =  28%) (Fig 2). Three studies evaluated the effect of prokinetics on ICU length of stay in the critical care setting [23,25,27]. These three studies, enrolling a total of 186 patients, showed that prokinetic agents appeared to have a positive effect on shortening ICU length of stay (MD -2.03, 95% CI -3.96, -0.10; P = 0.04; I 2 = 0%) (Fig 3). Additionally, the separate effects of different prokinetics on the ICU length of stay and hospital length of stay are presented in S3

1) GRV;
2) The percentage of the target enteral calorie intake achieved at the fifth day; 3) The plasma levels of ghrelin; 4) ICU length of stay; 5) Hospital length of stay; 6) Adverse events; 7) Mortality.
Inadequate enteral nutrition/failure to meet the enteral nutrition target at the fifth day (<60%) Preventive usage NG: nasogastric tube feeding; i.v.: intravenous injection; NA: not applicable; C max : peak paracetamol plasma levels; AUC: the area under the paracetamol concentration curve; t +15 , t +30 , t +60 , t +120 : plasma paracetamol concentrations at 15, 30, 60, and 120 minutes after administration of paracetamol and saline or metoclopramide in patients; SAPS, simplified acute physiology score; GRV, gastric residual volume.
# If the trials had more than 2 groups or factorial designs and permitted multiple comparisons, the subgroup in bold font was extracted in this study. ¶ The study did not provide useful data for meta-analysis.
https://doi.org/10.1371/journal.pone.0245317.t001 Gastric emptying in critically ill, sedated, and mechanically ventilated patients can be significantly improved by adding cisapride to a routine enteral feeding protocol.  Effect on reported adverse events. Seven studies reported events that met the definition of adverse events in 757 critically ill patients [23-25, 27-29, 32]. The meta-analysis showed no significant difference in the risk of reported adverse events between the prokinetic agent group and the control group (RR 1.13, 95% CI 0.92, 1.38; P = 0.25; I 2 = 0%) (S2 Fig). Effect on all-cause mortality. The effect of prokinetic agents on all-cause mortality was examined by six studies in 691 critically ill patients [23,25,28,29,31,32]. There was no significant difference in all-cause mortality between the prokinetic agent group and the control group (RR 0.96, 95% CI 0.81, 1.14; P = 0.64; I 2 = 0%) (S3 Fig).

Subgroup analysis
Although no significant heterogeneity was found, we performed subgroup analysis to determine whether important subgroup differences existed. In the subgroup analysis stratified by  type of prokinetic agents, no significant subgroup differences were detected in the clinical outcomes of hospital length of stay, ICU length of stay, reported adverse events and all-cause mortality (S4-S7 Figs). Furthermore, no study compared the use of a combination of prokinetics to placebo or no treatment. Only one study about the preventive usage of prokinetics for risk patients investigated reported adverse events [24]. The other studies investigated the preventive usage of prokinetics in all patients. The subgroup analysis of the preventive usage of prokinetics in all patients did not show important changes in the pooled effects of the reported adverse events.

Sensitivity analysis
The sensitivity analysis, which was performed by excluding the trials with a high risk of bias [25,28], did not show important changes in the pooled effects of hospital length of stay, ICU length of stay, reported adverse events, and all-cause mortality. (S8-S11 Figs).

Certainty of evidence
The certainty of evidence was moderate for the clinical outcome of all-cause mortality. However, the certainty of evidence was low for the clinical outcomes of ICU length of stay, hospital length of stay and reported adverse events. The details of the risk of bias and quality assessment are outlined in Table 3.

Discussion
In this systematic review, we conducted a comprehensive literature search and used objective study inclusion criteria. Fifteen studies were included in the final analysis. Because of the small sample sizes and a relatively small number of eligible studies, the pooled effects are lacking in accuracy in the quantitative analysis. Most studies (10 of 13, 76.92%) showed that prokinetic agents had beneficial effects on feeding tolerance in critically ill adults. The studies that did not show beneficial effects (3 of 13, 23.08%) investigated special populations of neuro-critical patients and critical traumatic brain injury patients taking metoclopramide and rikkunshito. Furthermore, the use of prokinetic agents in critically ill patients receiving gastric feeding may reduce the ICU or hospital length of stay, but the certainty of evidence was low due to the risk of bias and imprecision. Prokinetics did not significantly reduce the risks of reported adverse events or all-cause mortality.
In this study, we examined the effect of prokinetic agents on gastrointestinal symptoms, feeding tolerance and clinical outcomes. Compared to the control group, the group receiving prokinetics did not have a low risk of mortality; these results were the same as the results of the meta-analysis by Lewis, K. et al. [11], but our methods were different. Lewis, K. et al. [11] defined feeding intolerance as either GRV �150 mL, vomiting, or abdominal distention resulting in feeding interruption. This definition may be considered obsolete [12]. We defined gastric feeding intolerance as either a GRV �500 mL or concomitant symptoms of nausea, vomiting, abdominal distention, regurgitation or other symptoms resulting in feeding interruption in critically ill adult patients with gastric feeding tubes. We excluded studies that  Table 3. GRADE evidence profile of the efficacy and safety of prokinetics in critically ill adult patients receiving gastric feeding tubes. 3. Although the studies included any critically ill patient, we did not downgrade for indirectness. 4. We downgraded the quality of evidence for imprecision by one level because the total population size is less than 400. The 95% confidence interval contained a small benefit that did not meet the clinical decision threshold (min. one day).

5.
We did not downgrade for publication bias, although we could not assess this category reliably due to the small number of eligible studies. Not all included studies showed benefits of the studied intervention.
6. We downgraded the quality of evidence for risk of bias by one level. Most studies had an unclear risk of bias. In addition, one study lacked allocation concealment and blinding. 7. We did not downgrade for inconsistency, I 2 = 28% and Chi 2 = 5.52, P = 0.24.

8.
We downgraded the quality of evidence by one level for imprecision because the population size is less than 400. 9. We downgraded the quality of evidence for risk of bias by one level. Most studies had an unclear risk of bias. In addition, two studies lacked allocation concealment and/or blinding. 10. We did not downgrade for inconsistency, I 2 = 0% and Chi 2 = 1.64, P = 0.95.

11.
We downgraded the quality of evidence for imprecision by one level because the 95% confidence interval around the pooled effect included both no effect and appreciable harm (a relative risk increase greater than 25%).

12.
We downgraded the quality of evidence for risk of bias by one level. Most studies had an unclear risk of bias. In addition, two studies lacked allocation concealment, and two studies lacked blinding.
14. We did not downgrade for imprecision because the 95% confidence interval around the pooled effect did not include both no effect and an appreciable benefit (a relative risk reduction greater than 25%) or appreciable harm (a relative risk increase greater than 25%). https://doi.org/10.1371/journal.pone.0245317.t003 discontinued or interrupted gastric feeding prematurely following the disappearance of gastric feeding intolerance. Our meta-analysis included new studies that used this latest definition [23-27, 30, 31], and we identified 5 studies that investigated the administration of prokinetics, including herbal medicines/natural medicines, to critically ill adult patients with gastric feeding tubes [25-27, 34, 37]. Additionally, we found that prokinetic agents might reduce the ICU or hospital length of stay for critically ill patients receiving gastric feeding. However, the number of studies and the sample size were very small, and the certainty of the evidence was low. Furthermore, no significant difference was found between the prokinetic agent groups and placebo/no treatment groups with regard to the risks of reported adverse events and all-cause mortality. Therefore, we cannot draw a convincing conclusion that the use of prokinetics can improve clinical outcomes in critically ill adults. We recommend that more research should be conducted in this field.
This study has several limitations. First, 21 published original studies or trials registered in the International Clinical Trials Registry Platform (WHO) or clinicaltrials.gov were identified. However, 6 trials, although completed, did not have available results, which might have led to the omission of trials meeting the inclusion criteria, resulting in publication bias. Second, some included trials did not record the baseline status of feeding intolerance for all participants. The subgroup results might have been different if all individuals were evaluated. Third, we were unable to comprehensively evaluate the risk of bias in 12 studies due to a lack of information. Fourth, for each outcome, the total sample size was relatively small, which likely resulted in inadequate power to detect a difference in treatment effect. We recommend that more original studies on this topic be conducted.

Conclusion
As a class of drugs, prokinetic agents may improve gastric feeding tolerance in critically ill adults. However, the certainty of the evidence suggesting that prokinetic agents are effective at reducing the ICU or hospital length of stay is low. There was also no significant reduction in the risk of reported adverse events or all-cause mortality. Additional RCTs are needed to determine the effect of prokinetics on clinical outcomes in critically ill patients in the future.
Supporting information S1