Use of Electric Power Morcellation and Prevalence of Underlying Cancer in Women Who Undergo Myomectomy

Importance  Myomectomy, the excision of uterine leiomyoma, is now commonly performed via minimally invasive surgery. Electric power morcellation, or fragmentation of the leiomyoma with a mechanical device, may be used to facilitate extraction of the leiomyoma.

Objective  To analyze the prevalence of underlying cancer and precancerous changes in women who underwent myomectomy with and without electric power uterine morcellation.

Design, Setting, and Participants  We used a US nationwide database to retrospectively analyze women who underwent myomectomy at 496 hospitals from January 2006 to December 2012. Use of electric power morcellation at the time of myomectomy was investigated. The prevalence of uterine cancer, uterine neoplasms of uncertain malignant potential, and endometrial hyperplasia were estimated. Multivariable mixed-effects regression models were developed to examine predictors of use of electric power morcellation and factors associated with adverse pathologic outcomes.

Main Outcomes and Measures  Use of electric power morcellation at the time of myomectomy was examined. The occurrence of uterine cancer and precancerous uterine lesions was determined.

Results  The cohort consisted of 41 777 women who underwent myomectomy at 496 hospitals and included 3220 (7.7%) who had electric power morcellation. Uterine cancer was identified in 73 (1 in 528) women who underwent myomectomy without electric power morcellation (0.19%; 95% CI, 0.15%-0.23%) and in 3 (1 in 1073) women who underwent electric power morcellation (0.09%; 95% CI, 0.02%-0.27%). The corresponding rates of any pathologic finding (cancer, tumors of uncertain malignant potential, or endometrial hyperplasia) were 0.67% (n = 257) (95% CI, 0.59%-0.75%) (1 in 150) and 0.43% (n = 14) (95% CI, 0.21%-0.66%) (1 in 230), respectively. Advanced age was the strongest risk factor for uterine cancer.

Conclusions and Relevance  The prevalence of cancers and precancerous abnormalities of the uterus in women who undergo myomectomy with or without electric power morcellation is low overall, but risk increases with age. Electric power morcellation should be used with caution in older women undergoing myomectomy.

Uterine leiomyomas are benign smooth-muscle tumors of the uterus. These neoplasms are common and are thought to occur in 20% to 50% of women.^1-5 While leiomyomas are frequently asymptomatic, they may cause pain, menstrual bleeding and dysfunction, and reduced fertility.^1-6

First-line therapy for patients with symptomatic leiomyomata typically consists of medical management. A variety of hormonal modulating agents, including progestins, oral contraceptives, and gonadotropin-releasing hormone agonists, have demonstrated efficacy.^5-8 A number of nonsurgical procedures, including uterine artery embolization, are now also available as treatment options for some women with symptomatic leiomyoma.⁹ Definitive treatment is often surgical and consists of either hysterectomy or myomectomy in women who wish to preserve the uterus.^5,6,9

Myomectomy can be performed hysteroscopically, abdominally through a laparotomy, and, more recently, via a minimally invasive surgical approach with laparoscopic or robotic assistance.^10-12 Removal of large leiomyoma through the small incisions used for minimally invasive myomectomy often poses a challenge. Large leiomyoma can be removed through a small abdominal incision (minilaparotomy), vaginally by colpotomy, or through use of electric power morcellation to fragment the leiomyoma.¹³

Recently, the use of electric power morcellators has been subject to increased scrutiny after a patient with a presumed leiomyoma underwent hysterectomy with electric power morcellation and was noted to have a uterine sarcoma that was disseminated after the procedure.¹⁴ This case has led to increased recognition that while leiomyomas are most commonly benign, these neoplasms can be associated with unrecognized cancers. Use of electric power morcellation at the time of uterine surgery, either myomectomy or hysterectomy, disrupts the underlying malignant neoplasm and may potentiate tumor dissemination.^15,16 This case has led to reevaluation of the safety of electric power morcellators for gynecologic surgery.^17,18

While the controversy surrounding electric power morcellation predominantly stems from use of the device for hysterectomy, morcellators are also increasingly used for the performance of myomectomy. To date, there are limited data describing the safety or risk of pathologic abnormalities in women undergoing myomectomy with electric power morcellation. We performed a population-based analysis to determine the patterns of use of electric power morcellation at the time of myomectomy and to examine the prevalence and predictors of pathologic abnormalities in women who underwent myomectomy both with and without electric power morcellation.

Box Section Ref ID

Patient written informed consent and ethical approval for this retrospective analysis were waived by the Columbia University institutional review board.

We analyzed data from the Perspective database (Premier Inc), a large, all-payer database that includes hospitals from across the United States. Initially developed to measure resource utilization and quality, the Perspective database collects comprehensive data from over 500 acute care facilities, including clinical and demographic characteristics, diagnoses and procedures, and all billed services rendered during hospitalization. In 2006, Perspective captured approximately 5.5 million hospital discharges, representing approximately 15% of hospitalizations in the United States.¹⁹ The data undergo a rigorous, multistep quality-control process and have been utilized in a number of outcomes studies.^19-21

Our cohort consisted of women who underwent inpatient or outpatient myomectomy from January 2006 to December 2012. Patients with an International Classification of Diseases, Ninth Revision (ICD-9) code of 68.29 (excision or destruction of lesion of the uterus) in combination with a code for uterine leiomyoma (218.x) were identified. We excluded patients who underwent endometrial polypectomy (endometrial polyp 621.0, 622.7, 219.0) and those who underwent a hysteroscopic procedure (68.12) from the cohort. Patients who had concomitant ICD-9 codes for an abdominal, vaginal, laparoscopic, radical, or unspecified hysterectomy were also excluded. To capture uterine morcellation, we searched each patient’s recorded charge codes for use of any commercially available electric power morcellation device.²² Within the Perspective database, all drugs and devices used during a patient’s hospitalization are recorded. This method has previously been used and validated to capture drug and device use in patients recorded in the database.^20,21,23-26 Concurrent use of robotic assistance for the hysterectomy was recorded using ICD-9 procedure codes and charge codes as previously described.²¹

Demographic data including year of the procedure, patient age at the time of the procedure (<40, 40-49, 50-59, ≥60 years), marital status (married, single, or unknown), and insurance status (commercial, Medicare, Medicaid, uninsured, unknown) were recorded for each patient. Classification of race was based on patient self-report and grouped as white, black, and other. Risk adjustment for comorbid medical diseases was performed using the Elixhauser comorbidity index. Patients were classified by number of medical comorbidities into the following groups: 0, 1, and 2 or more.²⁷

Procedure hospitals were categorized by location (metropolitan or nonmetropolitan), region of the country (Northeast, Midwest, West, or South), size (<400 beds, 400-600 beds, or >600 beds), and teaching status (teaching or nonteaching). Hospital procedural volume was calculated in each case and estimated as the annualized procedural volume. Volume estimates were visually inspected and divided into 3 approximately equal volume-based tertiles: low (<24.7 procedures per year), intermediate (24.7-63.4 procedures per year), and high (>63.4 procedures per year).^28-30

The outcomes of the analyses were the occurrence of cancerous and precancerous abnormalities of the uterus. The primary outcome was the occurrence of uterine corpus cancer (ICD-9 codes 179.x and 182.x). We also quantified the occurrence of precancerous abnormalities and pathologic findings of undetermined long-term significance. All pathologic abnormalities were captured from review of coded claims at the time of the myomectomy.

Endometrial hyperplasia results from the aberrant growth of the endometrial glandular tissue that is associated with concurrent endometrial cancer and also poses a risk of progression to cancer.^31-33 Women with endometrial hyperplasia (ICD-9 codes 621.3, 621.30, 621.31, 621.32, and 621.33) were analyzed. Uterine neoplasms of uncertain behavior including, smooth-muscle tumors of uncertain malignant potential and atypical leiomyomas (ICD-9 code 236.0), were also examined. In addition, a composite outcome of any pathologic abnormality that consisted of uterine cancer, uterine neoplasms of uncertain behavior, and endometrial hyperplasia were analyzed.

Frequency distributions between categorical variables by use of electric power morcellation at the time of myomectomy were compared using χ² tests. To describe the prevalence of adverse pathologic outcomes, we report the unadjusted rate of each abnormality and the number of cases per 100 women. Data are presented for all of the outcomes and the composite of any pathologic abnormality.

The association between the clinical and demographic characteristics and use of electric power morcellation was examined using multivariable mixed effects log-linear regression models. To account for hospital-level clustering, these models included a random intercept for the hospital in which the procedure was performed. Similar models were developed to describe the association between the clinical characteristics and adverse pathologic findings. These models included use of electric power morcellation. Associations are reported as prevalence ratios with 95% CIs. All analyses were performed with SAS software, version 9.4 (SAS Institute Inc). All statistical tests were 2 sided, and P< .05 was considered statistically significant.

A total of 41 777 women who underwent myomectomy at 496 hospitals were identified. The cohort included 3220 (7.7%) patients who underwent myomectomy with use of electric power morcellation and 38 557 (92.3%) who underwent the procedure without electric power morcellation. Use of electric power morcellation increased each year from 2.6% in 2006 to 14.0% of cases in 2012 (P < .001) (Figure, A).

Table 1 lists the descriptive characteristics of the cohort. Among women who underwent electric power morcellation, 1988 (61.7%) were younger than 40 years; 976 (30.3%) were aged 40 to 49 years; 207 (6.4%) were aged 50 to 59 years; and 49 (1.5%) were 60 years or older (P = .01). Black women accounted for 26.7% of those who underwent morcellation (P < .001); 89.2% of all patients were commercially insured (P < .001). Electric power morcellation was more commonly used at intermediate- and higher-volume hospitals (P < .001).

In a multivariable model of factors associated with electric power morcellation, year of surgery was associated with morcellation (Table 2). Compared with women treated in 2006, the prevalence ratio (PR) for morcellation in 2012 was 2.12 (95% CI, 1.76-2.55). Black women (PR, 0.80; 95% CI, 0.72-0.88), uninsured patients (PR, 0.74; 95% CI, 0.57-0.95), and women with medical comorbidities (PR, 0.87; 95% CI, 0.77-0.98) were less likely to undergo electric power morcellation. Patients who underwent a robotic-assisted myomectomy were more likely to undergo power morcellation (PR, 6.86; 95% CI, 6.29-7.49), as were patients treated in the southern United States (PR, 1.62; 95% CI, 1.07-2.46). Patients treated at intermediate-volume hospitals were more likely to undergo electric power morcellation (PR, 1.52; 95% CI, 1.08-2.15).

Within the cohort, a total of 76 uterine cancers were identified (Table 3). The prevalence of uterine cancer was 0.19% (95% CI, 0.15%-0.23%) in women who underwent myomectomy without morcellation (1 in 528) and 0.09% (95% CI, 0.02%-0.27%) in those who had electric power morcellation (1 in 1073). Uterine tumors of uncertain malignant potential were identified in 0.08% (95% CI, 0.05%-0.11%) of patients who had myomectomy without electric power morcellation (1 in 1285) and 0.06% (95% CI, 0.01%-0.22%) of those with electric power morcellation (1 in 1610), while endometrial hyperplasia was seen in 0.41% (95% CI, 0.34%-0.47%) (1 in 246) and 0.31% (95% CI, 0.12%-0.50%) (1 in 322), respectively, of those without and with electric power morcellation. The composite end point of any pathologic outcome was found in 0.67% (95% CI, 0.59%-0.75%) of patients who did not undergo morcellation (1 in 150) and in 0.43% (95% CI, 0.21%-0.66%) of those who had electric power morcellation (1 in 230).

Among all women who underwent myomectomy, the prevalence of abnormal findings increased over time (Figure, B). Uterine cancer was noted in 0.13% (95% CI, 0.03%-0.22%) in 2006 and increased to 0.32% (95% CI, 0.18%-0.46%) in 2012 (P = .02). The prevalence of endometrial hyperplasia increased from 0.29% (95% CI, 0.15%-0.44%) in 2006 to 0.56% (95% CI, 0.38%-0.75%) in 2012 (P < .001). Similarly, the prevalence of any pathologic abnormality (uterine cancer, endometrial hyperplasia, or uterine neoplasms of uncertain malignant potential) rose from 0.57% (95% CI, 0.37%-0.77%) in 2006 to 0.93% (95% CI, 0.70%-1.16%) in 2012 (P < .001).

The prevalence of pathologic findings increased with age in women who underwent myomectomy with or without electric power morcellation (Table 4). Among women who had a myomectomy without morcellation, uterine cancer was seen in 0.05% (95% CI, 0.02%-0.07%) of women younger than 40 years and rose to 0.62% (95% CI, 0.30%-0.94%) in those aged 50 to 59 years and to 3.40% (95% CI, 2.22%-4.58%) in women 60 years or older. The prevalences of uterine cancer in women who had electric power morcellation was 0% (95% CI, 0%-0.19%) in those younger than 40 years, 0.97% (95% CI, 0.12%-3.45%) in patients aged 50 to 59 years, and 0% (95% CI, 0%-7.25%) in those 60 years or older. Similar trends were noted for endometrial hyperplasia and overall adverse pathologic findings.

In a multivariable model examining factors associated with uterine cancer, age was the strongest risk factor (Table 2). Compared with women younger than 40 years, the PR was 3.19 (95% CI, 1.50-6.80) in women aged 40 to 49 years, 14.86 (95% CI, 6.75-32.70) in those aged 50 to 59 years, and 54.33 (95% CI, 23.50-125.48) in women 60 years or older. Patients treated at nonmetropolitan hospitals were more frequently diagnosed with uterine cancer (PR, 2.81; 95% CI, 1.26-6.24). There was no statistically significant association between use of electric power morcellation and uterine cancer (PR, 0.66; 95% CI, 0.20-2.18). Similar findings were noted when any pathologic abnormality was examined as the end point.

The overall prevalence of uterine cancer was 0.09% in women who underwent myomectomy with electric power morcellation. These data have 2 important implications. First, the overall risk of malignant neoplasms associated with electric power morcellation at the time of myomectomy is lower than reports of the use of the device for hysterectomy.^15,16,22 Second, similar to hysterectomy, the risk of malignant neoplasms is significantly associated with age and is higher in older women.²² Reassuringly, the prevalence of pathologic abnormalities was lower in women who underwent electric power morcellation than in those who underwent the procedure without electric power morcellation.

Prior estimates of the risk of cancer in women undergoing myomectomy are largely lacking, and risks are often extrapolated from studies examining pathologic changes in women who underwent hysterectomy for leiomyoma. A series of over 1400 hysterectomy specimens noted leiomyosarcomas in 0.49%.³⁴ To date, few data have been reported to specifically examine pathologic outcomes among women who undergo myomectomy with the aid of electric power morcellation. Institutional series have noted sarcomas or smooth-muscle tumors of uncertain malignant potential in 1% to 1.2% of women who underwent hysterectomy with morcellation.^16,35 Epithelial endometrial tumors have also been found in women who underwent hysterectomy with morcellation.^22,36

We noted that the prevalence of pathologic abnormalities in women who underwent myomectomy with electric power morcellation was lower than previous estimates in women who underwent hysterectomy with the aid of power morcellation.²² A population-based analysis estimated the prevalence of uterine cancer at 0.27% in patients who underwent hysterectomy with morcellation (1 in 368).²² In contrast, we noted uterine cancer in 0.09% of those who underwent myomectomy with the assistance of electric power morcellation (1 in 1073). Myomectomy is generally a procedure performed in younger women who are typically considered to be at lower risk for invasive and preinvasive uterine disease. For those women who undergo hysterectomy with electric power morcellation and are ultimately found to have a uterine leiomyosarcoma, the risk of sarcomatous dissemination is substantial.^15,16,36,37 Little is known about the risk of possible tumor dissemination in women who undergo myomectomy either with or without morcellation. Further studies to examine long-term follow-up of these patients are warranted.

Age was one of the strongest risk factors for pathologic findings in our analysis. The role of myomectomy in perimenopausal and postmenopausal women remains controversial.^7,38 For women who have completed childbearing, hysterectomy is typically considered the surgical treatment of choice for leiomyoma given the risk of recurrence after myomectomy with uterine preservation. Furthermore, leiomyomas are often hormonally responsive and regress after menopause. These data have important implications for older women considering myomectomy and suggest that electric power morcellation should be approached with caution in older women.

While our study benefits from the inclusion of a large sample of women, we recognize a number of limitations. The most important limitation is the potential misclassification of pathologic abnormalities. Most hospitals report ICD-9 diagnosis codes for billing after completion of a surgical procedure once the operative findings are known. Although we were unable to verify our findings with review of pathologic records, it is unlikely that patients with high-grade precancers or cancers would intentionally undergo myomectomy or electric power morcellation; thus, any bias in our findings would likely be from undercapture of significant pathologic abnormalities and bias our findings toward the null hypothesis. Likewise, for women who undergo electric power morcellation, pathologic assessment of the tissue fragments is often difficult and may fail to diagnosis an underlying cancer, particularly if the lesion is small. As such, our estimates cannot capture tumors unrecognized at the time of surgery.³⁹ Furthermore, currently available ICD-9 codes are relatively nonspecific, and it is not possible to distinguish uterine sarcomas from endometrial tumors. However, given that electric power morcellation cannot be captured using ICD-9 coding alone, our data likely represent one of the few ways to assess morcellator-associated disease at the population level.

Second, use of electric power morcellators was determined through the use of hospital billing codes for the device. While use of billing codes from this database to capture the use of drugs and devices has been reported in a number of studies,^20,21,23-26 we cannot exclude the possibility that a small number of women were not captured. Although clinically uncommon, we cannot exclude the possibility that a small number of women intentionally underwent an excision of a known uterine cancer without hysterectomy. Currently available ICD-9 coding is inadequate to distinguish whether myomectomy was performed minimally invasively or via laparotomy; more detailed analysis of specific procedures would be of great value. While the a priori goal of our study was to examine the complications associated with electric power morcellators, some patients undoubtedly underwent manual morcellation procedures with sharp instrumentation that we were unable to capture. Finally, while prior work has failed to identify reliable factors predictive of cancer in women with presumed fibroids, unmeasured uterine factors may have influenced the use of electric power morcellation.

The increased scrutiny surrounding electric power morcellation has led to a number of proposals to improve safety, including more rigorous preoperative evaluation and thorough informed-consent guidelines.^13,14 Similarly, given the controversy surrounding electric power morcellation and the overall low prevalence of pathologic abnormalities, the development of national registries of patients who undergo the procedure would be of great utility. However, some have questioned whether these strategies go far enough to protect patients.¹⁴ A leading manufacturer of a commonly used power morcellator withdrew its device from the market while at least 1 commercial insurance carrier has eliminated coverage for procedures that use a power morcellator.⁴⁰ Recently, the US Food and Drug Administration recommended a black box warning for electric power morcellators and recommended that power morcellation not be performed in perimenopausal and postmenopausal women undergoing removal of uterine tissue for suspected fibroids.¹⁸

The current data provide prevalence estimates for patients considering myomectomy with use of electric power morcellation and suggest that 1 in 1073 women will be diagnosed with a uterine cancer and 1 in 230 with some form of pathologic abnormality. Given that older women are at the greatest risk for pathologic abnormalities, electric power morcellation should be approached with caution in patients older than 50 years undergoing myomectomy. The frequency of use of electric power morcellators for gynecologic surgery first increased rapidly with a relative lack of data and then abruptly decreased after an adverse outcome in a young woman. These events highlight the difficulty of evaluating, using, and marketing surgical devices. From a public health perspective, these findings highlight the need for more rigorous comparative effectiveness research and heightened regulatory oversight for new devices and procedures.

Accepted for Publication: December 11, 2014.

Corresponding Author: Jason D. Wright, MD, Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, Columbia University College of Physicians and Surgeons, 161 Ft Washington Ave, Eighth Floor, New York, NY 10032 (jw2459@columbia.edu).

Published Online: February 19, 2015. doi:10.1001/jamaoncol.2014.206.

Author Contributions: Dr Wright had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Wright, Tergas, Burke, Chen.

Acquisition, analysis, or interpretation of data: Wright, Cui, Hou, Ananth, Chen, Richards, Neugut, Hershman.

Drafting of the manuscript: Wright, Chen, Hershman.

Critical revision of the manuscript for important intellectual content: Wright, Tergas, Cui, Burke, Hou, Ananth, Chen, Richards, Neugut, Hershman.

Statistical analysis: Wright, Hou, Chen.

Obtained funding: Wright, Neugut.

Administrative, technical, or material support: Wright, Hou, Richards, Hershman.

Study supervision: Wright, Tergas, Burke, Hershman.

Conflict of Interest Disclosures: None reported.

Funding/Support: This work was supported in part by National Cancer Institute (NCI) grants R01 CA169121-01A1 (Dr Wright) and R01 CA166084 (Dr Hershman) and NCI fellowship R25 CA094061-11 (Dr Tergas).

Role of the Funder/Sponsor: The NCI had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Correction: This article was corrected on March 12, 2015, for errors in the Table 1 headings.