Authors: Hongsheng Wu, Biling Liao, Tiansheng Cao, Tengfei Ji, Yumei Luo, Jianbin Huang, Keqiang Ma
Categories: Systematic Review, Complicated appendicitis, Laparoscopic appendectomy, Meta-analysis, Peritoneal drainage
Source: BMC Gastroenterology
Peritoneal drainage (PD) following laparoscopic appendectomy(LA) has long been considered beneficial for appendicitis patients, especially those with complicated appendicitis. However, recent research has raised doubts about the advantages of PD, as it not only fails to reduce postoperative complications but also prolongs the operative duration and hospital stay and incurs higher medical expenses. Given this controversy, we conducted a meta-analysis to determine whether drainage is necessary after LA for complicated appendicitis. This meta-analysis had registered in PROSPERO(ID: CRD42023472382).
This study assessed current evidence regarding the efficacy, safety, and potential benefits of drainage versus no drainage following LA for complicated appendicitis.
We conducted a comprehensive search of PubMed, Springer, and the Cochrane Library using the search terms “appendicitis”, “laparoscopic appendectomy”, and “drain” or “drainage” for studies published between January 1, 2000, and December 31, 2022. We employed the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) criteria for study inclusion and exclusion. Primary outcomes included postoperative intra-abdominal abscess, postoperative intestinal obstruction, postoperative stump leakage, wound infection and postoperative visual analog scale(VAS) score, while secondary outcomes consisted of operative time, postoperative recovery time and total hospitalization duration. Studies with at least two outcomes were considered for meta-synthesis. Depending on I^2^ values, fixed- or random effects models were used for data synthesis. Pooled odds ratios (OR) and weighted mean differences (WMD) were calculated for outcome comparisons between PD and no peritoneal drainage (NPD). Sensitivity analysis and meta-regression were performed to assess and investigate inter-study heterogeneity.
After conducting our literature search and screening, twelve studies were analyzed, comprising 3374 cases. During the comparison of primary outcomes between PD and NPD, the incidence of wound infection and postoperative VAS score were significantly higher in the PD group(P < 0.05). While during the comparison of secondary outcomes, the operative duration, postoperative recovery time and hospitalization duration were significantly longer in the PD group than in the NPD group(P < 0.05).
PD following LA for complicated appendicitis not only increases the incidence wound infection and aggravate patients’ postoperative pain, but also prolongs the operative duration, postoperative recovery time and hospitalization duration. Therefore, routine PD after LA for acute complicated appendicitis is not recommended.
The online version contains supplementary material available at 10.1186/s12876-024-03500-8.
Keywords: Peritoneal drainage, Laparoscopic appendectomy, Complicated appendicitis, Meta-analysis
Acute appendicitis is a prevalent acute abdominal condition necessitating emergency surgical intervention. Laparoscopic appendectomy (LA) has rapidly evolved as the standard of care treatment for acute appendicitis [1, 2]. For decades, placing an intra-abdominal drainage tube following LA has been perceived as beneficial in averting the development of postoperative abdominal abscesses, monitoring postoperative abdominal bleeding, and preventing postoperative appendix stump leakage [3, 4]. The rationale behind peritoneal drainage stemmed from the widespread belief among surgeons that the accumulation of inflammatory exudate within the abdominal cavity escalated the risk of postoperative complications, including intra-abdominal abscesses, appendix stump leakage, wound infections, and intestinal obstructions. However, the past few years have witnessed unprecedented scientific progress, which has translated into improved surgical, and the necessity of peritoneal drainage following LA has come under increasing scrutiny [5, 6]. Consequently, controversies persist regarding the appropriateness and safety of peritoneal drainage after LA. To address this ongoing debate, we extensively reviewed available literature on the advantages and disadvantages of peritoneal drainage by searching multiple databases. We then performed a meta-analysis to assess the effectiveness, safety, and potential benefits of peritoneal drainage (PD) versus no peritoneal drainage (NPD) following LA.
Following a predefined protocol outlining objectives and inclusion and exclusion criteria, we meticulously designed this study. We followed the guidelines established by the Meta-analysis of Observational Studies in Epidemiology (MOOSE). This study had been reported in line with PRISMA [7] and AMSTAR [8] (Assessing the methodological quality of systematic reviews) Guidelines.
We conducted a comprehensive literature search using online databases, including PubMed, Springer Link, and the Cochrane Library. Our search terms included the following MeSH terms, limited to [Title/Abstract]: (((Appendicitis[Title/Abstract]) AND (Laparoscopic Appendectomy)) AND (drainage[Title/Abstract])) OR (Drain[Title/Abstract]). We also collected additional relevant materials, such as retrieved studies, review articles, conference papers, and the latest reports, to broaden the scope of our literature retrieval.
We included randomized controlled trials (RCTs) and retrospective comparative studies (cohort or case-control studies) that compared PD with NPD for postoperative LA across all age groups. Excluded from the analysis were studies involving animal experiments, case reports, letters to the editor, and review articles.
The primary outcome measures for this analysis contrasting PD with NPD after LA involved postoperative complications, including intra-abdominal abscess, wound infection, bowel obstruction, appendix stump leakage and postoperative VAS score. Secondary outcome measures encompassed operative duration(in minutes), postoperative recovery time(in days) and hospitalization stay duration(in days) reflecting both health and economic indicators for patients. Two independent reviewers (BL and TJ) extracted the outcomes from each including study, when controversial issues arose, any disagreement was resolved by the adjudicating senior authors(HW and KM).
The quality of research evidence was assessed using the criteria established by the Oxford Evidence-Based Medicine Center in the UK. We employed the Cochrane systematic evaluation tool to evaluate the quality of RCT literature, while the Newcastle-Ottawa Scale [9] was utilized to assess the quality of retrospective research. Three risk levels-low risk, uncertainty, or high risk-were employed in the assessment of “study methodological quality” [10].
We conducted a meta-analysis using the “meta” package. The “metabin” function was applied to synthesize binary variables, with the odds ratio (OR) serving as the combined statistical measure. We used the “metacont” function for data synthesis for continuous variables, employing the weighted mean difference (WMD) as the combined statistical measure. If I^2^ was less than 50%, the fixed-effects model was adopted to pool OR or WMD and their corresponding 95% confidence intervals (CI). In cases where I^2^ was 50% or higher, the random-effects model was chosen. To preliminarily investigate heterogeneity, we employed the “metainf” function for sensitivity analysis. Subgroup analysis and meta-regression were conducted to explore the sources of heterogeneity. Finally, we employed a funnel plot to assess publication bias. If funnel plot asymmetry was detected, the trim and fill method was applied using the “trimfill” function. All the statistical analyses mentioned were performed using R version 4.1.3.
In the initial search across PubMed, Springer, and the Cochrane Library databases, we retrieved a total of 959 studies. Following the removal of duplicate studies (n = 345), studies unrelated to this meta-analysis (n = 202), and studies for other reasons (n = 187), we screened 225 studies. After eliminating 75 studies that lacked full text or could not be retrieved, 119 studies were assessed. Ultimately, studies were excluded, including reviews, conference materials, editorials, letters, animal experiments, studies with unextractable data and studies involved uncomplicated appendicitis (n = 107). Therefore, the final analysis included twelve studies comprising 3374 patients (1117 PD cases and 2257 NPD cases). A detailed flowchart of the study screening and selection process is presented in Fig. 1.
Fig. 1 Flowchart of study searching strategy base on PRISMA
The included studies encompassed one RCT [11], two prospective studies [12, 13], and nine retrospective researches [6, 5, 14–20]. For primary outcomes, intestinal obstruction was reported in six studies [14, 15, 5, 18–20], intra-abdominal abscess in ten studies [6, 5, 12, 13, 15–20], stump leakage in three studies [12, 13, 17] and eight in wound infection [11, 12, 14, 17, 5, 18–20], respectively. For secondary outcomes, operative duration and hospitalization were reported in four studies [6, 17, 19, 20] and nine studies [6, 5, 11–13, 15–17, 19] respectively. Regarding patient age, five studies centered on appendicitis in children [11, 14, 5, 18, 19] and seven on adults [6, 12, 13, 15–17, 20]. Detailed characteristics of the included studies, encompassing demographic information, study design, selection criteria, matching, and quality scoring, are presented in Table 1.
The pooled results indicated no significant difference in postoperative bowel obstruction (OR 2.61, 95% CI 0.78–8.70, p = 0.1189) (Fig. 2A), postoperative intra-abdominal abscess (OR 1.17, 95% CI 0.70–1.98, p = 0.5451) (Fig. 2B), and postoperative appendix stump leakage (OR 2.71, 95% CI 0.52–14.18, p = 0.2368) (Fig. 2C) between the PD and NPD groups. However, the analysis results of postoperative wound infection showed a significant difference favoring the PD group (OR 2.38, 95% CI 1.74–3.27; p < 0.0001), with moderate heterogeneity (χ²=12.65, df 7, p = 0.08; I²=45%) (Fig. 2D), while the incidence of postoperative VAS score were significantly higher in the PD group, with WMD of 0.63 (95% CI 0.04–1.12; p = 0.0367) and obvious heterogeneity (χ²=36.61, df 3, p < 0.01; I²=92%) (Fig. 2E).
Fig. 2 Primary outcomes following peritoneal drainage versus no peritoneal drainage for laparoscopic appendectomy in term of A bowel obstruction, B intra-abdominal abscess, C appendix stump leakage, **D **wound infection and **E **postoperative VAS score Green squares represent the point estimates of the treatment effect OR, with 95% CI indicated by horizontal bars. Blue diamonds represent the summary estimate from the pooled studies with 95% CI
For secondary outcomes, the pooled data from four studies [6, 17, 19, 20] comprising 1631 patients assessing operative time revealed a significant difference favoring the PD group (WMD 13.17 min, 95% CI 2.93–23.41; p = 0.0117), with significant heterogeneity (χ²=132.61, df 3, p < 0.01; I²=98%) (Fig. 3A). Similarly, six studies [11, 13, 15, 5, 17, 19] with 797 patients included data on postoperative recovery time and nine studies [6, 11, 5, 12, 13, 15–17, 19] with 1366 patients on hospitalization duration were analyzed, the pooled results showed that the PD group was associated with significantly longer postoperative recovery time(WMD 1.08 days, 95% CI 0.82–1.34, p < 0.0001) (Fig. 3B) and hospitalization duration(WMD 1.47 days, 95% CI 0.78–2.15, p < 0.0001) than the NPD group (Fig. 3C). A summary of the meta-analysis comparing PD and NPD is presented in Table 2.
Fig. 3 Secondary outcomes following peritoneal drainage versus no peritoneal drainage for laparoscopic appendectomy in term of **A **operative duration, **B **postoperative recovery time and **C **hospitalization stay duration Green squares represent the point estimates of the treatment effect MD, with 95% CI indicated by horizontal bars. Blue diamonds are the summary estimate from the pooled studies with 95% CI
Given the significant heterogeneity observed in the meta-analysis of operative time and hospitalization duration for LA, sensitivity analyses were conducted to identify the specific studies contributing to the heterogeneity. The results indicated that when any single study was omitted from the sensitivity analysis of operative time (Fig. 4A) and hospitalization duration (Fig. 4B), the p-values of the pooled results remained less than 0.05, suggesting that the selection and inclusion of studies were not responsible for the heterogeneity observed in this meta-analysis.
Fig. 4 Sensitivity analysis of secondary outcomes included (A).operative duration and hospital stay duration Green squares are the point estimates of the omitting treatment effect MD, with 95% CI indicated by horizontal bars. While blue diamond are the heterogeneity from the pooled studies with 95% CI base on random effects model
To explore the heterogeneity observed in the meta-analysis of secondary outcomes, we conducted subgroup analyses. In the subgroup analysis of operative time for LA, we stratified the studies based on the population origin, the demographic characteristics and the surgeon experience. Subgroup analysis of population origin and demographic characteristics showed that both stratified indicators were responsible for the heterogeneity of the operative time (Fig. 5A and B). However, the subgroup analysis of operative time showed obvious heterogeneity (Tau^2^ = 216.61, I^2^ = 82%, p = 0.02) in junior surgeon experience subgroup but less heterogeneity (Tau^2^ = 0, I^2^ = 0, p = 0.87) in senior surgeon experience subgroup (Fig. 5C). Therefore, these subgroup analyses suggested that surgeon experience contributed to the heterogeneity observed in the synthesis of operative time.
Fig. 5 Operative duration subgroup analysis. **A **Population origin subgroup; **B **Demographic characteristics subgroup; **C **Surgeon experience subgroup
During subgroup analysis according to hospitalization duration, both the adult group (Tau^2^ = 1.07, I^2^ = 95%, p < 0.01) and the pediatric group (Tau^2^ = 1.03, I^2^ = 89%, p < 0.01) exhibited significant heterogeneity (Fig. 6B). Furthermore, regrouping based on population factors showed less heterogeneity in the non-Asian subgroup (Tau²=0, I^2^ = 0, p = 1.0). However, the Asian group indicated significant heterogeneity (Tau²=1.67, I^2^ = 90%, p < 0.01), indicating that the population factor was responsible for the heterogeneity observed in hospitalization duration (Fig. 6A). Finally, for appendicitis severity subgroup analysis, both perforated and suppurative appendicitis indicated obvious heterogeneity (Fig. 6C). Thus, the subgroup analyses results exhibited that population origin was responsible for the heterogeneity of hospitalization duration.
Fig. 6 Hospital stay duration subgroup analysis. **A **Population origin subgroup; **B **Demographic characteristics subgroup; **C **Appendicitis severity subgroup
Peritoneal drainage has long been thought to be beneficial for preventing abdominal infections, particularly in cases of infectious diseases or postoperative abdominal inflammatory conditions [21–24]. Indwelling PD catheters have been regarded as a useful method for preventing the accumulation of inflammatory effusions and promoting the formation of fistulas. Consequently, the concept of “no surgical infection, no drain” gained widespread acceptance among surgeons. However, the necessity of drainage following surgery has been under scrutiny. As research on the advantages and disadvantages of abdominal drainage for surgical patients has accumulated, an increasing number of studies have indicated that PD may have adverse effects on patients who undergo surgical procedures. These studies have shown that PD can increase the incidence of surgical site infections, intra-abdominal abscesses, and small bowel obstructions caused by adhesions, prolong patient hospitalization times, and increase expenses [14, 25–28]. Several studies on the evolving role of drainage in minimally invasive surgery beyond appendicitis had been reported recently. Mazzola, Michele et al. [29] evaluated the efficacy of PD in pancreatic surgery, through a systematic review of randomized controlled trials, the results indicated that the use of PD after minimally invasive distal pancreatectomy might not be necessary in all cases. Another research from Seung J.L et al. [30], the authors investigated whether PD was necessary after laparoscopic cholecystectomy, the results showed that routine PD was not recommended after laparoscopic cholecystectomy, and once PD was performed, early removal should be recommended to reduce the risk of postoperative complications. Indeed, the above studies just emphasize the evolutionary process of PD in minimally invasive surgical procedures rather than made comparisons between them, because the main purpose of PD among pancreatic surgery, laparoscopic cholecystectomy, and LA are different. Therefore, the question of whether to employ drainage or not after surgery has remained a topic of ongoing debate.
Currently, there are limited meta-analyses assessing the utility of PD following laparoscopic appendectomy. One relevant study in the Cochrane Database Systematic Review, published in 2021 [31], this study suggested that postoperative PD for appendectomy increased the overall complication rate and prolonged hospital stay, but the quality of evidence was low. Additionally, this study had limitations due to the small number of included studies, preventing the assessment of inconsistency and publication bias, as well as the conduct of subgroup analyses for investigating heterogeneity. Additionally, two other studies were published in 2022. Abu, Abduelraheim et al. [32] indicated that PD for complicated appendicitis not only significantly increased the risk of postoperative complications such as fistula, surgical site infection (SSI), bowel obstruction and ileus, but also prolonged length of hospital stay (LOS). However, due to the lack of available data, they were unable to performed subgroup analysis for investigating the source of heterogeneity in their meta-analysis. Another meta-analysis from Liao, Jiankun et al. [33], the result of this study showed that routine prophylactic PD after appendectomy should not be encourage, unless severe appendicitis indicated intraoperatively. Nevertheless, both LA and open appendectomy were included in this meta-analysis, and due to just a few studies included appendicitis grade in their baselines, so that they failed to evaluate the effect of grade of appendicitis on postoperative complications.
In our meta-analysis, pooled data on primary outcomes revealed that PD after LA was not only associated with a higher incidence of wound infections, but also aggravated postoperative pain of LA patients. One of the key advantages of Non-PD is the reduced risk of wound infection. This is because the placement of a drainage tube may introduce an additional site of potential bacterial contamination, which can increase the risk of infection at the surgical site [34, 35]. In addition, Non-PD may also be associated with lower postoperative VAS pain scores. This is because the placement and subsequent removal of a drainage tube can cause discomfort and pain for patients. Avoiding this additional source of pain can contribute to better postoperative outcomes and patient satisfaction. Our study, in line with other research, found that patients who did not receive peritoneal drainage reported significantly lower VAS pain scores compared to those who did. While analyzing secondary outcomes, we observed that intraoperative PD prolonged both operative time and hospital stay duration. Thus, for LA, placing a peritoneal drainage tube increased the risk of wound infections and prolonged the recovery time for patients with appendicitis. Similarly, in a predictive model of 160 complicated acute appendicitis patients who underwent LA, Martínez-Pérez et al. [36] found that PD was a significant risk factor for postoperative complications and prolonged hospital stays. However, in a recent meta-analysis by Neville, J. J. et al. [37], it was found that PD after LA for appendicitis significantly prolonged hospital stay, but there was no significant difference in the wound infection rate, which contradicted our findings. The discrepancy may be due to the inclusion of both open and laparoscopic appendectomy patients in their analysis, and no subgroup analysis according to wound infections was conducted to investigate the source of heterogeneity.
While the inter-study heterogeneity was not significant for most primary outcomes, it was significant for secondary outcomes. Subgroup analysis indicated that for operative time, surgeon experience contributed to the heterogeneity. There are several potential explanations for why surgeon experience might contribute to differences in operative time. First, senior surgeons may have a more refined surgical technique, allowing them to perform procedures more efficiently and with fewer complications. They may also be better able to anticipate and mitigate potential challenges during surgery, reducing the need for additional interventions or delays. Furthermore, senior surgeons may have a deeper understanding of surgical anatomy and physiology, enabling them to navigate the complexities of individual patient anatomies with greater precision and speed. This can lead to faster and more effective surgical procedures, ultimately resulting in shorter operative times. In the subgroup analysis of hospital stay duration, Asian patients with PD also had significantly longer hospital stays. However, subgroup analysis between adult and pediatric appendicitis patients showed no significant difference. These analyses suggest that population differences may be responsible for the heterogeneity observed in secondary outcomes.
Funnel plots, which represent the publication bias of this meta-analysis, were constructed. Since more than ten included studies were included, we conducted a publication bias analysis for intra-abdominal abscess and hospital stay duration. As shown in Fig. 7A, the funnel plot for intra-abdominal abscess exhibited a symmetrical distribution, suggesting no evidence of publication bias. However, the funnel plot for hospital stay duration was asymmetrical, implying the presence of publication bias. We adjusted the funnel plot using the trim and fill method, and the results indicated that to achieve funnel symmetry, two similar studies by Human M.J [11] should be included in future meta-analyses (Fig. 7B).
Fig. 7 Funnel plots for assessing publication bias. **A **Funnel plot showing symmetry indicative of no evidence of publication bias for postoperative intra-abdominal abscess; **B **Funnel plot showing asymmetry indicative of publication bias for hospital stay duration. After filling one studies base on trim and fill method which showed as the hollow circle dots, the funnel plot is basically symmetrical
This meta-analysis has certain limitations that should be considered. First, most of the included studies were retrospective, with only one small-sample RCT included, and the overall quality of evidence was low, which may be inevitably subject to potential selection bias. Second, significant heterogeneity was observed for operative and hospitalization duration. While sensitivity and subgroup analyses were used to explore the sources of heterogeneity, differences in the seniority and technical proficiency of surgeons, overall hospital conditions, preoperative treatments, the actual severity of appendicitis during surgery, and other factors among the studies may have contributed to the observed heterogeneity. Finally, publication bias was detected in one of the secondary outcomes (hospital stay duration), and despite adjustment using the trim and fill method, the results should be interpreted with caution.
In conclusion, routine PD should not be recommended for patients undergoing LA for complicated appendicitis. Given the limitations of this study, we look forward to more RCTs and multicenter research to further explore the use of PD after LA.
We acknowledge the support from the Construction of Major Subject of the People’s Hospital of Huadu District.
Not commissioned; externally peer reviewed.
H.W and T.C were responsible for the study concept and design of this meta-analysis; B.L and Y.L undertook the data acquisition; T.J and J.H were responsible for data analysis and interpretation; Statistical analysis and figures making were performed by H.W and K.M; Manuscript writing performed by H.W; All authors contributed to the article and approved the final submitted version.
This research was supported by the Construction of the Major Subject of People’s Hospital of Huadu District (YNZDXK202201) and the Internal Medicine Research Fund of Affiliated Huadu Hospital, Southern Medical University (No.2020A01).
Data is provided within supplementary information files.
Ethics of the Approval Committee of the research institution names Ethics Committee of Huadu District People’s Hospital of Guangzhou. This is a meta-analysis and do not involve any human and/ or animal studies, ethical approval was not applicable.
Not applicable.
The authors declare no competing interests.
Hongsheng Wu, Email: crazywu2007@126.com.
Keqiang Ma, Email: mkq9928@hotmail.com.
Data is provided within supplementary information files.