Effect of the decision to perform hysteroscopy on asymptomatic patients before undergoing assisted reproduction technologies—a systematic review and meta-analysis

OBJECTIVE Routine hysteroscopic evaluation before assisted reproductive technology treatment is a novel approach with the potential to reduce assisted reproductive technology failure even in the absence of evidence of uterine pathology. Following the publication of several relatively high-quality trials on this topic, we sought to determine if this practice is beneficial. DATA SOURCES We searched Web of Science, MEDLINE, PubMed, Scopus, Cochrane Library, and ClinicalTrials.gov from each database's inception until May 31, 2022 with our search strategy, attempting to locate all randomized controlled trials assessing the use of hysteroscopy in otherwise asymptomatic women undergoing assisted reproductive technology. STUDY ELIGIBILITY CRITERIA We included only randomized controlled trials that included at least one of our selected outcomes, and we excluded any studies with suspicion of pathology before the time of hysteroscopy, other than knowledge of the patient's infertility. We included all the aforementioned studies regardless of procedures or modifications performed as a result of hysteroscopic findings. Our initial search yielded 1802 results, which were reduced to 1421 after removal of duplicates. Ultimately, 11 studies were found to meet our criteria and were included in our quantitative synthesis. METHODS We used ReviewManager software, version 5.4.1 to analyze the data, which we imported after manually gathering from the 11 studies. Continuous and dichotomous outcomes were imported as standard deviations. Pooled analysis was described as a mean difference, relative to 95 % confidence interval in cases of continuous data. Dichotomous outcomes were analyzed using risk ratios and 95% confidence intervals. In homogeneous outcomes, we used a fixed-effects model, and in heterogeneous outcomes we used a random-effects model. RESULTS Our results showed that hysteroscopy was associated with significant improvement in the clinical pregnancy rate (risk ratio, 1.27 [1.11–1.45]; P<.001). We found no differences between the hysteroscopy group and the control group in live birth rate (risk ratio, 1.26 [0.99–1.59]; P=.06), miscarriage rate (risk ratio, 0.99 [0.81–1.19]; P=.88), fertilization rate (risk ratio, 1.01 [0.93–1.09]; P=.88), incidence of multiple gestations (risk ratio, 1.29 [0.98–1.71]; P=.07), number of embryos transferred (mean difference, 0.04 [−0.18 to 0.26]; P=.73), chemical pregnancy rate (risk ratio, 1.01 [0.86–1.17]; P=.93), and number of oocytes retrieved (mean difference, 0.44 [−0.11 to 0.98]; P=.11). CONCLUSION We observed an improvement in the clinical pregnancy rate, but no significant improvement in the live birth rate with routine hysteroscopy before assisted reproductive technology treatment. We believe this does not represent sufficient evidence to recommend routine hysteroscopy for otherwise asymptomatic patients before assisted reproductive technology treatment at this time.


Introduction
Infertility, defined as the inability to achieve pregnancy after 1 year of unprotected, timed intercourse, can result in severe psychological, mental, and even medical disease for patients. 1,2 It affects approximately 37% of couples worldwide, and originates from the male factor in 57% of cases, from the female factor in 35%, and from both in the remaining 8%. 3 The most common causes of female-factor infertility include tubal blockage, ovulatory disorders, endometriosis, tubal abnormalities, uterine abnormalities, and hormonal disorders. 4 A basic workup for infertility may include assessment of semen, ovarian reserve and function, the uterine cavity, the patency of the fallopian tubes, and endocrinology. 5 Regarding the assessment of the uterine cavity, several options exist, including ultrasonography, hysterosalpingography, and hysteroscopy. 6 Some clinicians may also choose to forgo the assessment in asymptomatic women. 6 Although invasive, hysteroscopy provides the surgeon the possibility of immediate surgical repair of discovered pathology, which may include endometrial polyps, leiomyomas, septums, or other intrauterine pathology. 7−9 In the recent months, we noticed several relatively high-quality randomized controlled trials (RCTs) published on the topic of performing hysteroscopy before assisted reproductive technology (ART) treatment in otherwise asymptomatic women. Thus, in this meta-analysis we sought to estimate the efficacy of performing hysteroscopy before ART application in improving the outcomes of the different ART techniques in infertile patients, and any resulting treatments or treatment plan modifications made as a result of the findings of that hysteroscopy.

Search strategy
This study was conducted in accordance with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. 10 We conducted our search in electronic databases, including Web of Science, MEDLINE, PubMed, Scopus, Cochrane Library, and ClinicalTrials.gov, from the inception of each database until May 31, 2022 using the following strategy: (Infertility OR Sterility OR Subfertility OR Sub-Fertility) AND (Hysteroscop* OR "Uterine Endoscopy" OR ureteroscopy) AND ("live birth" OR "pregnancy rate" OR miscarriage).

Study selection
Two authors performed title and abstract screening followed by full-text screening. This process was performed according to the following eligibility criteria: Population: infertile women undergoing any ART technique. Intervention: hysteroscopy in otherwise asymptomatic women. Comparator: control group. Outcomes: clinical pregnancy rate (defined as ultrasound and serologic confirmation of intrauterine pregnancy), live birth rate per cycle, miscarriage rate, fertilization rate, multiple pregnancies, number of transferred embryos, chemical pregnancy rate, and number of oocytes retrieved. Study design: we included RCTs only and excluded all other study designs, meta-analyses, and reviews.

Quality assessment
We evaluated the risk of bias of the included RCTs according to the Cochrane Handbook for Systematic Reviews of Interventions. 11 We assessed 7 domains in each study: (1) random sequence generation; (2) selective reporting; (3) blinding of participants and personnel; (4) blinding of outcome assessment; (5) incomplete outcome data; (6) allocation concealment; and (7) other biases.

Data extraction
Data were retrieved manually from the included studies and placed into spreadsheets. We extracted baseline data such as the demographic data of patients, the number of patients with primary infertility, the number of patients with secondary infertility, the duration of infertility, and the causes of infertility. Then, we extracted the following outcomes: clinical pregnancy rate, live birth rate per cycle, miscarriage rate, fertilization rate, multiple pregnancies, number of transferred embryos, chemical pregnancy rate, and number of oocytes retrieved. We also extracted additional data that were required to complete our quality assessment of the included studies.

Statistical analysis
We used ReviewManager (RevMan) software, version 5.4.1 (Cochrane, London, United Kingdom) to analyze the data. Continuous and dichotomous AJOG Global Reports at a Glance Why was this study conducted? A: Despite no widely accepted guideline, our researchers noticed the publication of many studies where reproductive endocrinologists would routinely perform hysteroscopy prior to assisted reproduction technologies, such as in vitro fertilization, on asymptomatic patients with no identified pathology. As a result we sought out to find if evidence existed of a benefit to this practice.
Key findings A: The decision to perform hysteroscopy prior to assisted reproduction technologies seems to be associated with a higher clinical pregnancy rate, but no difference was seen in the rates of live birth, the miscarriages rate, the incidence of multiple gestations or the chemical pregnancy rate.

What does this add to what is known?
This study does not definitively show benefit to the routine practice of hysteroscopy prior to assisted reproduction technologies, but adds to the body of evidence suggesting that there may be some improvement with this practice, as evidenced by the increased rate in clinical pregnancies. Systematic Review ajog.org outcomes were imported from the spreadsheet into the RevMan software as mean §standard deviation and percentage and total, respectively. Pooled analysis was described as mean difference (MD), relative to 95% confidence interval (CI) in cases of continuous data, and dichotomous data were analyzed using risk ratio (RR) and 95% CI. In homogeneous outcomes, we used a fixed-effects model, whereas heterogeneous outcomes were analyzed under the random-effects model. We measured heterogeneity among studies using I-squared (Higgins I 2 ). Outcomes with I 2 >50% or P<.1 in the pooled analysis were considered heterogeneous. 12 ajog.org

Summary of the included studies
The PRISMA flow diagram of our search is shown in Figure 1. We analyzed 3938 infertile women from 11 included RCTs 14−24 ; 1821 patients underwent hysteroscopy, whereas 2117 patients were allocated to the control group. The 2 groups were found to be similar in sample size, age, and body mass index. The demographic data of patients, the number of patients with primary infertility, the number of patients with secondary infertility, the duration of infertility, and the causes of infertility are described in Tables 1−3.

Results of the quality assessment
The quality assessment of the included trials yielded an overall moderate risk of bias. Concerning the randomization domain, all studies reported proper randomization, so they were categorized as having low risk of bias except Kilic et al, 21 which was categorized as at high   Figure 2.  (Figure 3, A). Regarding the multicenter subgroup, data from 2 trials were analyzed. The combined RR did not show any difference between the hysteroscopy and the control group (RR, 1.04 [0.93−1.16]; P=.53). Pooled analysis was homogeneous (P=.68; I 2 =0%) (Figure 3, B).

Analysis of outcomes
The combined analysis of all trials from both subgroups favored the hysteroscopy group over the control group (RR,    Figure 10).

Discussion
In this meta-analysis, we sought to determine whether hysteroscopy could increase the success rates of ARTs such as in vitro fertilization (IVF), intracytoplasmic sperm injection, and intrauterine insemination. Our analysis demonstrated that performing hysteroscopy before the different assisted conception techniques could increase the clinical pregnancy rate significantly. However, hysteroscopy did not improve the live birth rate, miscarriage rate, fertilization rate, chemical pregnancy rate, multiple pregnancy, number of transferred embryos, or number of oocytes retrieved.
It is notable that a statistically significant difference in the clinical pregnancy rate was observed when performing hysteroscopy before ART, but this did not result in a statistically significant difference in the live birth rate, a result that the authors cannot completely explain. Multiple conditions were diagnosed and treated in the study arms that included hysteroscopy before ART. These treatments included hysteroscopic polypectomy, hysteroscopic myomectomy, lysis of adhesions, and medical treatment for endometritis diagnosed at time of hysteroscopy. One possible explanation for this phenomenon would be the persistence of pathology that would have otherwise prevented implantation later resulting in miscarriage as a result of the inability of our current treatments to truly resolve the condition. Statistically speaking, if this was true for even one of the pathologies, it would explain the lack of significance in the clinical pregnancy data. Another possible explanation is that many patients suffer from >1 pathology, and that after treating a cause found on hysteroscopy (polyps, fibroids, adhesions, endometritis), a second pathology (eg, genetic) then results in miscarriage.
Many previous analyses have found similar results. In 2008, El-Toukhy et al 25 conducted a meta-analysis that was limited to office hysteroscopy but otherwise considered similar outcomes, and found improvements in almost all outcomes with routine hysteroscopy. They attributed this beneficial role to the ability of hysteroscopy to visualize and treat intrauterine abnormalities including polyps, fibroids, endometritis, and intrauterine adhesions. In researching the percentage of female patients suffering from hysteroscopically correctable uterine pathology, we found several authors estimating that such pathologies are found at the time of hysteroscopy in approximately 50%. 26−28 Endometritis in particular was of interest in many of these studies because many authors have previously referenced the increased sensitivity of hysteroscopic Systematic Review ajog.org visualization over ultrasound in the diagnosis of chronic endometritis, with some sources citing a >33% increase in sensitivity. 29,30 This would lead to greater opportunity for the use of antimicrobial treatments for chronic endometritis in infertile patients given that chronic endometritis is estimated to be a cause in up to 40% of women suffering from infertility worldwide. 30 Another previous meta-analysis by Chung et al 31    ajog.org Systematic Review hysteroscopy could improve IVF outcomes in patients regardless of the presence of hysteroscopic uterine abnormalities. This study, however, was limited to patients who had repeated failures of ARTs as opposed to the routine practice of hysteroscopy.
In more recent reviews, Pundir et al 32 in 2014 performed a meta-analysis of 6 studies evaluating the efficacy of routine hysteroscopy before the first IVF cycle in infertile women. They concluded that hysteroscopy could significantly increase the clinical pregnancy rate (RR, 1.44; P=.01). Although consistent with our findings, this study was limited by significant heterogeneity in all outcomes. In addition, only 1 of the 6 included studies was randomized.
Before this study, the most recent analysis by Mao et al 33     Marchand. Effect of the decision to perform hysteroscopy on asymptomatic patients. Am J Obstet Gynecol Glob Rep 2023.

AJOG Global Reports May 2023
Systematic Review ajog.org 3932 patients and showed that hysteroscopy was associated with a better clinical pregnancy rate (P<.001) and implantation rate (P=.025) compared with the control group. As in our study, they found no significant difference between the hysteroscopy group and the control group in terms of live birth rate and miscarriage rate. This analysis   Although our findings were largely consistent with previous analyses, the exact mechanism of improving the clinical pregnancy rate after hysteroscopy is still unclear. Raju et al 23 suggested that hysteroscopy may treat small intrauterine lesions to increase the pregnancy rate. Shohayeb et al 34 reported that it was likely that endometrial "scratching" or performing a biopsy could change the features of the endometrium, facilitating embryo implantation. We do not believe that there is any clear consensus on the mechanism of action at this time.

Strengths and limitations
This was a large meta-analysis that had sufficient evidence to include only RCTs because of the newly published studies since the last analysis. With regard to limitations, 3 of the 8 outcomes were heterogeneous. Although heterogeneity decreases the certainty of evidence according to GRADE (Grading of Recommendations, Assessment, Development and Evaluations) guidelines, we were able to solve the heterogeneity in all cases. This was accomplished by subgroup analysis and the "leave-one-out method," as described in the Cochrane handbook. 12

Conclusion
We found that hysteroscopy performed before different ART procedures could improve the clinical pregnancy rate. A trend was observed toward an increased live birth rate in the hysteroscopy group, but statistical significance was not reached. We did not observe an increase in the miscarriage rate, fertilization rate, chemical pregnancy rate, multiple pregnancy, number of transferred embryos, or number of oocytes retrieved. Thus, we consider that the beneficial role of hysteroscopy should be assessed in further RCTs with more high-quality evidence before considering it a routine procedure in infertile women. Analysis of the cost-effectiveness of routine hysteroscopy before ART treatment would also provide valuable data. &