An open letter to the FDA regarding the use of morcellation procedures for women having surgery for presumed uterine fibroids

In November 2014, the FDA ruled that power morcellation was contraindicated in "the majority of women" having surgery for uterine fibroids due to the potential risk of spreading occult uterine sarcoma.¹ Although problems with this ruling were immediately apparent, the passage of time has allowed for more clarity on the related medical issues.

Prevalence of leiomyosarcoma among women having surgery for presumed uterine fibroids
The prevalence of occult leiomyosarcoma among women with fibroids is critical for every patient. All medical procedures have potential risk and the patient's understanding of risk is the foundation of medical decision making.

The FDA estimated that for every 458 women having surgery for fibroids, one woman would be found to have an occult leiomyosarcoma (LMS). We challenge this calculation. To estimate this risk, the FDA searched medical databases using the terms “uterine cancer” AND “hysterectomy or myomectomy.” Because “uterine cancer” was required, studies where cancer was not found or discussed were not identified. Nine studies, all but one of which were retrospective, were analyzed including a non–peer-reviewed Letter to the Editor and an abstract from an unpublished study.^2,3

Additionally, 3 leiomyosarcoma cases identified by the FDA do not meet current pathologic criteria for cancer and would now be classified as benign "atypical" leiomyomas. If atypical leiomyomas and non–peer-reviewed data are excluded, the FDA identified 8 cases of LMS among 12,402 women having surgery for presumed leiomyomas, a prevalence of 1 in 1,550 (0.064%).

Pritts and colleagues⁴ recently published a more rigorous meta-analysis of 133 studies and determined that the prevalence of LMS among women having surgery for presumed fibroids was 1 in 1,960, or 0.051%. All peer-reviewed reports in which surgery was performed for presumed fibroids were analyzed, including reports where cancer was not found. Inclusion criteria required that histopathology results be explicitly provided and available for interpretation. Among the 26 randomized controlled trials analyzed, 1,582 women had surgery for fibroids and none were found to have LMS.

Bojahr and colleagues⁵ recently published a large population-based prospective registry study and reported 2 occult LMS among 8,720 women having surgery for fibroids (0.023%).

In summary: The re-analyzed FDA dataset yields a prevalence of 1 in 1,550 (0.064%); the Pritts study reports a prevalence of 1 in 1,960 (0.051%), with the RCTs having a prevalence of 0; and the Bojahr study reports a prevalence of 2 of 8,720 (0.023%). We acknowledge that with rare events statistical analysis may be uncertain and confidence intervals may be wide. However, these numbers do not support the FDA's estimated prevalence of LMS among women having surgery for presumed fibroids and those at risk for morcellation of an LMS.

Prognosis for women with morcellated LMS
Women with LMS, removed intact without morcellation, have a poor prognosis. Based on SEER data, the 5-year survival of stage I and II LMS is only 61%.⁶ Whether morcellation influences the prognosis of women with LMS is not known, and the biology of this tumor has not been well studied. Distant metastases occur early in the disease process, primarily hematogenous dissemination. Four frequently quoted published studies examine survival following power morcellation. Surprisingly, virtually none of the women in these studies had power morcellation. Furthermore, the data presented in these reports are poorly analyzed and patient numbers are very small.

Park and colleagues⁷ reported only one of the 25 morcellated cases had laparoscopic surgery with power morcellation. Eighteen women had a laparoscopically-assisted vaginal hysterectomy with scalpel morcellation performed through the vagina, one had a vaginal hysterectomy with scalpel-morcellation, and 5 had mini-laparotomy with scalpel morcellation through small lower abdominal incisions. Seventeen of the 25 patients plotted in the published survival curve were referred to the hospital after initial diagnosis or the discovery of a recurrence at another institution. Since the number of nonreferred women with less aggressive disease or without recurrence is not known, it is not possible to determine differences in survival between patients with and without morcellation.

In a study by Perri and colleagues,⁸ none of the patients had power morcellation: 4 women had abdominal myomectomy; 4 had hysteroscopic myomectomy with tissue confined within the uterine cavity; 2 had laparoscopic hysterectomy with scalpel morcellation; 4 had supracervical abdominal hysterectomy with cut-through at the cervix; and 2 had abdominal hysterectomy with injury to the uterus with a sharp instrument.

When comparing the outcomes for women with morcellated and nonmorcellated LMS, Morice and colleagues⁹ found no difference in recurrence rates or overall and disease-free survival at 6 months.

In the only study to compare use of power with scalpel morcellation in women with LMS, Oduyebo and colleagues¹⁰ found no difference in outcomes for the 10 women with power morcellation and 5 with scalpel morcellation followed for a median of 27 months (range, 2–93 months). Notably, a life table analysis of the above studies showed no difference in survival between morcellation methods.¹¹