Criteria for study selection

This study was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines. All published cohort studies and randomized trials (including at least 5 patients) that investigated the impact of excision of benign nonendometriotic ovarian cysts on ovarian reserve as determined by changes in postoperative serum anti-Müllerian hormone concentration were included in this metaanalysis.

Outcome measures

Search strategy

An extensive electronic database search was performed using MEDLINE, Scopus, Embase, and ScienceDirect to identify published research articles between January 2000 and Jan. 31, 2016, on the impact of excision of benign ovarian cysts (excluding endometriomas) on ovarian reserve as determined by serum anti-Müllerian hormone concentration. No restrictions were placed on language. A combination of the following search terms was used: laparoscopy, laparotomy, ovarian cystectomy, excision, anti-Müllerian hormone, benign ovarian cysts, and ovarian reserve. All searches were carried out by the first author and then independently repeated using the same criteria by an accredited clinical librarian. All relevant reports were retrieved, and their reference lists were reviewed manually to identify further studies. A manual search of related articles on PubMed was also performed. We also considered published abstracts from conferences.

Data extraction

All the identified papers were evaluated according to a standardized format including study design, methods, participant characteristics, intervention, and results. Two investigators scored the studies and collected the information independently. In the case of discrepancies in scoring between the 2 investigators, a consensus was reached after discussion or after involvement of the senior investigator. In 5 studies the mean ± SD was not presented. The authors of these studies were contacted, but only 2 replied, providing the missing data, which were used in our analysis.

In another study, which was a conference abstract, the authors did not describe the methods of recruitment (inclusion and exclusion criteria), nor they specified the type of anti-Müllerian hormone assay kit. This study was included in the initial analysis but was excluded from the sensitivity analysis. The authors were contacted to provide the missing information, but no response was received.

Quality of included studies and risk of bias assessment

The quality and risk of bias of the included studies were assessed using the modified Newcastle-Ottawa Scale, as previously described. The original Newcastle-Ottawa Scale assesses 3 main categories including selection, comparability, and outcomes, giving a maximum of 4, 2, and 3 stars for each category, respectively. This scale was modified to suit the nature of this study, giving a maximum of 3 stars for selection, 4 for comparability, and 2 for outcome criteria.

Selection was rated according to recruitment bias, selection of consecutive patients, and power calculation. Comparability was assessed based on studies adjusting their analysis for 4 confounders including patients’ age (<40 years), cyst diameter (>5 cm), baseline serum anti-Müllerian hormone (≥ 3.1 ng/mL), and cyst laterality.

Outcome was scored according to completeness of at least 3 months of follow-up after surgery. It is generally agreed that a limit of 5 stars could identify studies at low risk of bias. However, in this study, we have given more weight to comparability factors and used the cutoff level of 6 stars, with a minimum of 3 stars in the comparability category. Table 1 shows the results of quality scores of the studies included in this analysis.

Data analysis

Pre- and postoperative data including mean ± SD serum concentrations of anti-Müllerian hormone (nanograms per milliliter) and follicle-stimulating hormone (milliinternational units per milliliter) and antral follicle count were extracted from the individual studies and pooled using RevMan software (Review Manager, version 5.1, The Cochrane Collaboration, 2011; The Nordic Cochrane Centre, Copenhagen, Denmark).

The weighted mean difference between the pre- and postoperative values was calculated. Statistical heterogeneity was assessed by a χ ² test and I ² statistics. A χ ² statistic larger than its degree of freedom or an I ² higher than 50% was indicative of significant heterogeneity between studies. When heterogeneity was significant, a random-effect model was used for metaanalysis. A fixed effect metaanalysis was used when there was no significant heterogeneity.

The initial analysis included data from all studies, irrespective of length of follow-up and cyst characteristics (diameter, laterality, and pathological type). In studies with multiple postoperative measurements at different follow-up points, we used the latest anti-Müllerian hormone level. Further subgroup analyses of anti-Müllerian hormone levels were then performed based on the laterality of the excised cysts, anti-Müllerian hormone kits used, and duration of follow-up. To examine and account for heterogeneity, a sensitivity analysis was carried out based on modified the Newcastle-Ottawa Scale for risk of bias as described above. Studies with the lowest risk of bias were defined as those with a score of ≥6 with at least 3 stars on comparability score and using the same surgical approach (laparoscopy).

Excluded studies

After the initial screening on the basis of the title and abstract, 10 studies did not use the anti-Müllerian hormone to investigate ovarian reserve after surgery for benign nonendometriotic cyst and were therefore excluded. Five further studies were excluded, one because of the small number of patients (n = 3), another study because of extremely low anti-Müllerian hormone levels (below the sensitivity of anti-Müllerian hormone assays), and 3 studies because of missing the mean ± SD of serum anti-Müllerian hormone concentrations (data were either presented as median [interquartile range]) or mean ± SE percentage. The authors of the latter 3 studies were contacted to provide the anti-Müllerian hormone data, but no response was received.

Included studies

Details of the included 10 studies are shown in Table 2 .

Study design

The included studies were all cohort studies except one, which was a randomized controlled trial. However, both arms of the randomized controlled trial (laparoscopy and laparotomy) were combined and included in the initial metaanalysis as a cohort study. One arm (laparoscopy) was then included in the sensitivity analysis.

Participants

Selection criteria were appropriate for all studies. All participants underwent the same type of surgery (cystectomy) through laparoscopy except one study, in which the patients were randomly allocated to either laparoscopy or laparotomy. Patients of this randomized controlled trial were consecutive and were followed up within their particular group, and the results were given separately for each arm of the randomized controlled trial. All studies reported inclusion and exclusion criteria that were appropriate except one. The author of that study was contacted but did not reply. All patients were accounted for in all studies.

Ovarian cyst diagnosis

Most of the studies reported that the initial diagnosis of the cysts was achieved through transvaginal ultrasound. The ultrasound scans were performed by gynecologists with sufficient experience. Postoperatively, the nature of the cyst was confirmed with histopathological examination. Six studies reported the mean ± SD cyst diameter and 6 studies determined the side and laterality of the cysts.

Surgery and length of follow-up

All studies included patients undergoing ovarian cystectomy carried out laparoscopically except 1 study, which was an randomized controlled trial comparing laparoscopic vs open cystectomy. The length of follow-up was 1 month in 4 studies, 3 months in 6 studies, and 6 months in 1 study.

Anti-Müllerian hormone kits

Anti-Müllerian hormone concentration was measured using 1 of the 4 available kits. The first one was Immunotech anti-Müllerian hormone/Müllerian-inhibiting substance enzyme immunoassay kit (Immunotech, Beckman Coulter, Marseille, France). The intra- and interassay coefficients of variation for the anti-Müllerian hormone assay were below 12.3% and 14.2%, respectively, with a detection limit of 0.14 ng/mL.

The second anti-Müllerian hormone kit was Diagnostic Systems Laboratories active Müllerian–inhibiting substance/anti-Müllerian hormone enzyme-linked immunosorbent assay kit (Diagnostic Systems Laboratories, Webster, TX). The intra- and interassay coefficients of variation for the anti-Müllerian hormone were 4.6% and 8.0%, respectively, with a detection limit of 0.017 ng/mL.

The third kit was the AMH Gen II enzyme-linked immunosorbent assay (Beckman Coulter, Chaska, MN). The intra- and interassay coefficients of variation for the anti-Müllerian hormone assay were both below 10%, with a detection limit of 0.08 ng/mL. Two studies used this kit including Ergun et al, who used the modified AMH Gen II kit and Kwon et al, who used the original Gen II assay. The fourth kit was the ultrasensitive anti-Müllerian hormone enzyme-linked immunosorbent assay Ansh Labs assay (Ansh Labs, Glasgow, United Kingdom). The intra- and interassay coefficients of variation for the anti-Müllerian hormone were 0.02 (2.22/95) and 7.81 (15.62/2), respectively, with a detection limit of 0.06 ng/mL.

Potential source of bias

In all studies, patients were selected in a consecutive fashion. The selection method was clearly described, making it easy to assess selection bias.

Overall pooled results for all studies

The initial analysis of the 10 studies included all 367 patients who underwent cystectomy for unilateral or bilateral benign nonendometriotic ovarian cysts. The analysis revealed a statistically significant postoperative fall in serum anti-Müllerian hormone concentrations (weighted mean difference, –1.14 ng/ml; 95% confidence interval, –1.36 to –0.92). Heterogeneity between studies was low (I ² = 43 %) ( Figure 2 ).

WMD in serum AMH concentrations after ovarian cystectomy

WMD in serum anti-Müllerian hormone concentrations after excision of benign nonendometriotic ovarian cysts: pooled results for all 10 studies.

AMH, anti-Müllerian hormone; CI , confidence interval; WMD, weighted mean difference.

Mohamed. Impact of ovarian cystectomy on ovarian reserve. Am J Obstet Gynecol 2016 .

Subgroup analysis

Sensitivity analysis

Pooled analysis of 8 studies with a low risk of bias (as defined in previous text) including 297 patients showed a statistically significant fall in postoperative serum anti-Müllerian hormone concentration (weighted mean difference, –1.05; 95% confidence interval, –1.29 to –0.81; I ² = 43%). Heterogeneity between studies was low.

Studies with ovarian cyst >5 cm

Seven studies including 276 patients were identified. Pooled analysis revealed a statistically significant fall in postoperative serum anti-Müllerian hormone concentration (weighted mean difference, –1.13; 95% confidence interval, –1.56 to –0.70; I ² = 62%). Heterogeneity between studies was high.

Studies with different histological types

Analysis of 6 studies including 158 patients with dermoid cysts revealed a statistically significant fall in serum anti-Müllerian hormone concentration (weighted mean difference, –1.27; 95% confidence interval, –1.93 to –0.62; I ² = 55%). Similarly, an analysis of 4 studies including 84 patients with cystadenomas showed a statistically significant decline in serum anti-Müllerian hormone concentration (weighted mean difference, –1.59; 95% confidence interval, –2.00 to –1.17; I ² = 0%).

Secondary outcomes

Three studies measured changes in serum follicle-stimulating hormone concentrations, but only 2 (including 95 patients) of these provided full pre- and postoperative data. Pooled analysis of these 2 studies revealed no significant change in circulating follicle-stimulating hormone following ovarian cystectomy (weighted mean difference, –0.50; 95% confidence interval, –1.28 to 0.28; I ² = 0%). The authors of the other study reported that follicle-stimulating hormone levels did not change after surgery, but they failed to present the actual data.

With regard to antral follicle count, although 1 study included this ovarian reserve marker as an outcome measure, the authors provided the postoperative antral follicle count data only and failed to present the preoperative values.