Uterine Size and Route of Hysterectomy: Measuring the Association Using the National Surgical Quality Improvement Program Database (NSQIP)

INTRODUCTION: Advances in minimally invasive surgery (MIS) have allowed for complex pathology to be done without requiring open surgery. This includes minimally invasive hysterectomy (MI-H) through laparoscopic or vaginal approaches for large uteri. Many factors influence the route of hysterectomy and uterine size is important. Case series of MI-H for large uteri have demonstrated its feasibility. However, there is wide variation in the use of MI-H for removing the large uterus. To our knowledge, the probability of MI-H or AH based on uterine size is unknown. OBJECTIVE: To quantify the association of uterine size with the route of hysterectomy. METHODS: The American College of Surgeons National Surgical Quality Improvement Program (NSQIP) database was used to identify patients who underwent hysterectomy from 2014 to 2021. Malignancies, emergencies, non-hysterectomy procedures, cases missing uterine weight, and cases with concomitant procedures were excluded. Route of hysterectomy was classified as MI-H (laparoscopic and vaginal route) or AH. 116,019 patients were included in the analysis. To determine the association of uterine size with the likelihood AH (the primary outcome), univariable and multivariable Poisson regression models with robust standard errors were conducted. Uterine weight was grouped into 6 categories and relative risks (RRs) were calculated for each uterine category compared to the reference (<250 g). The multivariable model was adjusted for 14 clinically significant potential confounders. To identify a uterine weight with at least a 50% predicted probability of AH, a logistic regression model was run with the same outcome. Continuous uterine weight and the14 variables were used as predictors. The model was fit on a training dataset (80% of the data), with predicted probabilities calculated on a test dataset (20% of the data). Continuous predictors were held at their medians and other predictors were held at a constant category. RESULTS: There was a higher proportion of AH with each increase in the uterine weight category (Table 1). AH is more common than MI-H once the uterine weight reaches a category of 750 g or more. At the largest weight category (>1500 g), only 12% of patients undergo MI-H. The relative risk of AH based on uterine weight category adjusted for other variables (aRR) increases with each category of uterine weight, compared to the reference level. A uterine weight of 750 to 1000 g was associated with a 383% increase in the risk of AH (aRR [95% CI]: 4.83 [4.69, 4.98]; p<0.001). For cases from 2014, patients had a predicted probability over 50% of undergoing AH when uterine weight was greater than 750 g. This increased to 1,064 in 2021, the most recent year of data (Figure 1). CONCLUSIONS: Increasing uterine size is a significant factor associated with the risk of AH even when controlled for other variables. Progression in MIS has allowed for MI-H in increasingly large uteri, although only 23.5% of patients with a uterus of more than 750 g are undergoing MI-H. The RR of AH increases significantly even when adjusted for other factors. For the most recent data, a patient with a uterus larger than 1,064 g had a predicted probability over 50% of undergoing AH.