Even though minimally invasive surgery has improved outcomes for hysterectomy, the procedure requires removal of the uterus through small incisions. Morcellation, or fragmentation of the uterus into smaller pieces, is one method to remove the uterus. Recently, concern has been raised that morcellation may result in the spread of undetected malignancies.1
Despite the commercial availability of electric power morcellators for 2 decades, accurate estimates of the prevalence of malignancy at the time of electric power morcellation (herein referred to as morcellation) are lacking,1,2 with single-center studies reporting prevalences from 9 to 100 in 10 000.3,4 We used a large insurance database to investigate the prevalence of underlying cancer in women who underwent uterine morcellation.
The Perspective database was used to identify women who underwent a minimally invasive hysterectomy from 2006-2012. Perspective is an all-payer database including more than 500 hospitals capturing 15% of hospitalizations. Hospitals within this database are more frequently urban teaching centers and located in the southern United States. Data undergo a rigorous quality control process. Use of commercially available morcellators was captured by identification of charge codes.5 The analysis was deemed exempt by the Columbia University institutional review board.
The primary outcome was identification of uterine corpus cancer (all histologies) based onInternational Classification of Diseases, Ninth Revision, coding at surgery. We also examined the occurrence of uterine neoplasms of uncertain malignant potential; malignancies of other parts of the uterus, including cervical cancer, and surrounding adnexal structures (other gynecologic cancer); and endometrial hyperplasia.
Multivariable mixed-effects log-linear models, including clinical and demographic covariates and a random-intercept for the procedural hospital, were developed for uterine cancer and endometrial hyperplasia. The other outcomes were rare and the models did not converge.
All statistical analyses were 2-sided and performed with SAS version 9.4 (SAS Institute Inc). A Pvalue of <.05 was considered statistically significant.
Within the cohort of 232 882 women who underwent minimally invasive hysterectomy from 2006-2012, morcellation was performed in 36 470 (15.7%). Women who underwent morcellation differed in clinical and demographic characteristics from women who did not (eTable in the Supplement). Among those who underwent morcellation, 99 cases of uterine cancer were identified, a prevalence of 27/10 000 (95% CI, 22-32/10 000). Twenty-six cases of other gynecologic malignancies were found (a prevalence of 7/10 000 [95% CI, 4-10/10 000]), 39 uterine neoplasms of uncertain malignant potential (11/10 000 [95% CI, 7-14/10 000]), and 368 cases of endometrial hyperplasia (101/10 000 [95% CI, 91-111 per 10 000]).
Our data demonstrate that uterine cancers occurred in 27 per 10 000 women undergoing morcellation. Other malignancies and precancerous abnormalities were also detected. Although morcellators have been in use since 1993, few studies have described the prevalence of unexpected pathology at the time of hysterectomy.2– 4 Prevalence information is the first step in determining the risk of spreading cancer with morcellation. Although data are limited, women with apparent uterine-confined neoplasms at the time of morcellation have been found to have intraperitoneal tumor dissemination at the time of reexploration.3,6
We recognize a number of limitations including the inability to verify pathological findings, possible misclassification of pathology, potential undercapture of morcellation, and the fact that our findings may not be generalizable to all hospitals. Last, we lack data on long-term follow-up, and the outcome of women with pathological abnormalities who underwent morcellation requires further study. Patients considering morcellation should be adequately counseled about the prevalence of cancerous and precancerous conditions prior to undergoing the procedure.
PubMed | Link to Article
PubMed | Link to Article
PubMed | Link to Article
PubMed | Link to Article
PubMed | Link to Article
PubMed | Link to Article