If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
School of Population Health, University of Queensland, Brisbane, AustraliaDepartment of Family and Community Medicine, King Saud University, Riyadh, Saudi Arabia
Department of Reproductive Health and Department of Gynecology, Shen-Zhen City Maternity and Child Healthcare Hospital, Shen-Zhen, Guang-Dong Province, People’s Republic of China
School of Population Health, University of Queensland, Brisbane, AustraliaDepartment of Endocrinology, Princess Alexandra Hospital, Brisbane, Australia
Obesity is known to interfere with reproductive outcomes in polycystic ovary syndrome. There is no consensus regarding the impact of obesity on reproductive outcomes after ovarian ablative therapy (OAT) and there is no level I evidence to answer this question. This systematic review and meta-analysis assessed the strength of the association between obesity and ovulation or pregnancy rates after OAT. MEDLINE and several other databases were searched from 2000 to September 2011 for studies reporting on OAT and reproductive outcomes. Data were synthesized to determine the relative risk of reproductive outcomes (ovulation and pregnancy) in lean (body mass index <25 kg/m2) compared with overweight or obese women. The study obtained 15 data sets (14 articles) for analysis, which included 905 subjects in the obese group and 879 subjects in the lean group. Lean women had increased ovulation rates (RR 1.43, 95% CI 1.22–1.66) compared with obese women. Pregnancy rates also showed a similar trend (RR 1.73, 95% CI 1.39–2.17). Reproductive outcomes were generally better in younger women, more recent studies and randomized controlled trials. It is concluded that lean women respond better to OAT than their obese counterparts. These epidemiological observations indicate that obesity alters reproductive outcomes after OAT negatively.
Obesity is known to interfere with reproductive outcomes in polycystic ovary syndrome. There is no consensus regarding the impact of obesity on ovarian ablative therapy (OAT) and there is no level I evidence to answer this question. We therefore undertook a systematic review and meta-analysis to assess the strength of the association between obesity and ovulation or pregnancy rates after OAT. We searched MEDLINE and several other databases from 2000 to September 2011 for studies reporting on OAT and reproductive outcomes. Data were synthesized to determine the risk ratio of reproductive outcomes (ovulation and pregnancy) in lean (BMI <25 kg/m2) as opposed to overweight or obese women. We obtained 15 datasets (14 articles) for analysis, which included 905 subjects in the obese group and 879 subjects in the lean group. Lean women had increased ovulation rates (RR 1.43, 95% CI 1.22–1.66) as compared to obese women. Pregnancy rates also showed a similar trend (RR 1.73, 95% CI 1.39–2.17). Reproductive outcomes were generally better in younger women, more recent studies and randomized controlled trials. We conclude that lean women respond better to OAT than their obese counterparts. These epidemiological observations indicate that obesity alters reproductive outcomes after OAT negatively.
Polycystic ovary syndrome (PCOS) is a medical condition considered to be one of the most common causes of irregular menstrual cycles, anovulation and infertility which manifests at adolescence (
). It is characterized by heterogeneous clinical and endocrinological manifestations including hyperandrogenism, neuroendocrine dysfunction, insulin resistance and the metabolic syndrome (
). In many of these women, anovulation leads to infertility, which can be managed with clomiphene citrate but approximately 25% fail to ovulate and require alternative treatment (
). Since then, there has been renewed interest in other related ovarian ablative therapies (OAT), such as laparoscopic ovarian drilling (LOD) when it became evident that they were less invasive and were associated with fewer complications (e.g. decreased multiple pregnancies and ovarian hyperstimulation compared with pharmacological therapy (
An economic evaluation of laparoscopic ovarian diathermy versus gonadotrophin therapy for women with clomiphene citrate-resistant polycystic ovarian syndrome.
). There have been many publications reporting the outcomes after ovarian drilling and other ablative therapies (e.g. ultrasonographic-guided ovarian stroma hydrocoagulation and transvaginal ultrasound-guided ovarian interstitial laser treatment) in the literature that suggest that they are also both effective and safe (
Transvaginal, ultrasound-guided, ovarian, interstitial laser treatment in anovulatory women with clomifene-citrate-resistant polycystic ovary syndrome.
) but there is a paucity of data regarding the effect of obesity on efficacy, even though obesity is frequently associated with PCOS.
While the mechanism of action of OAT is still unknown, it has been suggested that, in PCOS, steroid hormone excess might result in anovulation due to a direct effect on follicular maturation (
). Given this hypothesis, OAT could act presumably through sufficient destruction of the ovarian tissue which helps in decreasing the production of such ovarian steroids (
) and thus in obese women with additional extraovarian steroidogenesis, the response rate can be expected to be different compared with lean women with ovarian steroid excess.
To date, as far as is known, there has been only one study that examined this relationship, which, notably, found that obese women with PCOS do better than lean women in terms of reproductive outcomes (
Predictive value of serum androstenedione basal levels on the choice of gonadotropin or laparoscopic ovarian electrocautery as ovulation induction in clomiphene citrate-resistant patients with polycystic ovary syndrome.
). It was felt that this study might have been flawed as mean basal LH and androstenedione concentrations were reported to be higher in obese women while it is well known that they are reportedly lower in obese women (
). This collaborative meta-analysis of individual patient data specifically addresses this issue.
Materials and methods
Data sources
The literature (English and otherwise) was searched for articles that studied the reproductive outcome among PCOS patients following OAT with special reference to body mass data. PreMEDLINE and MEDLINE was searched using PubMed, Embase, Cochrane Central Register of Controlled Trials (CENTRAL), CINAHL and Scopus from 2000 to September 2011 for studies on OAT and reproductive outcomes; relevant articles were then hand searched for additional references. Articles were found with the use of controlled vocabulary terms (MeSH terms) as well as keywords. The database search strategies used are described in Supplementary Appendix 1 (available online only).
Study selection and patient outcomes
Studies were included that: (i) included a detailed and acceptable diagnosis of PCOS as per current guidelines; (ii) examined the effect of OAT in PCOS; (iii) assessed either the spontaneous reproductive outcomes (ovulation and pregnancy) or after additional co-interventions; and (iv) reported the body mass index (BMI) of these women. The BMI was defined as the patient’s weight in kg divided by the height in m2, with a cut-off point <25 kg/m2 for lean subjects and ⩾25 kg/m2 for overweight or obese subjects.
The primary reproductive outcome was defined as the ovulation rate and the secondary outcome was the pregnancy rate. Both were primarily computed per woman, but per cycle rates were also collected when available. Per woman data was calculated as the number of women that ovulated or had a pregnancy respectively out of the total group studied over the period of follow up. The per cycle data was computed as the number of ovulatory cycles (or pregnancies) observed in relation to the total number of cycles observed during the study.
Data abstraction
The process of data abstraction examined a wide range of variables and these included characteristics of patients such as case definition, description of patient’s characteristics, study population description, study design, methods of OAT, co-intervention used and the outcomes sought (Table 1). These data were extracted and tabulated with data supplied by the authors of the included studies who were sent a request to provide these data. Studies were excluded if the follow-up data were either not available, not extractable, not documented in the studies or if the authors did not respond. Additional exclusion was made if the studies were inconsistent. Although one of the studies fulfilled the inclusion criteria (
Predictive value of serum androstenedione basal levels on the choice of gonadotropin or laparoscopic ovarian electrocautery as ovulation induction in clomiphene citrate-resistant patients with polycystic ovary syndrome.
), it was excluded because the hormone concentrations reported were inconsistent with the classification of obese versus lean (see discussion for details).
Table 1Publications meeting the inclusion criteria.
CC-resistant PCOS women, OHSS with recombinant FSH; PCOS based on Rotterdam criteria including oligo- or anovulation, clinical or biochemical signs of hyperandrogenism and polycystic ovaries on ultrasound; medical and surgical history are considered; mean age 30.2 years
Bilateral ovarian drilling by fertiloscopy: punctures evenly spaced (6–10) on surface of each ovary with diameter 5–8 mm and depth up to 5 mm
Ovulation stimulation by FSH
Ovulation: Progesterone >0.9 ng/ml Pregnancy: Index pregnancy (first pregnancy following surgery)
Laparoscopic ovarian diathermy vs. clomiphene citrate plus metformin as second-line strategy for infertile anovulatory patients with polycystic ovary syndrome: a randomized controlled trial.
Primary infertile anovulatory CC-resistant women with PCOS; desiring pregnancy; PCOS diagnosed by presence of both clinical and biochemical hyperandrogenism and oligoanovulation and fulfilling the revised ESHRE and ASRM criteria; mean age 28.2 versus 27.5 years
LOD: 3–6 punctures, made by inserting needle cautery of 36 mm into ovary with diameter of 3 mm and depth of 4–5 mm
None
Ovulation: Plasma progesterone assay >10 ng/ml 7 days after expected menses Pregnancy: Increasing β-HCG and presence of intrauterine gestational sac by ultrasound
13.5
Palomba, unpublished observations
Total = 245 LOD = 245 Obese = 143 Lean = 102
PCS
Primary infertile anovulatory CC-resistant women with PCOS; desiring pregnancy; PCOS fulfilled the revised ESHRE and ASRM criteria of 2004 based on Rotterdam criteria; mean age 29.7 versus 30.1 years
LOD: 3–6 punctures, made by inserting needle cautery of 36 mm into ovary with diameter of 3 mm and depth of 4–5 mm
None
Ovulation: Plasma progesterone assay >10 ng/ml 7 days after the expected menses Pregnancy: Increasing β-HCG and presence of intrauterine gestational sac by ultrasound
CC-resistant women with PCOS; PCOS fulfilling the revised ESHRE and ASRM criteria of 2004 based on Rotterdam criteria; mean age 28.2 years; mean duration of infertility was 27.4
Bilateral LOD: monopolar electrocoagulation technique and monopolar hook electrode and with 5–10 mm incisions of 2–3 mm length of ovarian capsule
Metformin
Ovulation: Serum progesterone >8.0 ng/ml; ultrasound monitoring of follicular growth and measuring LH, FSH and progesterone Pregnancy: Confirmed by serum β-HCG and vaginal ultrasound
CC-resistant PCOS women with anovulatory infertility; some women received additional gonadotrophin therapy or IVF treatment or bilateral wedge resection; PCOS, by the criteria of 1990 NIH/NICHD consensus conference; all women and their partners were investigated for other common causes of infertility; mean age 30.5 years; mean duration of infertility 4.6 years
LOD: number of punctures related to size of ovary and number of subcapsular cysts
CC and gonadotrophins
Ovulation: Plasma progesterone ⩾30 nmol/l at mid-luteal phase, day 21 Pregnancy: Urine pregnancy test and serum β-HCG
Total = 200 LOD = 200 Obese = 120 Lean = 72 BMI not available = 8
RCS
CC-resistant PCOS women with anovulatory infertility >1 year; some women failed to conceive after HMG therapy; PCOS diagnosed by hormonal and ultrasonographic evidence; mean age 28.9 years
LOD: 3–10 punctures, depending on ovary size with diameter of 4 mm and depth of 7–8 mm
None
Ovulation: progesterone ⩾30 nmol/l at day 21 Pregnancy: Pregnancy at 12-month follow up
CC-resistant PCOS women with anovulatory infertility >1 year; most women were CC resistant and some were HMG resistant; PCOS diagnosed by the presence of the hormonal and ultrasonographic evidence of PCOS; mean age 33.3 versus 32.2 years
LOD: punctures, depending on ovary size with diameter of 4 mm and depth of 5–7 mm
None
Ovulation: Progesterone ⩾30 nmol/l at day 21 Pregnancy: Any pregnancy following the procedure over comparable follow up
Transvaginal ultrasound-guided ovarian interstitial laser treatment in anovulatory women with polycystic ovary syndrome: a randomized clinical trial on the effect of laser dose used on the outcome.
CC-resistant PCOS women seeking pregnancy; oligomenorrhoea or amenorrhea and anovulation for at least 2 years; polycystic ovaries by ultrasound; PCOS diagnosed by Rotterdam criteria; mean age 29.1 years; normal hysterosalpingography or laparoscopy in the last 3 years
Transvaginal ultrasound-guided ovarian interstitial laser treatment: coagulation point per ovary ranged from 1–5 points with diameter of nearly 10 mm
None
Ovulation: Series of transvaginal ultrasound to monitor the dominant follicle (with diameter of 16 mm) and urine LH measurement Pregnancy: Urine pregnancy test
CC-resistant PCOS women with infertility for at least 2 years due to anovulation, patent Fallopian tubes and normal semen analysis; PCOS diagnosed by Rotterdam criteria; age 18–38 years
Unilateral (right ovary) LOD: 4 punctures of 4 mm depth in the cortex
HCG
Ovulation: Transvaginal ultrasound to monitor the dominant follicle (with diameter of ⩾18 mm) and endometrial thickness on days 10, 12, 14; elevated serum progesterone measured at cycle days 21 and 23 Pregnancy: Serum β-HCG >25 mlU/ml; or intrauterine gestational sac detected by transvaginal ultrasound
CC-resistant PCOS women; patent Fallopian tubes and normal semen analysis; PCOS diagnosed by Rotterdam criteria; baseline hormonal and ultrasound tests; mean age 25.4–29.3 years
LOD: four punctures of 4 mm depth made using mixed current
CC and HCG if patients become anovulatory again
Ovulation: Serum progesterone ⩾5 ng/ml at cycle days 21–23 Pregnancy: Rising serum β-HCG with absence of menstruation; intrauterine gestational sac seen by transvaginal ultrasound at 6 weeks
Letrozole versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
CC-resistant PCOS women; patent Fallopian tubes and normal semen analysis; PCOS diagnosed by Rotterdam criteria; baseline hormonal and ultrasound tests were performed; mean age 26.4–27.3 years
LOD: four punctures of 4 mm depth made using mixed current
None
Ovulation: Serum progesterone ⩾5 ng/ml at cycle days 21–23 Pregnancy: Serum β-HCG ⩾50 mlU/ml with absence of menstruation; intrauterine gestational sac seen by transvaginal ultrasound at 6 weeks
Combined metformin and clomiphene citrate versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
CC-resistant PCOS women; patent Fallopian tubes and normal semen analysis; PCOS diagnosed by Rotterdam criteria; baseline hormonal and ultrasound tests were performed; the mean age 26.5 years
LOD: four punctures of 4 mm depth made using mixed current
None
Ovulation: Serum progesterone ⩾5 ng/ml at cycle days 21–23 Pregnancy: rising serum β-HCG with absence of menstruation; intrauterine gestational sac seen by transvaginal ultrasound at 6 weeks
CC-failure PCOS women; patent Fallopian tubes and normal semen analysis; PCOS diagnosed by Rotterdam criteria; baseline hormonal and ultrasound test ; mean age 26.3 years
LOD: four punctures of 4 mm depth made using mixed current
None
Ovulation: Serum progesterone ⩾5 ng/ml at cycle days 21–23 Pregnancy: Serum β-HCG ⩾50 mIU/ml in the absence of menstruation; intrauterine gestational sac seen by transvaginal ultrasound at 6–7 weeks
Ovulation: Having regular cycle associated with ovulation symptoms. Detecting mature follicle by regular transvaginal ultrasound on day 10–14. Serum progesterone levels at day 20 of cycle. Pregnancy: Serum β-HCG
CC-resistant PCOS women with infertility for at least 2 years due to primary or secondary infertility, patent Fallopian tubes, no hormonal treatment for 3 months prior to the study and normal semen analysis; PCOS diagnosed by Rotterdam criteria; age 18–38 years
LOD: 4–6 punctures into each ovary through the capsule
HCG
Ovulation: Transvaginal ultrasound to monitor the mean follicular diameter and endometrial thickness on days 10, 12, 14; elevated serum progesterone measured at cycle days 21 and 23 Pregnancy: Increased serum β-HCG; or intrauterine gestational sac with a beating fetal heart detected by transvaginal ultrasound
11.5
All studies designated obesity as BMI ⩾25 kg/m2 and lean at BMI <25 kg/m2, except for
Laparoscopic ovarian diathermy vs. clomiphene citrate plus metformin as second-line strategy for infertile anovulatory patients with polycystic ovary syndrome: a randomized controlled trial.
, which designated obesity as BMI 25–35 kg/m2 since BMI >35 was an exclusion criterion.
ASRM = American Society for Reproductive Medicine; BMI = body mass index; CC = clomiphene citrate; ESHRE = European Society of Human Reproduction and Embryology; HCG = human chorionic gonadotrophin; HMG = human menopausal gonadotrophin; LOD = laparoscopic ovarian diathermy; NIH/NICHD = National Institute of Health/National Institute of Child Health and Development ; OHSS = ovarian hyperstimulation syndrome; PCOS = polycystic ovary syndrome; PCS = prospective cohort study; RCS = retrospective cohort study; RCT = randomized controlled trial.
As different methodologies were used to answer a common question in the included studies, it is essential to differentiate between the higher- and lower-quality reports. Using an explicit model to pool results based on the differences between studies will facilitate clinical application of the findings if it is known which of the included studies used designs that provided less biased results with a high precision (
). Therefore, incorporating quality in a more explicit way in the current analysis was considered better than using the conventional random-effects model based on a random effect variance component (
) was tailored for use in the assessment of the methodological quality of the included studies (Supplementary Table S1). The balance of important prognostic indicators affecting reproductive outcomes across BMI groups were also considered. The prognostic score created looked at six items, namely age, BMI, duration of infertility, diet or exercise status, diabetic status and use of co-interventions. If five or six of these items were balanced, the score given was 1; if three or four items were balanced, the score was 0.5; if none, one or two of these items were balanced across comparison groups, or not documented in the study, the score given was 0.
The quality-effects model (QE) as well as (for comparison purposes) the random-effects model (RE) were used for all meta-analyses. The QE model pooled estimate has the advantage over the RE model of a meaningful probabilistic interpretation of the results (
). As the outcomes of interest are predefined and binary (ovulation or pregnancy), the relative risk (RR) was used to measure the association between BMI (lean versus obese) and the reproductive outcomes after the surgical management of PCOS. A value of 1 indicates equivalence. On the other hand, a value greater than 1 indicates that there is greater benefit for lean women in terms of the reproductive outcomes after OAT. To have a clearer presentation of the forest plot, the RR was plotted on a log scale as this would allow the confidence interval to become symmetrical about the point estimates of the RR. MetaXL version 1.2 was the software used for all analyses (www.epigear.com). It is well known that statistical tests for heterogeneity (e.g. the Q statistic and its variants) have low statistical power and thus, in addition to the application of statistical techniques, it has been recommended that common sense and a-priori biological knowledge, to the extent that it exists, must be vigilantly utilized when synthesizing the results of many studies (
). Thus, statistical heterogeneity was assumed across the various trial groups if tau-squared was >0 in favor of taking into account the P-value associated with the Cochran’s Q statistic. The confidence intervals in the QE model included a correction for over-dispersion as well as added uncertainty for decrements in quality (
). Finally, detecting discrepancy between big and small studies (which could be due to publication bias) was done by assessing the presence of asymmetry in the funnel plot. If the intercept on Egger’s regression deviated from zero with a P-value <0.05, the funnel plot was considered asymmetric.
Results
Characteristics of studies and subjects
After the search of the electronic databases, 557 studies were retrieved and an additional study was identified through other sources (Figure 1). After reviewing the abstracts of the identified articles, those not meeting the inclusion criteria, were excluded (n = 487). Emails were sent to authors of the remaining studies (n = 71) to ask them to participate in this collaboration. Fifty-one authors either did not have data or did not respond and four articles were excluded due to translation difficulty as they were written in Polish, Romanian and Russian. One study (
Predictive value of serum androstenedione basal levels on the choice of gonadotropin or laparoscopic ovarian electrocautery as ovulation induction in clomiphene citrate-resistant patients with polycystic ovary syndrome.
) was excluded due to the inconsistency noted previously. Although the author was contacted, there was no response by the time of submission of this manuscript. Therefore, 14 published research reports and one unpublished paper were included in this meta-analysis (
Laparoscopic ovarian diathermy vs. clomiphene citrate plus metformin as second-line strategy for infertile anovulatory patients with polycystic ovary syndrome: a randomized controlled trial.
Transvaginal ultrasound-guided ovarian interstitial laser treatment in anovulatory women with polycystic ovary syndrome: a randomized clinical trial on the effect of laser dose used on the outcome.
Letrozole versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
Combined metformin and clomiphene citrate versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
). The final 15 studies meeting the inclusion criteria and included in this meta-analysis are summarized in Table 1. In two studies, only ovulation per women data (
Letrozole versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
Combined metformin and clomiphene citrate versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
Laparoscopic ovarian diathermy vs. clomiphene citrate plus metformin as second-line strategy for infertile anovulatory patients with polycystic ovary syndrome: a randomized controlled trial.
Transvaginal ultrasound-guided ovarian interstitial laser treatment in anovulatory women with polycystic ovary syndrome: a randomized clinical trial on the effect of laser dose used on the outcome.
), the rest being cohort studies. Data were available for 1784 PCOS women (879 in the lean arms and 905 in the obese arms). These PCOS women received four different types of OAT: namely, ovarian drilling by fertiloscopy, laparoscopic ovarian diathermy, ultrasonographic-guided ovarian stroma hydrocoagulation and transvaginal ultrasound-guided ovarian interstitial laser treatment (Table 1). Women included in the pilot study of
The mean age of PCOS women included ranged between 26.3 and 30.5 years. The mean BMI of lean subjects ranged from 21.4 to 23.3 kg/m2 and in the overweight/obese arm ranged from 26.7 to 32.9 kg/m2 (Table 2). The mean duration of infertility ranged between almost 2 years and 5.7 years. In seven studies (
Letrozole versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
), anovulatory PCOS women who underwent OAT received further treatments. There were 377 PCOS women who had clomiphene citrate (CC) or gonadotrophins or both as co-interventions (not all women in these studies received co-interventions) and 100 participants were pretreated with metformin (
Laparoscopic ovarian diathermy vs. clomiphene citrate plus metformin as second-line strategy for infertile anovulatory patients with polycystic ovary syndrome: a randomized controlled trial.
Transvaginal ultrasound-guided ovarian interstitial laser treatment in anovulatory women with polycystic ovary syndrome: a randomized clinical trial on the effect of laser dose used on the outcome.
Letrozole versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
Combined metformin and clomiphene citrate versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
Laparoscopic ovarian diathermy vs. clomiphene citrate plus metformin as second-line strategy for infertile anovulatory patients with polycystic ovary syndrome: a randomized controlled trial.
Transvaginal ultrasound-guided ovarian interstitial laser treatment in anovulatory women with polycystic ovary syndrome: a randomized clinical trial on the effect of laser dose used on the outcome.
Letrozole versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
Combined metformin and clomiphene citrate versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
) favoured obese PCOS women in terms of the ovulation outcome. Nevertheless, the strongest effect size was for the studies by Abu Hashim et al., which were large trials with good quality scores. The QE model RR for all 14 data sets that reported this outcome was 1.43 (95% CI 1.22–1.66) suggesting that the lean PCOS women had 43% increase in ovulation rate after OAT compared with obese PCOS women (Figure 2). The RE model RR is depicted for comparison in Supplementary Figure S1.
Figure 2Forest plot depicting the relative risk of ovulation after OAT in lean versus obese PCOS women using the quality-effects model.
Letrozole versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
Combined metformin and clomiphene citrate versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
Laparoscopic ovarian diathermy vs. clomiphene citrate plus metformin as second-line strategy for infertile anovulatory patients with polycystic ovary syndrome: a randomized controlled trial.
) and one unpublished data set (Palomba, unpublished observations) were analyzed. All of these studies favoured lean PCOS patients as regards the ovulation outcome, and the QE model RR for all five studies was 1.90 (95% CI 1.46–2.48) (Figure 3). This suggests that there is twice the ovulation rate per cycle among lean patients compared with obese ones. The RE model RR is depicted for comparison in Supplementary Figure S2.
Figure 3Forest plot depicting the relative risk of ovulation per cycle after OAT in lean versus obese PCOS women using the quality-effects model.
In the per woman analysis, there was also a similar increase in pregnancy rate among lean patients compared with obese ones (RR 1.73, 95% CI 1.39–2.17; Supplementary Figure S3). All included studies (
Laparoscopic ovarian diathermy vs. clomiphene citrate plus metformin as second-line strategy for infertile anovulatory patients with polycystic ovary syndrome: a randomized controlled trial.
Transvaginal ultrasound-guided ovarian interstitial laser treatment in anovulatory women with polycystic ovary syndrome: a randomized clinical trial on the effect of laser dose used on the outcome.
Letrozole versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
Combined metformin and clomiphene citrate versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
) in terms of this outcome. For comparison, the respective RE model RR is depicted in Supplementary Figure S4. In the per cycle analysis, the QE model for the pregnancy outcome showed that lean PCOS women who underwent OAT had a 4-fold increase in pregnancy rate compared with obese patients (RR 4.14, 95% CI 2.08–8.23) (Supplementary Figure S5). Again, for comparison, the RE model RR is depicted in Supplementary Figure S6.
Sensitivity analysis and publication bias
The possible effect on the pooled result of various factors deemed to influence reproductive outcomes was determined by subgrouping the included studies, to find out the degree of sensitivity of the results to such changes. Some of the important factors that might influence the pooled results were considered in the sensitivity analysis. The pooled analysis was redone after subgrouping for mean age of participants (⩽30 years or >30 years), mean duration of infertility (⩽3.5 years or >3.5 years), year of publication (in or prior to 2005 or after 2005), and study design used (RCT or non-RCT). The results presented in Table 3 show that lean PCOS women have greater benefit from the surgery if: they are young (⩽30 years); duration of infertility is shorter (⩽3.5 years); were reported in studies that were randomized; and were reported in studies published after 2005.
Table 3Sensitivity analysis.
Parameter
Ovulation outcome
Pregnancy outcome
QE model
RE model
QE model
RE model
Mean age of subjects (years)
⩽30
1.57 (1.34–1.83)
1.56 (1.31–1.85)
1.89 (1.47–2.42)
2.06 (1.45–2.92)
>30
1.00 (0.86–1.16)
0.98 (0.85–1.13)
1.07 (0.69–1.67)
1.00 (0.53–1.88)
Mean duration of infertility (years)
⩽3.5
1.40 (1.13–1.73)
1.39 (1.08–1.77)
1.67 (1.30–2.15)
1.71 (1.25–2.35)
>3.5
1.46 (1.16–1.85)
1.43 (1.10–1.87)
1.89 (1.28–2.78)
1.45 (0.80–2.64)
Year of publication
⩽2005
1.15 (0.98–1.35)
1.13 (0.97–1.32)
1.09 (0.81–1.46)
1.00 (0.69–1.44)
>2005
1.54 (1.28–1.85)
1.54 (1.24–1.92)
2.11 (1.61–2.76)
2.28 (1.56–3.34)
Study design
RCT
1.59 (1.29–1.95)
1.59 (1.24–2.02)
2.70 (1.93–3.78)
2.96 (1.76–4.95)
Non-RCT
1.33 (1.09–1.62)
1.31 (1.04–1.65)
1.21 (1.00–1.46)
1.18 (0.96–1.45)
Co-interventions
Used
1.35 (1.09–1.67)
1.33 (1.04–1.69)
1.79 (1.23–2.59)
1.76 (1.20–2.58)
Not used
1.54 (1.25–1.91)
1.51 (1.19–1.92)
1.71 (1.28–2.27)
1.76 (0.98–3.17)
Values are pooled relative risk (95% CI).
QE = quality-effects; RCT = randomized controlled trial; RE = random-effects.
Predictive value of serum androstenedione basal levels on the choice of gonadotropin or laparoscopic ovarian electrocautery as ovulation induction in clomiphene citrate-resistant patients with polycystic ovary syndrome.
and the pooled result for the ovulation outcome per woman was RR 1.39 (95% CI 1.20–1.62) and per cycle RR 1.80 (95% CI 1.38–2.34). The pregnancy outcome results were per woman RR 1.68 (95% CI 1.35–2.09) and per cycle RR 3.53 (95% CI 1.95–6.39). These results were not significantly different from the original results.
To examine the possibility of publication bias, funnel plots were examined as missed or unpublished studies (publication bias) can be one of the reasons for funnel plot asymmetry. The funnel plots (each with a minimum of 10 studies) obtained for both reproductive outcomes were symmetric indicating that there was no obvious imbalance of studies in one direction that could have affected the pooled results across the 14 studies included (Supplementary Figures S7 and S8). Egger’s regression revealed an intercept not significantly different from zero (intercept = 1.6), consistent with a symmetrical funnel plot.
Discussion
While there has previously been some evidence that supports the hypothesis that ovulation rates after OAT are higher in lean compared with obese women (
), this meta-analysis adds to this evidence by confirming, as far as is known for the first time, that obesity is associated with a poorer outcome following OAT. Both ovulation rates per woman and per cycle were greater in lean women after this intervention but more so per cycle, suggesting that OAT results in a consistently better chance of a cycle being ovulatory in lean women after this intervention (Figure 3).
It is well known that obesity even without PCOS is associated with poor reproductive capacity (
) as compared with their average-build counterparts. There is also evidence that weight loss in obese or overweight women leads to improvement in reproductive outcomes (
It is not clear what obesity-related factor is responsible for this phenomenon. It could be argued that hormonal changes seen in obesity, such as elevated oestradiol, hyperinsulinaemia, lower androstenedione and lower basal LH (
), might be responsible for the reproductive failure here. It is now known that the decreased LH in obese PCOS women occurs at the level of the pituitary (
Inverse relationship between luteinizing hormone and body mass index in polycystic ovarian syndrome: investigation of hypothalamic and pituitary contributions.
Inverse relationship between luteinizing hormone and body mass index in polycystic ovarian syndrome: investigation of hypothalamic and pituitary contributions.
) but have not specifically looked at the putative role of androstenedione. It is plausible thus that androstenedione acts at the pituitary level to block a negative effect of oestradiol, which itself has been shown to act at the pituitary level (
) and explains the schema in Figure 4. Also, both LH and androstenedione are lowered after OAT (similar to what happens with obesity) but are increased after bariatric surgery (
Why does ovarian surgery in PCOS help? Insight into the endocrine implications of ovarian surgery for ovulation induction in polycystic ovary syndrome.
). In terms of insulin resistance, fasting insulin concentrations in PCOS patients remained unaltered in the first weeks to months after ovarian surgery in the majority of the studies (
Why does ovarian surgery in PCOS help? Insight into the endocrine implications of ovarian surgery for ovulation induction in polycystic ovary syndrome.
) and there is a growing body of evidence that serum hormone concentrations of insulin-resistant and non-insulin-resistant PCOS women are determined by obesity rather than insulin resistance since they are similar after controlling for bodyweight (
Figure 4Schema demonstrating the relationship between obesity and hormonal alterations in PCOS. The various mechanisms lead to insulin resistance and excess which triggers extraglandular oestradiol production as well as conversion of androstenedione to testosterone. This leads to a relative decrease in serum androstenedione and this lessens the inhibition imposed on oestradiol negative feedback on the hypothalamus. This decreases disinhibition and LH drops. In addition, the excess extraovarian oestradiol is postulated to be responsible for anovulation (
It thus seems that extraovarian oestradiol, by default, may be implicated in this failure of response to OAT since none of the other factors mentioned above have been linked to reproductive outcomes independently of obesity except for oestradiol concentrations (
). While oestradiol concentrations do decline after ovarian surgery, studies are inconclusive as obesity has never been factored in, and many studies did not also factor in time of measurement within the cycle (
Why does ovarian surgery in PCOS help? Insight into the endocrine implications of ovarian surgery for ovulation induction in polycystic ovary syndrome.
). However, ovarian surgery would not be expected to alter excess extraovarian oestradiol and it is well known that adipose tissue is a place for metabolizing and converting androgens to oestrogens via the aid of steroidogenic enzymes (
). Two such enzymes are expressed in adipose tissues namely, cytochrome P450-dependent aromatase and 17β-hydroxysteroid dehydrogenase/17-ketosteroid reductase with the former being responsible for aromatization of androstenedione and testosterone to oestrone and oestradiol respectively while the latter facilitates conversion of androstenedione to testosterone (
). Therefore, obese women have an extraovarian source of circulating oestrogens that remains intact even after OAT. The current study proposes the schema in Figure 4 as a likely sequence of events that maintains anovulation after surgery in obese PCOS. Indirect parallels to such a follicular-phase steroid abnormality also come from studies of follicular-phase oestradiol in PCOS (
Subnormal follicular-phase serum progesterone levels and elevated follicular-phase serum estradiol levels in young women with insulin-dependent diabetes.
Laparoscopic ovarian diathermy vs. clomiphene citrate plus metformin as second-line strategy for infertile anovulatory patients with polycystic ovary syndrome: a randomized controlled trial.
Transvaginal ultrasound-guided ovarian interstitial laser treatment in anovulatory women with polycystic ovary syndrome: a randomized clinical trial on the effect of laser dose used on the outcome.
Letrozole versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
Combined metformin and clomiphene citrate versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
Predictive value of serum androstenedione basal levels on the choice of gonadotropin or laparoscopic ovarian electrocautery as ovulation induction in clomiphene citrate-resistant patients with polycystic ovary syndrome.
, who concluded that PCOS women benefited more from LOD if they were obese. Given the inverse biochemical profile in this study, this would suggest a methodological error or inadvertent mix up of study groups in
Predictive value of serum androstenedione basal levels on the choice of gonadotropin or laparoscopic ovarian electrocautery as ovulation induction in clomiphene citrate-resistant patients with polycystic ovary syndrome.
Despite the majority of pooled studies in this analysis being concordant for better outcomes in lean women, obese women fared marginally better in terms of ovulation in two studies (
) were discordant for ovulation, both effect sizes were close to unity and the current study attributes this to the possibility of design-related factors. In
, there was a greater imbalance between potential confounders in both groups. In particular, obese women were younger and this likely confounded the effects of obesity and would also explain why they report poorer pregnancy outcomes in lean women as well (
). Indeed when a sensitivity analysis was conducted, age was a significant confounder and the body mass-related differences were considerably attenuated in older women. Several other factors that might have had impact on the pooled results were looked at in the sensitivity analysis, including duration of infertility, year of publication and study design. All these factors showed a significant association with reproductive outcomes. It was found that shorter duration of infertility (⩽3.5 years) was associated with increased success after OAT. This sort of increase in success was also seen for RCTs and studies published after 2005. This is not surprising, since duration of infertility is known to be inversely correlated with the success rate of OAT in terms of reproductive outcomes (
). Finally, studies after 2005 reported better outcomes after OAT compared with older studies probably as a result of refinement of surgical technique for OAT over time. Despite these influences, the trend of lean women having better outcomes was retained.
In this meta-analysis, there is a difference between the QE and RE weights. Assessing the differences in quality of each included study is the main difference characterizing the former model. This numerical assessment of methodological quality facilitates the non-random adjustment of weights (
). By looking at the reproductive outcomes of per woman results, the pooled RR for the QE model were similar to that of the RE model with a value above 1 and the confidence intervals were narrow; this is because effect size heterogeneity was minimal. However, when per cycle data were pooled for this meta-analysis, the confidence interval for QE result for only the pregnancy outcome was different from that of the RE model, with the former being statistically significant. This could be explained by the fact that the RE variance component acts as a nuisance variable and artificially inflates model variance (
In conclusion, it is clear from the results of this meta-analysis that obesity alters the efficacy of OAT and that BMI should be carefully considered before this procedure to define the optimal management of ovulation in PCOS patients.
Letrozole versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
Combined metformin and clomiphene citrate versus laparoscopic ovarian diathermy for ovulation induction in clomiphene-resistant women with polycystic ovary syndrome: a randomized controlled trial.
An economic evaluation of laparoscopic ovarian diathermy versus gonadotrophin therapy for women with clomiphene citrate-resistant polycystic ovarian syndrome.
Why does ovarian surgery in PCOS help? Insight into the endocrine implications of ovarian surgery for ovulation induction in polycystic ovary syndrome.
Inverse relationship between luteinizing hormone and body mass index in polycystic ovarian syndrome: investigation of hypothalamic and pituitary contributions.
Laparoscopic ovarian diathermy vs. clomiphene citrate plus metformin as second-line strategy for infertile anovulatory patients with polycystic ovary syndrome: a randomized controlled trial.
Predictive value of serum androstenedione basal levels on the choice of gonadotropin or laparoscopic ovarian electrocautery as ovulation induction in clomiphene citrate-resistant patients with polycystic ovary syndrome.
Transvaginal, ultrasound-guided, ovarian, interstitial laser treatment in anovulatory women with clomifene-citrate-resistant polycystic ovary syndrome.
Transvaginal ultrasound-guided ovarian interstitial laser treatment in anovulatory women with polycystic ovary syndrome: a randomized clinical trial on the effect of laser dose used on the outcome.
Subnormal follicular-phase serum progesterone levels and elevated follicular-phase serum estradiol levels in young women with insulin-dependent diabetes.
Declaration: The authors report no financial or commercial conflicts of interest.
Footnotes
Dr Leena Baghdadi graduated from medical school at Umm Al-Qura University, Makkah, in 2005. She joined the Clinical Epidemiology Unit of the University of Queensland in Brisbane under the supervision of A/Professor Suhail Doi, where she completed her Masters and is now working towards a PhD. She also holds a joint appointment as a lecturer at the Department of Family and Community Medicine at King Saud University in Riyadh, Saudi Arabia. She is keenly interested in evidence-based health care and aims to further her academic career in clinical research.
00336-7&id=gr1.jpg){kind=link}
00336-7&id=gr2.jpg){kind=link}
00336-7&id=gr3.jpg){kind=link}
00336-7&id=gr4.jpg){kind=link}