Savage, Ulrike K. MD; Masters, Steven J. MD; Smid, Marcela C. MD; Hung, Yun-Yi PhD; Jacobson, Gavin F. MD
Sterilization procedures are the most common form of contraception in the United States.1 There are several transabdominal methods for tubal occlusion, and although efficacy varies by method, the U.S. Collaborative Review of Sterilization (CREST) study found that the overall 10-year probability of pregnancy was approximately 18.5 per 1,000 procedures.2 Any new sterilization technique should be compared with these well-established and highly effective methods.
In November 2002, the U.S. Food and Drug Administration approved a new method of interval tubal sterilization using hysteroscopically placed microinserts. The Essure Permanent Birth Control Device (Conceptus Inc., San Carlos, CA) involves transcervical insertion of a microcoil that expands to lock into place, and over time tissue ingrowth causes tubal occlusion. Alternative contraception is necessary until a hysterosalpingogram performed 3 months after microinsert placement demonstrates satisfactory device placement and tubal occlusion.
Phase 2 and 3 clinical trials assessed the safety and efficacy of the new device.3,4 These vendor-sponsored trials reported 88–92% placement rates and 92–96% bilateral occlusion rates by hysterosalpingogram at 3 months. Subsequent descriptive studies that use hysterosalpingography for follow-up have reported occlusion rates of at least 98%; however, most of these seem to be performed by or under the direct guidance of a few highly experienced providers.5–9 Kaiser Permanente Northern California includes multiple practice sites and a large number of physicians, and our experience may more accurately reflect real-life use and effectiveness of this method when first introduced into a clinical practice. We sought to estimate device placement and tubal occlusion rates for hysteroscopic sterilization since its adoption at Kaiser Permanente Northern California and identify risk factors for failure.
MATERIALS AND METHODS
This study was conducted at Kaiser Permanente Northern California, a prepaid group-model integrated health care delivery system that provides comprehensive medical services to approximately 30% of the Northern California population and is demographically representative of that population. Hysteroscopic microinsert sterilization was introduced at Kaiser Permanente Northern California in 2004 but not widely adopted until 2005. Women who underwent this procedure from January 2004 to December 2006 were identified by procedure coding in our electronic medical records system. Procedures performed in the clinic were considered outpatient, and those performed in the hospital were considered inpatient. For procedures occurring in this time frame, follow-up was assessed through December 31, 2008. Standards of practice at Kaiser Permanente Northern California generally included timing procedures with suppression by hormonal contraception or during the follicular phase of the menstrual cycle. Demographic data were collected by review of electronic medical records and included patient age, parity, and body mass index (BMI). The primary surgeons for procedures were only attending physicians, not fellows or residents. To assess experience with the technique, we determined the chronologic order of procedures for each physician and calculated the success rate for each insertion order by sequence (eg, success rate for all first insertions, second insertions, third insertions, and so forth). For patients who had more than one procedure attempt, each procedure was treated as a separate experience. For procedures where outcome was not known, we counted the procedure toward an individual provider’s experience, but it was not included in calculation of success rate.
Placement was assessed from review of electronic notes, and in cases where none were available, we assumed that performance of a follow-up hysterosalpingogram was indicative that the physician judged the placement as successful. Hysterosalpingograms were performed per local facility protocol and interpreted by radiologists at each facility, and electronically stored interpretations were reviewed. For women who underwent more than one hysterosalpingogram, the final hysterosalpingogram result was used for analysis. Statistical analysis was performed with SAS 9.1 (SAS Institute Inc., Cary, NC). For univariable analyses to identify factors predictive of successful placement and occlusion, we used the χ2 and Fisher exact test where appropriate for categorical variables and the Student t test for continuous variables. Changes over time were tested by the Cochran-Armitage test for linear trend.10 The Kaiser Permanente Northern California Institutional Review Board approved this study.
We identified 884 women who underwent 897 hysteroscopic microinsert sterilization procedures (13 underwent two attempts) at Kaiser Permanente Northern California from April 2004 to December 2006. Procedures were performed at 30 facilities by 118 physicians. The initial placement attempt was successful in 850 (96.2%), failed in 31 (3.5%), and could not be determined in three (0.3%). There were no statistically significant differences in mean age, nulliparity, or mean BMI between women who had successful device placement compared with those who did not (Table 1). The rate of successful placement was significantly higher in the outpatient setting than in the inpatient setting (97.3% compared with 92.8%, P=.004). There was no statistically significant differences in age, BMI, or nulliparity between women who had device placement in the outpatient compared with the inpatient setting (data not shown).
Of the original cohort of 884, four (0.4%) had failed hysterosalpingogram attempts, in 26 (2.9%) hysterosalpingography was not indicated primarily because of failed insertion, and 115 (13.0%) were lost to follow-up. There were 739 women who returned for at least one hysterosalpingogram after their final hysteroscopic placement attempt, including 54 who had two hysterosalpingograms and 4 who had three hysterosalpingograms. Bilateral tubal occlusion was reported in 687 of 739 (93.0%) who had hysterosalpingography completed. There were no statistically significant differences in age, nulliparity, BMI, or operative location between women who had bilateral occlusion compared with those who did not (Table 2). Women who did not have bilateral occlusion had a higher mean number of hysterosalpingographic procedures (P<.001) and longer intervals from procedure to hysterosalpingogram (P<.001).
One hundred eighteen physicians performed from 1 to 54 procedures during the study period; however, the number of procedures per physician quickly decreased as sequence increased: 57 physicians performed at least five procedures, 26 performed at least 10 procedures, 20 performed at least 15 procedures, 10 performed at least 20 procedures, five performed at least 29 procedures, and four or fewer performed at least 30 procedures. Trend analysis was performed for occlusion rate by sequence for which at least five physicians performed procedures (through the first 29 procedures). Figure 1 shows no significant change in the bilateral tubal occlusion rate by increasing experience (P for trend=.6). There was no significant change in the bilateral occlusion rate by increasing experience when sequence data were analyzed by groups of two (P for trend=.87), three (P for trend=.75), four (P for trend=.87), five (P for trend=.98), or 10 (P for trend=.68). The last failed occlusion by sequence was the 42nd procedure.
Figure 1 also shows the percent lost to follow-up by sequence, which ranged from 5.6% to 33.3%. Overall, 13% were lost to follow-up before a hysterosalpingogram could be obtained, with the 11.2% loss rate through the first 10 procedures significantly lower than the 18.5% loss rate in subsequent procedures (P=.003). Of the 115 lost to follow-up, 38.3% were not Kaiser Permanente Northern California members at the end of the study period.
Thirteen patients had known prior unilateral salpingectomy, and one patient had known congenital absence of one fallopian tube. Thirteen had successful device placement, and all 13 had successful bilateral occlusion. We identified eight women who had a pregnancy after initial placement: one never returned for hysterosalpingography, four had a hysterosalpingogram with at least one patent tube, and three had the hysterosalpingogram interpreted as bilaterally occluded. Among women with a hysterosalpingogram showing tubal patency, 42 had subsequent laparoscopic sterilizations, and two partners chose vasectomy. At least three women with successful occlusion had return visits to discuss reversal of sterilization.
Our successful device placement rate, 96.2%, is comparable to rates ranging from 88–96% reported in the literature.3–9 Our 93.0% tubal occlusion rate is comparable to that in published controlled trials but lower than that in smaller series reports.3,4,6,8 The discrepancy between initial device placement and subsequent bilateral tubal occlusion rates highlights the need for hysterosalpingographic follow-up. Our assessment of hysterosalpingograms focused strictly on occlusion and did not address confirmation of placement because this was inconsistently reported on hysterosalpingographic reports. Among the three women who conceived after the hysterosalpingogram was read as occluded, subsequent internal review suggested that none were properly placed. It is possible that our occlusion rate was artificially increased by inclusion of improperly placed devices.
Approximately 13% of women did not have appropriate hysterosalpingographic follow-up. Reasons for incomplete follow-up may reflect patient reluctance to undergo another invasive procedure, inadequate counseling, or an inadequate system of tracking hysterosalpingographic follow-up. Vendor-sponsored studies have more complete follow-up, whereas others have 6–27% lost to follow-up.3,4,8,11 Overall, only 77.7% of all women in our study group had bilateral occlusion on hysterosalpingogram, similar to rates that can be extrapolated from the literature of 58–87%.3,4,6,8,11 These lower rates include both placement failures and loss to follow-up and exemplify how significantly these problems plague most centers.
Few data directly address appropriate training to competency. The formal training process requires completion of hands-on training with a vendor-designated preceptor “until competency is established” and completion of initial procedures under observation of vendor-designated preceptor “until competency is established.”12 Although competency is not defined, before enrolling for training, it is suggested that two or three patients be identified and scheduled for the procedure. The average number of procedures per physician was 45 in the phase 2 trial and 25 in the phase 3 trial.3,4 In the phase 2 trial, Kerin et al3 noted that the five providers involved became more confident and time-efficient with the device over the first five procedures; however, they refer to unpublished data to support this. In the phase 3 trial, Cooper et al4 reported that bilateral placement rates did not improve substantially in relation to experience and that procedure time decreased over the first five cases, but no data are provided to evaluate these conclusions. The Food and Drug Administration–mandated U.S. postapproval study included 31 physicians without prior experience who averaged 10 procedures each and reported a 90% placement rate.13 However, all of these physicians not only had completed didactic and model training, but also had completed their preceptorship before contributing, and no occlusion data were presented. Although time to complete procedure is no measure of competency, studies refer to improvement in time over the first 5–13 procedures.3,4,8 Of note, we found no significant increase in occlusion rates when trend analysis was performed in sequences of 5 or 10.
Reports of case series tend to involve small numbers of experienced providers performing or directly supervising larger numbers of procedures, and only cumulative success rates are reported.6,9 Rosen5 reported on a learning curve for the first 80 cases and suggested that surgical time was reduced with each 25 procedures. His first failure in this series was the 14th procedure. However, these cases were collected after the approval of the device in Australia and do not seem to account for the procedures performed by the author during the previous phase 3 trial, and no occlusion data are presented.4,5 Sinha et al7 reported on 112 procedures by two physicians and reported that six of nine failures were in the first 14. Levie et al8 reported on 102 procedures all performed by or under the direct supervision of a single experienced physician. They reported the greatest improvement in time after the first 13 procedures.
Local standards of practice among Kaiser Permanente Northern California physicians varied, after an initial training most observed from one to four procedures before inserting their first device, and most were proctored for their first two to six procedures. This system seemed to be successful with regard to an initial high device placement rate and seemed to translate to high initial occlusion rates. The proctoring practices could explain the high initial occlusion rates, which remained unchanged by trend analysis. It is surprising that the occlusion rate did not increase with experience; however, the high loss to follow-up creates a wide margin of error for each point. This negative finding could be related to our high initial occlusion rate, which exceeded 95% through the first three procedures; to show a statistically significant increase in this rate by trend analysis would require an exceptionally large cohort. The significance of an occlusion rate greater than 95% may not be clinically meaningful. Additionally, we are unable to explain the higher placement rate when the procedure was performed in the outpatient compared with the inpatient setting. It is possible that other variables not measured prompted operators to perform procedures they assessed as more difficult in the hospital.
Strengths of this study include the large number of procedures performed by a large number of different physicians at different facilities within our health care system. We believe our real-life experience will be valuable to any physician, physician group, or health care organization that is considering offering hysteroscopic sterilization. A limitation of our study is the large group lost to follow-up. We found no significant trend of occlusion rate as experience increases; however, given the low number of providers performing a larger volume of procedures, we cannot exclude a type 2 error in the trend analysis. Additional limitations are the limited data on specific procedural techniques and procedural sedation protocols used at the different facilities and the limited demographic variables available for analysis.
Completing the didactics, model training, and short preceptorship lead to a placement rate of 96% and subsequent occlusion rate of 93%, which seems lower than comparable contraceptive options, such as laparoscopic sterilization or intrauterine device. Although competency can be difficult to define, our findings suggest that after the initial training and observation period, consistently high occlusion rates can be achieved quickly. Patient counseling regarding success of the procedure should include estimates of failure that include inability to insert the devices bilaterally as well as those subsequently found to be improperly placed. Attention must also be given to the number who may need two procedures, two or more hysterosalpingograms, or both, possibly requiring up to 9 months of alternate contraception use before being able to rely on the this method. Most importantly, a multicheck system to encourage appropriate follow-up is needed.
1. Mosher WD, Martinez GM, Chandra A, Abma JC, Willson SJ. Use of contraception and use of family planning services in the United States: 1982–2002. Adv Data 2004;350:1–36.
2. Peterson HB, Xia Z, Hughes JM, Wilcox LS, Tylor LR, Trussell J. The risk of pregnancy after tubal sterilization: findings from the U.S. Collaborative Review of Sterilization. Am J Obstet Gynecol 1996;174:1161–70.
3. Kerin JF, Cooper JM, Price T, Herendael BJV, Cayuela-Font E, Cher D, et al. Hysteroscopic sterilization using micro-insert device: results of a multicentre phase II study. Hum Reprod 2003;18:1223–30.
4. Cooper JM, Carignan CS, Cher D, Kerin JF; Selective Tubal Occlusion Procedure 2000 Investigators Group. Microinsert nonincisional hysteroscopic sterilization. Obstet Gynecol 2003;102:59–67.
5. Rosen DMB. Learning curve for hysteroscopic sterilisation: lessons from the first 80 cases. Aust N Z J Obstet Gynaecol 2004;44:62–4.
6. Chern B, Siow A. Initial Asian experience in hysteroscopic sterilization using the Essure permanent birth control device. BJOG 2005;112:1322–7.
7. Sinha D, Kalathy V, Gupta JK, Clark TJ. The feasibility, success and patient satisfaction associated with outpatient hysteroscopic sterilization. BJOG 2007;114:676–83.
8. Levie MD, Chudnoff SG. Prospective analysis of office-based hysteroscopic sterilization. J Minim Invasive Gynecol 2006;13:98–101.
9. Mino M, Arjona JE, Cordon J, Pelegrin B, Povedano B, Chacon E. Success rate and patient satisfaction with the Essure sterilisation in an outpatient setting: a prospective study of 857 women. BJOG 2007;114:763–6.
10. Armitage P. Tests for linear trends in proportions and frequencies. Biometrics 1955;11:375–86.
11. Duffy S, Marsh F, Rogerson L, Hudson H, Cooper K, Jack S, et al. Female sterilisation: a cohort controlled comparative study of Essure versus laparoscopic sterilisation. BJOG 2005;112:1522–8.
13. Nichols M, Carter JF, Fylstra DL, Childers M; Essure System U.S. Post-Approval Study Group. A comparative study of hysteroscopic sterilization performed in-office versus a hospital operating room. J Minim Invasive Gynecol 2006;13:447–50.