Comparing Female Permanent Contraception Options in High Resource Countries: A Systematic Review

Background: Multiple options for permanent or long-acting contraception are available, each with adverse effects and non-contraceptive benets. A comprehensive comparison of methods to support decision-making for people seeking to end their fertility and their healthcare providers is needed. We aimed to understand what is known from high quality studies about the comparability of options for permanent contraception. We sought studies comparing these methods of permanent or long-acting contraception: laparoscopic tubal ligation, hysteroscopic tubal occlusion, bilateral salpingectomy, and levonorgestrel-releasing intrauterine contraception (LNG-IUC), for effectiveness, adverse events, tolerability, patient recovery, non-contraceptive benets and healthcare system costs among females in high resource countries seeking to permanently avoid conception. Methods: We followed PRISMA guidelines, searched EMBASE, Pubmed (Medline), Web of Science, and screened retrieved articles to identify additional relevant studies. We extracted data on population, interventions, outcomes, follow-up, health system costs, and study funding source. We assessed risk of bias using the Newcastle-Ottawa Scale, excluding studies with high risk of bias. Due to considerable heterogeneity, we performed a narrative synthesis. Results: Our search identied 6,612 articles. We reviewed the full text of 154 studies, yielding 35 studies which met inclusion criteria. We excluded 10 studies with high risk of bias, retaining 25 at low or medium risk for bias in our synthesis. Most studies assessed hysteroscopic tubal occlusion and/or laparoscopic tubal ligation. Most comparisons reported on effectiveness and adverse events, with fewer reporting tolerability, patient recovery, non-contraceptive benets, and healthcare system costs. No comparisons reported accessibility, eligibility, or follow-ups required. We found inconclusive evidence comparing the effectiveness of hysteroscopic tubal occlusion to laparoscopic tubal ligation. All studies reported adverse events. All forms of tubal interruption reported a protective effect against cancers. Tolerability appeared greater among tubal ligation patients compared to hysteroscopic tubal occlusion patients. No medium or high-quality studies compared LNG-IUC to other methods. Conclusions: High-quality studies comparing outcomes relevant for those seeking female permanent contraception are needed to support informed decision-making. is to compare surgical forms of permanent contraception, as tubal ligation or with alternative options, such as intrauterine contraception. at three 1.04 Sixteen studies assessed adverse effects using prospective/retrospective cohort designs. Results are organized by intervention: six compared hysteroscopic tubal occlusion and laparoscopic tubal ligation, 24, 25, 28–31 three compared laparoscopic tubal ligation and a control, 32–34 ve compared laparoscopic tubal ligation and bilateral salpingectomy, 11, 35–37, 40 one compared hysteroscopic tubal occlusion, laparoscopic tubal ligation with controls, and bilateral salpingectomy with controls, 38 and one compared bilateral salpingectomy with laparoscopic tubal ligation and with matched controls. One and risk among women undergoing tubal at 24 months post-procedure. laparoscopy minilaparotomy, laparotomy, hysteroscopic occlusion using Essure™ compared to an unspecied destruction or occlusion of fallopian tubes. Hazard ratios for ectopic pregnancy did not remain signicant for laparoscopy with Filshie clip, minilaparotomy, and laparotomy, no ectopic pregnancies reported for bilateral salpingectomy nor hysteroscopic tubal occlusion with Essure TM Previous meta-analyses have compared outcomes between hysteroscopic tubal occlusion and laparoscopic tubal ligation, yet few have considered other options available for permanent contraception. Several studies illustrate an increasing trend of using bilateral salpingectomy for permanent contraception, and there are many calls to consider LNG-IUC as a permanent contraception alternative, as it offers similar effectiveness rates with fewer adverse effects. High quality studies that compare traditional forms of permanent contraception, such as the laparoscopic tubal ligation, with new and alternative methods, are urgently needed to provide evidence on the implications for all options available to people seeking permanent female contraception.


Introduction
Permanent contraception is the most common method of fertility control worldwide. 1,2 Globally, nearly one in four females in high income countries use either intrauterine contraception or female sterilization. 3 Female permanent contraceptive methods are the fourth most commonly relied upon method for preventing pregnancy among people in Canada,4 and the second most common method in the United States. 5 Permanent contraception is traditionally achieved using laparoscopic tubal ligation, with the clipping, coagulation, or other blocking of the fallopian tubes to prevent sperm from travelling to an ovulated oocyte. 2 However, in the last two decades, other methods to achieve permanent contraception have emerged including bilateral salpingectomy and the levonorgestrelreleasing intrauterine contraceptive (LNG-IUC), while Essure,™ micro inserts used in hysteroscopic tubal occlusion, has been taken off the market in select countries. [6][7][8] Bilateral salpingectomy is increasingly being considered as an alternative option to laparoscopic tubal ligation. In the Canadian province of British Columbia, after an educational campaign led by the Ovarian Cancer Research Team highlighting emerging evidence that ovarian cancer originates in the fallopian tubes 9 , the province saw an increased trend in salpingectomy from 0.4% of female sterilization procedures in 2008 to 33.0% in 2011. 10 Similar increases were seen in Texas and New York, ranging from 4.8 to 77.5% over a similar time period. 11 In June 2017, the Society of Obstetricians and Gynaecologists of Canada released a committee opinion that physicians should discuss the protective bene t of tubal ligation and the "fact that the complete removal of the fallopian tube may provide additional bene t." 12 The additional bene t is primarily the anticipated greater protective effect against ovarian cancer with no additional side-effects over those with laparoscopic tubal ligation. 9,13 Two Markov simulation models, creating theoretical cohorts of patients using previously published data, suggested that bilateral salpingectomy could theoretically reduce ovarian cancer risk, diagnoses, 14 and death. 15 Bilateral salpingectomy was theorized to contribute to additional quality years of life when compared to laparoscopic tubal ligation, with a mean incremental cost of $152 per person. 14 Additional decades of follow up after salpingectomy are still needed to understand how closely reality will compare to this simulation.
Although a reversible contraceptive method, insertion of the LNG-IUC has e cacy and effectiveness rates similar to tubal ligation. [1][2][3]5 Although the LNG-IUC requires re-purchase and re-insertion every 5 or more years, 16 this option may be attractive due to the avoidance of surgery and of the faster recovery time compared to that required for other methods. 17 For this reason, we have included LNG-IUC as a permanent contraception alternative for the purpose of this analysis.
Despite the many options available, we were unable to nd a guide for clinicians or for people seeking female permanent contraception that systematically compares available methods according to important outcomes, nor any that include comparable long-acting reversible contraception. Permanent contraception decision-making can be complex. It is important that the realities and options are outlined so that people can make an informed choice that is aligned with their reproductive goals. Ultimately, the choice of which contraceptive method to utilize should be based on an informed understanding of not only the effectiveness, but also the potential risks, additional bene ts, and accessibility in terms of associated costs and recovery time for the patient.

Objectives
We aimed to understand what is known from high quality studies about the comparability of permanent methods of contraception. In particular we included laparoscopic tubal ligation, hysteroscopic tubal occlusion, bilateral salpingectomy, and insertion of the LNG-IUC among people seeking permanent female contraception in high resource countries. Outcomes included effectiveness at preventing pregnancy, adverse events, tolerability, patient recovery, and noncontraceptive bene ts. Secondary outcomes included the length of procedure, costs to the healthcare system, eligibility, accessibility, and follow-ups required to ensure completion or for safety monitoring. Thorough de nitions of study objectives are explained in the systematic review protocol. 18

Methods
We followed PRISMA guidelines 19 for this analysis in accordance with the accompanying explanation and elaboration paper. 20 The PRISMA checklist is available in Additional le 1.

Protocol registration
We pre-speci ed and previously published 18 objectives and analyses in a protocol registered on PROSPERO (CRD42016038254).

Eligibility criteria
We included studies that met the following criteria: Population comprised of females of reproductive age , without major comorbidities; Prospective and retrospective cohort, case-control, or randomized control trial methodology; The paper was published in a peer reviewed journal in English; The analysis took place in high resource countries as de ned by the World Bank Country and Lending groups; 21 The interventions included comparisons of any two or more of: laparoscopic tubal ligation, hysteroscopic tubal occlusion, bilateral salpingectomy, or insertion of the LNG-IUC performed as permanent contraception, and/or controls; The outcomes assessed included at least one of the following: effectiveness, adverse events, tolerability, non-contraceptive bene ts, patient recovery, accessibility, length of the procedure, follow-ups required, eligibility, or costs to the healthcare system.
We excluded studies with these characteristics: A case study or case series design; Conducted outside of high resource countries; Interventions included concomitant procedures.

Information Sources
We searched EMBASE, Pubmed (Medline), and Web of Science using a combination of MeSH terms and key words related to hysteroscopic tubal occlusion, laparoscopic tubal ligation, the LNG-IUC, and bilateral salpingectomy. We also reviewed the references of relevant articles as secondary screening. We did not set date restrictions, and performed the last search on January 30th, 2019.

Study selection
Three authors (RG, BC, BV) independently reviewed titles and abstracts of initial articles based on relevance. RG reviewed all identi ed articles, and BC and BV each reviewed a subset of articles and together reviewed all articles. After comparison of articles for relevance based on titles and abstracts, RG reviewed full text articles for inclusion or exclusion, noting reasons for exclusion. Data collection process RG created the data extraction form and initially pilot-tested the form on a randomly selected subset of studies to determine comprehensiveness. We extracted data from each study that met the inclusion criteria including: population, intervention, comparisons, outcomes, and study design (PICOS); follow-up period; and funding source for the study, where available.
RG extracted data from all relevant articles, and BV independently extracted data from a sample of articles. We compared the data extraction forms for accuracy. Any discrepancies were adjudicated by the senior author (WVN).

Data items
We de ned all data items, including de nitions of the variables sought, in detail in our protocol. 18

Risk of bias in individual studies and across studies
We used the Newcastle Ottawa Scale (NOS) to determine risk of bias for cohort and case-control studies. 22 We assessed risk of bias for each included study, and presented the results in a table strati ed by study design. Any articles assessed to be at high risk of bias (NOS score < 7) were excluded from the narrative analysis.
We assessed the cumulative risk of bias based on the risk of bias found in individual studies, along with careful consideration of any outcome reporting bias, incomplete study data, or overall quality of the evidence presented and synthesized.

Synthesis of results
We aimed to perform a network meta-analysis, but heterogeneity was assessed as substantial (I 2 ≥ 80%), with wide variability in outcome reporting that precluded a valid pooling of results. Therefore, we undertook a narrative synthesis in accordance with the guidelines of the Cochrane Consumers and Communication Review Group. 23 We organized the narrative synthesis by outcome, describing the similarities, differences, and patterns of results.

Study Selection
Figure 1 details our study selection process, including reasons for exclusion, following PRISMA guidelines. 19 Our database search of EMBASE, Pubmed (Medline), and Web of Science revealed 6,826 documents, and we identi ed an additional 25 documents through secondary screening. After we excluded duplicates (239), RG, BC, BV reviewed titles and abstracts based on relevance. RG reviewed all 6,612, and BC and BV independently reviewed 1,647 and 5,088 studies respectively. Of these, we excluded 6,458 because they did not meet the inclusion criteria. RG reviewed the full text of the remaining 154 studies. BV reviewed 30 full text studies, and compared them with the relevant articles assessed by RG to check for accuracy. As discussed in the protocol, we did not include unpublished studies in this review. We extracted data and assessed risk of bias for 35 studies. We found medium to high risk of bias in 10 studies, leaving a total of 25 studies included in the narrative review synthesis.
People who had severe comorbidities were excluded. All studies were pairwise comparisons of females of reproductive age, who underwent an interval contraceptive procedure including either laparoscopic tubal ligation, hysteroscopic tubal occlusion, bilateral salpingectomy, or insertion of the LNG-IUC, or were selected as a control comparison.

Risk of bias within studies
We excluded studies where risk of bias was determined to be medium to high (NOS 0-6) in at least one domain of assessment of risk, largely due to the observational study designs, non-random allocation of interventions, and differences in baseline characteristics between comparator groups. Our assessment of risk of bias for each study can be found in Additional File 2.

Results of individual studies
Results of individual studies can be found in Table 1.

Effectiveness
Six studies reported the rate of pregnancy, all of which were cohort studies comparing hysteroscopic tubal occlusion and laparoscopic tubal ligation. [24][25][26][27][28][29] Among the included studies, there was a wide range of follow-up time to assess effectiveness (from one year to a maximum of ten years) 24 and signi cant variance in directionality and strength of the outcome; this likely explains the considerable heterogeneity observed.
Three analyses found no signi cant difference in the risk or reported number of unintended pregnancies between laparoscopic tubal ligation and hysteroscopic tubal occlusion. 24,27,28 A retrospective cohort in the United States found that the cumulative rate of pregnancy was 1.02 pregnancies per 100 person years for hysteroscopic tubal occlusion and 0.88 pregnancies per 100 person years for laparoscopic tubal ligation (p = 0.003). Patients who underwent hysteroscopic tubal occlusion were at 1.2 times higher risk of becoming pregnant compared to those who underwent laparoscopic tubal ligation [aHR = 1.20 (95% con dence interval = 1.09-1.33)]. 29 Two studies, both conducted in France, found a higher risk of pregnancy among laparoscopic tubal ligation than hysteroscopic tubal occlusion; 25

Adverse Effects
Sixteen studies assessed adverse effects using prospective/retrospective cohort designs. Results are organized by intervention: six compared hysteroscopic tubal occlusion and laparoscopic tubal ligation, 24, 25, 28-31 three compared laparoscopic tubal ligation and a control, 32-34 ve compared laparoscopic tubal ligation and bilateral salpingectomy, 11,[35][36][37]40 one compared hysteroscopic tubal occlusion, laparoscopic tubal ligation with controls, and bilateral salpingectomy with controls, 38 and one compared bilateral salpingectomy with laparoscopic tubal ligation and with matched controls. 39 Hysteroscopic tubal occlusion vs. Laparoscopic tubal ligationTwo studies 24,30 found no statistically signi cant difference in rates of adverse effects including abnormal uterine bleeding, pelvic pain, or opioid managed pain between the two interventions. One study found no difference between chronic pelvic pain, risk of hysterectomy, or abnormal uterine bleeding at 6 or 12 months post-procedure among women who underwent hysteroscopic tubal occlusion compared to laparoscopic tubal ligation, nding only a signi cantly lower risk of chronic pelvic pain and risk of hysterectomy among women undergoing hysteroscopic tubal occlusion at 24 months post-procedure. 31 Hysteroscopic tubal occlusion was associated with a lower risk of surgical complications than laparoscopic tubal ligation [aOR = 0.18 (0.14-0.23)] 25 and a lower risk of iatrogenic complications after surgery [OR = 0.35 (0.20, 0.61)]. 28 Two analyses found higher rates of gynecological complications with hysteroscopic tubal occlusion compared to laparoscopic tubal ligation, 25,29 with one analysis nding that menstrual dysfunction was signi cantly higher for people undergoing hysteroscopic tubal occlusion [aHR = 1.23 (1.20-1.27)]. 29 However, pelvic pain incidence was found to be signi cantly lower in hysteroscopic tubal occlusion patients compared to laparoscopic tubal ligation patients [21.0% compared with 25.6% at 2 years, aHR = 0.83 (0.80-0.85)]. 29 Laparoscopic tubal ligation vs. Comparison Compared to controls, included studies did not nd a statistically signi cant change in menstrual cycle after undergoing laparoscopic tubal ligation 32 nor any differences in blood LH and E2 levels, ovarian volume, or number of antral follicles on third day of cycle. 33 One study found an increase in risk for anal cancer among those who underwent laparoscopic tubal ligation compared to those who did not [RR = 1.34 (1.11-1.63)]; however, no associations between laparoscopic tubal ligation and risk of endometrium, breast, cervix, or colorectal cancers, nor all cancers combined, were signi cant. 34 Laparoscopic tubal ligation vs. Bilateral salpingectomy Three studies found that there was no signi cant difference when comparing risk of readmission, blood transfusion, or intraoperative complications between laparoscopic tubal ligation and bilateral salpingectomy, 11,35,36 nor any difference in post-procedure physician visits for surgical infection or complication. 37 No signi cant differences in complications were found when assessing both immediate (2.9% vs. 2.5%, p = 1.0) and short-term (within 30 days) adverse events (14.7% vs. bilateral salpingectomy: 4.9%, p = 0.51) among people undergoing laparoscopic tubal ligation and bilateral salpingectomy, respectively. 40 However, there was a higher risk among people who underwent bilateral salpingectomy who required prescription analgesic use after surgery compared to those who underwent laparoscopic tubal ligation [(aOR = 1.21 (1.14-1.29)]. 37 Hysteroscopic tubal occlusion vs. Laparoscopic tubal ligation vs. Bilateral salpingectomy vs. controlsOne study performed a retrospective cohort study using administrative data to assess risk of ectopic pregnancy among people who underwent surgical sterilizations including bilateral salpingectomy, laparoscopy with Filshie clip, minilaparotomy, laparotomy, and hysteroscopic tubal occlusion using Essure™ compared to an unspeci ed destruction or occlusion of fallopian tubes. 38 Hazard ratios for ectopic pregnancy did not remain signi cant for laparoscopy with Filshie clip, minilaparotomy, and laparotomy, and there were no ectopic pregnancies reported for bilateral salpingectomy nor hysteroscopic tubal occlusion with Essure TM . 38 Bilateral salpingectomy vs. Tubal ligation vs. Historical Controls When comparing bilateral salpingectomy and laparoscopic tubal ligation to historical controls, there was no signi cant difference when comparing estimated median blood loss between bilateral salpingectomy (5 ml) with laparoscopic tubal ligation current (7 ml, p = 0.18) or historical (10 ml, p = 0.31) controls. 39

Patient Recovery
Two studies comparing laparoscopic tubal ligation and bilateral salpingectomy found no signi cant difference in terms of length of hospital stay, although there was a wide variance in reported length between the two studies (median of 1.8 hours 36 to 1.31 days 35 ).
One study found that women who underwent hysteroscopic tubal occlusion required fewer sick days compared to women who underwent laparoscopic tubal ligation at one-year (.590 days vs. 6.50 days, p < 0.001) and at three years (28.3 vs. 32.3, p < 0.001). 25 Non-contraceptive bene ts Five studies measured non-contraceptive bene ts, primarily assessing preventative bene ts in reducing the risk of developing various cancers. Three compared bilateral salpingectomy or laparoscopic tubal ligation against controls, 41-43 one compared laparoscopic tubal ligation against controls, 34 and the last compared hysteroscopic tubal occlusion and laparoscopic tubal ligation directly. 44 Two were a case-control studies 42,43 while the rest were cohorts using large administrative databases. 34,41,44 When

Tolerability
Nine studies assessed the ability to perform the intended method without requiring other procedures to either x the procedure due to an unsuccessful rst attempt, or a secondary procedure to remove hardware (micro inserts). All studies compared the tolerability of hysteroscopic tubal occlusion vs. laparoscopic tubal ligation with a wide variation in follow-up time, from an average of thirty days 24 to seven years. 44 Six studies found that there was a signi cantly higher risk of re-operation among those who underwent hysteroscopic tubal occlusion compared to laparoscopic tubal ligation, 24,25,[27][28][29]44 with studies reporting an increased adjusted hazard ratio from 2.05, 29 to 3.26 (1 year post-procedure), 25 and reporting odds of re-operation as high as 10-fold the odds at one year post-procedure. 28 At three years follow up, one study found that the increased risk of re-operation remained (aHR = 1.62, 1.51-1.73). 25

Accessibility
We did not nd an eligible study that systematically measured or compared the out-of-pocket costs for the procedure, wait times, or the locations where the procedure can be performed.

Eligibility
We did not nd an eligible study that systematically measured or compared eligibility for the procedures.

Follow-up required
We did not nd an eligible study that compared the number of follow-up visits needed, or required, to ensure that the method was completed or for safety monitoring.
Costs to the healthcare system Four studies measured costs to the healthcare system by index cost. All studies compared total costs between hysteroscopic tubal occlusion and laparoscopic tubal ligation, either reporting the mean or the median index costs per patient when undergoing each procedure. Three studies found that hysteroscopic tubal occlusion was less costly to perform than laparoscopic tubal ligation, [45][46][47] with costs for hysteroscopic tubal occlusion ranging between $1374 46 and $3964. 45 Costs for laparoscopic tubal ligation ranging between $2264 47 and $5163. 45 One study found that total medical and prescription costs ($7093 vs. $7568, p < 0.0001) and procedure-related costs ($4971 vs. $5407, p < 0.0001) were lower among women who underwent hysteroscopic tubal occlusion compared to tubal ligation. 45 However, costs related to complications or failures were higher with hysteroscopic tubal occlusion compared to laparoscopic tubal ligation ($272 vs. $176). 46 One study found higher total charges for hysteroscopic tubal occlusion compared to laparoscopic tubal ligation (median $7832 vs. $5068, P < 0.01). 28

Length of the Procedure
Six studies compared the length of the procedure, with four comparing bilateral salpingectomy and laparoscopic tubal ligation, and two comparing laparoscopic tubal ligation and hysteroscopic tubal occlusion. Four studies found that the bilateral salpingectomy procedure took signi cantly longer than laparoscopic tubal ligation to complete (3-11 minutes longer). 11,35,36,40 Median/mean procedure times for bilateral salpingectomy ranged between 44 40  Two studies comparing laparoscopic tubal ligation and hysteroscopic tubal occlusion found that the laparoscopic tubal ligation procedure took signi cantly longer (means and medians ranging between 27-52 minutes) than hysteroscopic tubal occlusion (means and medians ranging between 18-36 minutes). 28,47 Risk of bias across studies Overall, the cumulative evidence presented remains at low to medium risk of bias due to the observational study design used in all cases. Due to the observational study designs used, we found that there were signi cant sociodemographic differences between comparator groups that were not able to be adjusted for. Some studies did not fully report their patient demographic, leading to questions about comparability. With high heterogeneity found, our interpretation of evidence must be balanced and cautious. Our conclusions focus on comparisons between laparoscopic tubal ligation and hysteroscopic tubal occlusion, and we described tentative conclusions with other comparisons.

Additional Analysis
We did not conduct any additional analyses.

Summary of Evidence
There is insu cient data to compare available options for people seeking female permanent contraception, especially comparing to the LNG-IUC. The majority of studies eligible for our review compared laparoscopic tubal ligation to hysteroscopic tubal occlusion using Essure™ micro inserts. The latter method is currently no longer available for use in some jurisdictions. [6][7][8] As trends for alternatives to laparoscopic tubal ligation continue to increase, more research is needed to compare the options available so that people seeking female permanent contraception and their healthcare providers are able to make an informed decision.
While our review found that the majority of studies comparing hysteroscopic tubal occlusion and laparoscopic tubal ligation did not nd a signi cant difference in effectiveness, it is important to note that for hysteroscopic tubal occlusion, success of effectiveness relied on participants using another form of contraception, or abstinence, before tubal occlusion could be con rmed with a hysterosalpingogram and correct bilateral placements of the micro inserts.
Other options, including laparoscopic tubal ligation, bilateral salpingectomy, and insertion of the LNG-IUC are successful at preventing pregnancy immediately.
Although studies that assessed the effectiveness of the LNG-IUC were not included, other reviews have demonstrated the high e cacy of this method, with a cumulative pregnancy rate of 0.5 per 100 users, 48 which appears comparable to laparoscopic tubal ligation. 49 Non-contraceptive bene ts primarily looked at protective effects against various types of cancers. While the magnitude of the protective effect differed between methods of permanent contraception, it appears that undergoing some form of tubal interruption -whether it be occlusion, ligation, or removal -has a protective effect against several types of gynecologic cancers. In separate reviews, the LNG-IUC is also suggested to have protective effects against gynecological cancers, 50 as well as menorrhagia, endometriosis, adenomyosis, and broids. 48 Longer term cohort studies will be required to effectively compare these protective effects among all available options for female permanent or long acting contraception.
All options for female permanent contraception included in our review had risks of adverse effects; however, we did not nd signi cant differences in opioid managed pain, pelvic pain, menstrual dysfunction, or intraoperative complications when comparing surgical methods and/or controls. Notably no studies comparing pain included patients selecting the LNG-IUC. One study found that hysteroscopic tubal occlusion patients had a lower risk of post-procedure hysterectomy 24-months post-procedure, 31 but the strength of the evidence is diminished with potential bias in funding from Bayer, the company that created the Essure™ device. Multiple studies have found an association between an increased risk of anal cancer and a history of laparoscopic tubal ligation. 34,51 Although we did not nd articles assessing adverse events of the LNG-IUC to include in our review, previous studies found minimal adverse effects, with some attributed to the device itself such as dysmenorrhea or irregular bleeding or to the levonorgestrel such as weight gain; 48 although there is con icting evidence with weight gain due to levonorgestrel. 52 The more serious complications, such as uterine perforation, are found to occur rarely (estimated 2.6 per 1000 insertions). 53 Patient recovery was only assessed by length of hospital stay between salpingectomy and tubal ligation, where no signi cant difference was found.
Considerations including patient satisfaction, time to return to work, and other factors need to be explored further. Separate reviews have found that those who undergo the insertion of the LNG-IUC have high patient satisfaction and an almost immediate recovery time. [54][55][56] Lastly, we considered costs to the healthcare system. While three out of four studies assessing hysteroscopic tubal occlusion and laparoscopic tubal ligation found that hysteroscopic tubal occlusion was signi cantly less costly to perform than laparoscopic tubal ligation, 45-47 costs related to complications or failures were higher after hysteroscopic tubal ligation. 46 Costs to the healthcare system also should balance preventative costs, such as savings per life-year gained with prevention of cancer cases. No studies calculated preventative cost-savings accompanying non-contraceptive bene ts in each method, despite evidence that laparoscopic tubal ligation 34,[41][42][43][44] and the LNG-IUC 57 provide potential cancer risk reduction. Three Markov models predicted a theoretically signi cant cost-effectiveness when bilateral salpingectomy is employed in place of laparoscopic tubal ligation in terms of ovarian cancer prevention and lifeyears gained. 14,15,58 Actual patient cohort studies will be needed to determine if this savings is realized in practice.

Limitations
We found signi cant heterogeneity between the included articles. This high heterogeneity is likely driven by their methodological diversity and observational study designs, which did not allow for randomized allocation of participants. Therefore, signi cant differences in study population likely existed between the different intervention types, such as age, socioeconomic status, or underlying health conditions that were not excluded as a major comorbidity. Additionally, we found that follow-up times varied widely between the included studies, with some only allotting minimal weeks for follow-up time, which biases individual studies by not allowing for an accurate assessment of possible outcomes. Additionally, outcomes may thus be attributable to baseline differences and although associated with the interventions, may not necessarily be causally linked to the interventions.
Secondly, due to high heterogeneity, we were not able to complete a network meta-analysis and instead performed a narrative synthesis of results. Limitations to narrative synthesis include the potential biasing of results by overemphasizing the outcomes of particular studies, and the inability to objectively compare the different options available.
Thirdly, results primarily focused on ndings comparing laparoscopic tubal ligation and/or hysteroscopic tubal occlusion, with 10 out of 25 studies assessing bilateral salpingectomy and no articles assessing the LNG-IUC. Therefore, for bilateral salpingectomy results of studies may be overemphasized as our outcomes are based on less available evidence.

Conclusions
Previous meta-analyses have compared outcomes between hysteroscopic tubal occlusion and laparoscopic tubal ligation, yet few have considered other options available for permanent contraception. Several studies illustrate an increasing trend of using bilateral salpingectomy for permanent contraception, and there are many calls to consider LNG-IUC as a permanent contraception alternative, as it offers similar effectiveness rates with fewer adverse effects. High quality studies that compare traditional forms of permanent contraception, such as the laparoscopic tubal ligation, with new and alternative methods, are urgently needed to provide evidence on the implications for all options available to people seeking permanent female contraception.