Systematic review and meta-analysis of safety of laparoscopic versus open appendicectomy for suspected appendicitis in pregnancy

Background Laparoscopic appendicectomy has gained wide acceptance as an alternative to open appendicectomy during pregnancy. However, data regarding the safety and optimal surgical approach to appendicitis in pregnancy are still controversial. Methods This was a systematic review and meta-analysis of studies comparing laparoscopic and open appendicectomy in pregnancy identified using PubMed and Scopus search engines from January 1990 to July 2011. Two reviewers independently extracted data on fetal loss, preterm delivery, wound infection, duration of operation, hospital stay, Apgar score and birth weight between laparoscopic and open appendicectomy groups. Results Eleven studies with a total of 3415 women (599 in laparoscopic and 2816 in open group) were included in the analysis. Fetal loss was statistically significantly worse in those who underwent laparoscopy compared with open appendicectomy; the pooled relative risk (RR) was 1·91 (95 per cent confidence interval (c.i.) 1·31 to 2·77) without heterogeneity. The pooled RR for preterm labour was 1·44 (0·68 to 3·06), but this risk was not statistically significant. The mean difference in length of hospital stay was − 0·49 (−1·76 to − 0·78) days, but this was not clinically significant. No significant difference was found for wound infection, birth weight, duration of operation or Apgar score. Conclusion The available low-grade evidence suggests that laparoscopic appendicectomy in pregnant women might be associated with a greater risk of fetal loss. Copyright © 2012 British Journal of Surgery Society Ltd. Published by John Wiley & Sons, Ltd.


Introduction
Suspected appendicitis is the most common indication for surgery for non-obstetric conditions during pregnancy, and occurs in approximately one in 635 to one in 500 pregnancies per year 1,2 . Appendicitis occurs more frequently in the second trimester than in the first or third trimester of pregnancy 2 -6 . Abdominal surgery during pregnancy, particularly appendicectomy 7 , may increase the risk of poor pregnancy outcomes 8 . Fetal loss usually occurs in 3-15 per cent of women with complicated appendicectomy during the first trimester. However, the rate may be as high as 20-30 per cent 9 -11 , with a premature delivery rate of 15-45 per cent 12 , and a significantly increased risk of spontaneous abortion, premature labour, and perinatal morbidity and mortality 13 . Miscarriage and infant mortality occur more frequently in women with perforated appendicitis 14 . However, the maternal mortality rate is very low 15,16 as a result of the use of advanced antibiotics, close perioperative monitoring, cooperation between specialties and improvements in perioperative management 17,18 .
Although guidelines for laparoscopic procedures during pregnancy have been established 19 , concern remains over the safety of the procedure, with reports of an increased risk of intra-abdominal abscess, particularly in perforated appendicitis. Assessment for open appendicectomy is related to gestational age as the appendix progressively relocates during pregnancy, typically from McBurney's point upwards from the iliac crest to near the gallbladder. Open appendicectomy is an established and safe operation with acceptable morbidity and low mortality rates.
Previous studies 11,13 -15,17,18 were underpowered to detect any benefit of laparoscopic appendicectomy over the traditional open approach, resulting in conflicting results regarding the efficacy of laparoscopic appendicectomy for appendicitis during pregnancy. Only one randomized clinical trial comparing laparoscopic with open appendicectomy in pregnant women has been performed, with quality of life as the primary outcome 20 . One previous systematic review, including 28 observational studies that documented 637 laparoscopic appendicectomies, suggested that the laparoscopic procedure was associated with a higher rate of fetal loss but a similar or lower rate of preterm delivery compared with open appendicectomy 21 . However, the magnitude of treatment effects was not quantified. A systematic review and meta-analysis was therefore carried out, with the primary aim of estimating and comparing pregnancy outcomes including rates of fetal loss, preterm delivery, Apgar score and low birth weight.

Study selection
Studies published between January 1990 and 11 July 2011 were identified from MEDLINE and Scopus databases using PubMed and Scopus search engines respectively. Search terms used were: pregnancy, pregnant women, laparoscopy, laparoscopic appendectomy, laparoscopic management, open appendectomy, conventional appendectomy, maternal outcome, premature labor pain, preterm labor, abortion, fetal loss, gestational age, fetal outcome, birth weight, Apgar score, surgical outcome, hospital stay, length of stay, hospitalization length, operative time, operation time, duration of operation, infection, wound infection, surgical infection and negative appendectomy. Search strategies are described in Table S1 (supporting  information).

Inclusion criteria
Studies were included in the review if they met the following criteria: studied patients were pregnant women with suspected appendicitis; the intervention and comparator were laparoscopic and open appendicectomy respectively; at least one pregnancy (for example preterm delivery, fetal loss, birth weight, Apgar score) or surgical (duration of operation, length of hospital stay, wound infection, negative appendicectomy) outcome was reported; and the study was published in English. Studies were excluded if a hybrid procedure or single-trocar technique was used rather than the standard laparoscopic appendicectomy.
The reference lists of all relevant studies were also reviewed. If studies were duplicates, the one with the most complete data was included. For studies that reported insufficient data, the corresponding authors were contacted and invited to provide more information. Two attempts were made to contact authors and if no response was obtained the study was excluded from the review.

Outcomes
The primary outcomes of interest were the pregnancy outcomes fetal loss and preterm delivery. Secondary outcomes were: birth weight, Apgar score and surgical outcomes, including duration of operation, length of hospital stay, wound infection and negative appendicectomy (appendicitis not proven pathologically).

Data extraction
Two investigators independently extracted the data from each study using a standard data extraction form. Information extracted included general data (author, year of publication, journal), study characteristics (study design, setting), patient characteristics (age, gestational age at surgery, gravida, body temperature, white blood cell count, number of subjects per group), and outcome as described above. Any disagreement was discussed and resolved by consensus with a third reviewer.

Assessment of risk of bias
The quality of studies was assessed independently by three reviewers on the basis of representativeness of studied subjects, information bias (ascertainment of outcome and interventions) and confounding bias ( Table S2, supporting information). Each item was graded as having a low risk of bias, a high risk of bias, or an unclear risk if there was insufficient information to judge 22 . Any disagreement between the reviewers was discussed and resolved by consensus.

Statistical analysis
All analyses were performed using Stata version 11.1 (StataCorp LP, College Station, Texas, USA) 23 . For dichotomous outcomes (preterm delivery, fetal loss and wound infection), the relative risk (RR) of the outcome between laparoscopic versus open appendicectomy and its 95 per cent confidence interval (c.i.) were estimated for each study. If one cell in the 2 × 2 table contained zero, a continuity correction was performed by adding 0·5 to each cell 24 . Heterogeneity of RRs across studies was assessed using Cochran's Q test and the degree of heterogeneity was estimated using the I 2 statistic. If the heterogeneity was significant or I 2 exceeded 25 per cent, a random-effects model using the DerSimonian and Laird method was applied for pooling ORs; otherwise the inverse variance method was used 25,26 .
For continuous variables (duration of operation, length of hospital stay, Apgar score, birth weight) the unstandardized mean difference in outcomes between groups along with its 95 per cent c.i. was estimated and the values were pooled. Heterogeneity of the mean difference across studies was assessed as described above.
Meta-regression analysis was used to assess the source of heterogeneity by fitting age and gestational age at surgery in the meta-regression model. A funnel plot with or without contour enhancement was used to detect publication bias owing to small study effects. The asymmetry of the funnel plot was assessed by means of Egger's test. The trim-andfill method was used to impute missing studies if there was evidence of asymmetry of the funnel. P < 0·050 was considered statistically significant, except for the heterogeneity test, for which P < 0·100 was used.

Results
The initial literature search identified 88 and 196 studies from MEDLINE and Scopus databases respectively. Sixtyone studies were duplicates, leaving 223 for title or abstract review. After exclusion of 212 ineligible articles, 11 studies remained for analysis ( Fig. 1). Agreement in data extraction between the two reviewers was 93·9 per cent (κ = 0·93, P < 0·001) and 92·2 per cent (κ = 0·0·92, P = 0 < 0·001) for dichotomous and continuous outcomes respectively.  The 11 included studies contained a total of 3415 patients (599 in laparoscopic and 2816 in open group) ( Table 1) 12,27 -36 . Eight studies were comparative prospective cohort studies and three were comparative retrospective medical record reviews. Nine of the 11 studies were from the USA. The mean patient age ranged from 23·4 to 29·5 years. Gestational age at surgery was mostly in the second trimester, except in the study by Upadhyay and colleagues 30 . Four studies reported failure of laparoscopic appendicectomy, and the need to convert to open surgery in between one and three patients in each study 27,29,33,36 . Two of these studies carried out intention-to-treat analysis, Corneille et al. 27 Low risk Low risk* High risk High risk 2 operations in LA group converted to OA Sadot et al. 28 Low risk Low risk* Low risk High risk Kirshtein et al. 29 Low risk Low risk High risk High risk 1 operation in LA group converted to OA McGory et al. 12 Low risk Low risk* Low risk Low risk Applied logistic regression, adjusted for age and race Upadhyay et al. 30 Unclear Low risk* Low risk High risk Carver et al. 31 Low risk Unclear Low risk High risk Lyass et al. 32 Low risk Unclear Low risk High risk Affleck et al. 33 Low risk Low risk* High risk High risk 2 operations in LA group converted to OA Conron et al. 34 Unclear Unclear Low risk High risk Gurbuz and Peetz 35 Unclear Unclear Low risk High risk Curet et al. 36 Unclear  including the patient in the laparoscopic group 27,29 . One study applied a per-protocol analysis 33 and the other excluded patients whose operation was converted 36 . The negative appendicectomy rate ranged from 12 to 24 per cent. Fetal loss was reported in all 11 studies 12,27 -36 and preterm labour in seven 27,28,30 -33,35 . Four studies reported on Apgar scores 28,29,31,34 , but only three had sufficient data to pool 28,31,34 , and four reported birth weight 28,29,31,34 . Table 2. The agreement between two reviewers was 95·5 per cent with a κ statistic of 0·94 (P < 0·001). Among 11 studies, the risk of selection bias from the use of non-representative cases was low in seven and unclear in four studies. The ascertainment of all outcomes was clearly described (except for wound infection) in six studies. Ascertainment of surgical technique was clear in seven studies. Unclear ascertainment Curet et al. 36 Gurbuz and Peetz 35 Affleck et al. 33 Conron et al. 34 Lyass et al. 32 Carver et al. 31 McGory et al. 12 Upadhyay et al. 30 Kirshtein et al. 29 Corneille et al. 27 Sadot et al. 28 Overall (I 2 = 0·0%, P = 0·992) Gurbuz and Peetz 35 Affleck et al. 33 Lyass et al. 32 Carver et al. 31 Upadhyay et al. 30 Corneille et al. 27 Sadot et al. 28 Overall (I 2 = 0·0%, P = 0·995)  in four studies was due to conversion from laparoscopic appendicectomy to an open technique. Confounding bias was likely to be present in ten studies.
Egger's test suggested asymmetry of the funnel (coefficient −0·47, s.e. 0·13, P = 0·005). A contourenhanced funnel plot was therefore created (Fig. 3a); this showed that all studies were in the non-significant area except that by McGory and colleagues 12 in which laparoscopic appendicectomy had a significantly higher risk. Despite this asymmetry, application of a nonparametric trim-and-fill method could not identify any missing study.

Surgical outcomes
Wound infection, duration of operation and hospital stay were also pooled across studies ( Table 4). Pooling wound infection in three studies (n = 127) 28,29,31 yielded a pooled RR of 0·91 (0·12 to 7·18), suggesting little difference in the risk of wound infection between interventions. The duration of operation was longer in the laparoscopy group, by a mean of 5·88 (−1·58 to 13·33) min, but the difference was not significant. The length of hospital stay was significantly shorter in the laparoscopy group by almost half a day (95 per cent c.i. −1·76 to −0·78 days).

Discussion
The results of this systematic review and meta-analysis suggest that laparoscopic appendicectomy in pregnancy results in an almost twofold significantly higher risk of fetal loss compared with open appendicectomy. No significant differences were observed between groups in preterm delivery, birth weight, Apgar score, wound infection after surgery or duration of operation. The higher risk of fetal loss after laparoscopic compared with open appendicectomy needs to be addressed in the era of laparoscopic surgery, and has been discussed in many reports of the relative safety of laparoscopy in pregnancy 10,13,37 . However, this finding was largely dominated by the study of McGory and colleagues 12 , which had largest sample size and greatest power in detection of an association. After exclusion of this study from the pooled analysis, there was no effect of laparoscopic appendicectomy on fetal loss.
The major consideration in laparoscopic appendicectomy in pregnancy is the effect of increased intraabdominal pressure and fetal acidosis during carbon dioxide pneumoperitoneum. Increasing abdominal pressure from the pneumoperitoneum can lead to decreased venous return, especially in women with impaired cardiac output 38 , and result in maternal hypotension and hypoxia 39 . In addition, it has been reported that carbon dioxide is also absorbed across the peritoneum, which leads to fetal acidosis 40 . However, this is in contrast the findings of another study that reported no substantial adverse effect on the fetus when the maximum pneumoperitoneal pressure was as high as 10-12 mmHg and the duration less than 30 min 41 .
Although not statistically significant, the present results suggest that there may be an increased risk of preterm delivery in those undergoing laparoscopic appendicectomy compared with open appendicectomy. It is likely that this analysis did not have sufficient statistical power to detect a significant difference, given that a sample size of 749 would be required in each group to detect a RR of 1·44.
Although the mean operating time was 5·88 min longer in the laparoscopic group, this was not statistically significant. The length of hospital stay was approximately half a day shorter after laparoscopic compared with the open appendicectomy, but this result depends heavily on one outlier study and cannot be considered robust. This requires further investigation for health service use planning, but a shorter hospital stay after a laparoscopic appendicectomy might not be advantageous clinically because of the need to monitor the patient for the adverse events noted above.
This meta-analysis quantified the effects of laparoscopic and open appendicectomy on pregnancy and surgical outcomes. A previous review did not pool data and most included studies were non-comparative, with only one group 21 . The present review included the most relevant pregnancy and surgical outcomes.
One major limitation is that all studies included in the pooled analysis were observational, and summary data published within each article were included in the review. Many other factors (such as patient age, duration of pregnancy, weight gain, complicated appendicitis, surgeon's skill, clinical setting) may affect the outcomes following appendicectomy, and confounding bias cannot be ruled out as the studies were not randomized. To adjust for confounding bias, individual patient data would be required from each study. There were no available data on complicated appendicitis (perforated and gangrenous) and it was not possible to assess whether the effects of laparoscopic appendicectomy on fetal loss were confounded by complicated appendicitis. Agreement between the present results and a meta-analysis of randomized trials or a subsequent large-scale trial is needed to confirm the present findings 42 . Given that pregnant women are subject to human-subject protection in clinical studies, it will be difficult to conduct a randomized trial. However, the authors believe that the direction of bias is probably conservative: those with more co-morbidity and who are considered high risk are likely to undergo open appendicectomy, making the laparoscopic approach look spuriously superior. The increased risk of fetal loss seen here is therefore likely to be an underestimate. It was not possible to identify a statistically significant difference for preterm delivery and infection owing to the limited number of studies available for pooling. Finally, as the severity of appendicitis was not reported consistently in the pooled studies, a subgroup analysis to identify specific subgroups of women who might benefit from, or be harmed by, laparoscopic appendicectomy was not possible.

Disclosure
The authors declare no conflict of interest.   When appendicitis is suspected in a pregnant patient, the attending surgeon is faced with a dilemma. An operation might lead to fetal loss or preterm delivery. A delay in operative treatment carries the same risks. So, if there is a high degree of suspicion, surgeons will proceed to appendicectomy.
Appendicitis occurs most frequently during the second trimester. Although the appendix migrates to the right upper quadrant of the abdomen during pregnancy, in the first and second trimester both laparoscopic and open appendicectomies are technically feasible. The present meta-analysis summarizes the available evidence on laparoscopic and open appendicectomy in pregnant women. Laparoscopy seems to be associated with an almost twofold higher rate of fetal loss compared with open appendicectomy. These results are of potential importance for surgeons worldwide, but, in my opinion, need to be interpreted with caution.
The study that dominates the meta-analysis is a large study from the Californian patient registry 1 . In this study, a multivariable analysis showed that negative appendicectomy, perforated appendicitis and laparoscopic appendicectomy