Utilizations and characteristics of ovarian conservation at hysterectomy for cervical carcinoma in situ

To examine the trends and characteristics of ovarian conservation at time of hysterectomy in cervical carcinoma in situ.

carcinoma in situ who have completed childbearing and are medically fit for surgery, total hysterectomy is commonly offered as definitive surgical treatment. 1,2 One consideration during total hysterectomy for cervical carcinoma in situ is whether a concurrent oophorectomy should be performed, taking into consideration the risks and benefits of this additional procedure.
The provider needs to consider the risks of (1) infrequent occult invasive cervical cancer that may result in ovarian micrometastasis after ovarian conservation 1-3 and (2) the morbidity of early surgical menopause following oophorectomy. [4][5][6][7][8] The 2020 Society of Gynecologic Oncology (SGO) recommendation with endorsement from the American Society for Colposcopy and Cervical Pathology (ASCCP) stated that the decision regarding oophorectomy at time of hysterectomy for cervical carcinoma in situ is left to the judgment of the clinician based on patient age, risk factors, and hormonal status with an implied understanding that the impact of early surgical menopause is not benign. 2 As a premalignant condition, cervical carcinoma in situ falls on the spectrum between benign and malignant, therefore nuanced counseling regarding ovarian conservation is required in these patients. Providers are asked to use their clinical judgment regarding ovarian conservation in patients with cervical carcinoma in situ, which may be challenging in the absence of concrete guidelines or national-level data on these practices. This study examined the contemporary trends and characteristics of ovarian conservation at time of hysterectomy in cervical carcinoma in situ.

| Data source
The National Inpatient Sample (NIS) was used for this study. The NIS program is a publicly available and de-identified population-based all-payer database that is sponsored by the Agency for Healthcare Research and Quality and was developed as a part of the Healthcare Cost and Utilization Project. 9 The NIS collects information regarding inpatient utilization, cost, quality, and outcomes by randomly sampling 20% of admission records in each hospital, with the weighted sample representing more than 90% of the US population. 9 This study used publicly available deidentified data, and does not contain any studies with human participants performed by any of the authors. The University of Southern California Institutional Review Board deemed this study exempt because of the use of publicly available deidentified data (exemption HS-16-00481).

| Inclusion and exclusion
This is a retrospective cohort study examining the NIS program from January 2016 to December 2019. This study point was chosen as a result of the introduction of WHO International Classification of Disease 10th revision (ICD-10) codes into the NIS program. Patients aged 65 years or less with carcinoma in situ of the uterine cervix who had hysterectomy were eligible. This age cut-off was chosen based on the results of a previous analysis demonstrating the possible decrease in all-cause mortality with ovarian conservation at benign hysterectomy in this age demographic. 7 The case identification for carcinoma in situ of the uterine cervix was based on the ICD-10 Clinical Modification diagnosis code of D06, which was consistent throughout the study period. This ICD-10 code includes cervical adenocarcinoma in situ and cervical intraepithelial neoplasia III.
Patients were excluded if they did not have a hysterectomy or if surgical information was unavailable, were older than 65 years, or had adnexal pathology, and/or gynecologic malignancy such as uterine, cervical, or ovarian cancer. Patients were also excluded if nodal evaluation was performed at the time of surgery. These exclusions were selected to ensure assessment of the effect of carcinoma in situ of the uterine cervix on adnexal surgery.

| Outcome measures
The co-primary outcomes were (1) temporal trends and (2) patient, surgical, and hospital characteristics related to ovarian conservation at time of hysterectomy for carcinoma in situ of the uterine cervix.
These end points are clinically relevant and were previously examined in other types of gynecologic premalignant disease. 10 Patients were allocated to the oophorectomy group if they had an ICD-10 code for bilateral adnexectomy at the time of surgery whereas those without the code were assigned to the ovarian conservation group. This strategy for exposure assignment followed the same definition as previous analysis. 11

| Study covariates
Baseline information including patient demographics, surgical treatment, and facility parameters were collected from the NIS program for eligible participants. This study followed the same ICD-10 codes for the extraction of information that was unchanged during the study period. 10,12 Abstracted patient characteristics included age, year, race, and ethnicity determined per the NIS program, primary expected payer, median household income, Charlson Comorbidity Index calculated according to previous study, 12,13 and obesity. Information was collected regarding the treatment facility such as the hospital bed capacity, location and teaching status, and regional area. Study covariates for surgical treatment included hysterectomy modality and use of robotic-assisted surgery.

| Statistical consideration
The first step of analysis was to examine the trend of ovarian conservation over time. Temporal trends of ovarian conservation over time were assessed with the Cochrane-Armitage trend test. An age-specific trend of ovarian conservation was assessed with a linear segmented regression with log-transformation. 14 The identified inflection points were determined in an automated fashion in the analysis that was used for the age clustering. In each segment, statistical significance of the slope was determined.
The second step of the analysis was to identify the independent characteristics related to ovarian conservation at hysterectomy for cervical carcinoma in situ. In the multivariable analysis, a binary logistic regression model was fitted to determine independent characteristics associated with ovarian conservation. Initial covariate selection was set at a P < 0.05 in the univariable analysis.
Conditional step-forward selection was then performed with the stopping rule of P < 0.05 in the final model. 15 Effect size for ovarian conservation compared with oophorectomy was expressed with adjusted odds ratio with a corresponding 95% confidence interval.
The last step of analysis was to assess the utilization patterns of ovarian conservation at hysterectomy for cervical carcinoma in situ. A classification-tree was constructed by fitting recursive partitioning analysis. 16 All the independent factors for ovarian conservation were entered in the modeling and the χ 2 automatic interaction detector method was used with a stopping rule of maximum three layers. In the determined patterns, the utilization rate of ovarian conservation was computed.
The weighted values for national estimates provided by the NIS program were used, and statistical interpretation was based on a two-tailed hypothesis. A P value less than 0.05 was considered statistically significant. SPSS version 28.0 (IBM) was used for all analyses. The STROBE reporting guidelines were consulted to summarize the performance of the cohort study.

| Study cohort
A total of 6605 women aged 65 years or less underwent hysterectomy for carcinoma in situ of the uterine cervix during the study period for national estimates ( Table 1). The median age was 43 years (interquartile range 36-51 years). Most patients were White individuals, privately insured, non-obese, had no comorbidities, and underwent abdominal hysterectomy at large urban teaching centers.

| Temporal trends of ovarian conservation
More than half of included participants had ovarian conservation at the time of hysterectomy (57.2%). The median age for those who had ovarian conservation was 39 years (interquartile range 34-45 years). The performance of ovarian conservation was unchanged during the study period, ranging from 59.0% to 56.6% Ovarian conservation rates remained stable until age 40 years, ranging from 88.0% to 78.6% (P-trend = 0.236; Figure 1), after which time the rate sharply and significantly decreased from 78.6% to 19.1% (P-trend < 0.001). After 40 years of age, each 1-year age increment resulted in a 6.0% (95% confidence interval 4.4%-7.5%) decrease in the rate of ovarian conservation.

| Characteristics of ovarian conservation
In a univariable analysis (Table 1), all the measured study covariates except for the year and presence of obesity were associated with ovarian conservation (all, P < 0.05). In a multivariable analysis ( Table 2), (1) patient characteristics including younger age and higher household income, (2) hospital characteristics including urban non-teaching or urban teaching centers, and (3) surgical characteristics including vaginal hysterectomy or laparoscopyassisted vaginal hysterectomy were independently associated with increased utilization of ovarian conservation (all, P < 0.05). In contrast, decreased rates of ovarian conservation were seen in (1) patients with more comorbidities and (2) hospitals with larger bed capacity (all, P < 0.05).

| Utilization patterns of ovarian conservation
In the evaluation of utilization patterns of ovarian conservation ( Figure 2 and Table 3), a classification-tree model identified 17 unique patterns of ovarian conservation based on patient factors (age, comorbidity, and household income), treatment facility factors (hospital bed capacity and hospital teaching status), and surgical factors (hysterectomy approach). Patient age was the first indicator for allocation in the first layer of the classification-tree (≤40 vs. >40 years, 80.3% vs. 41.2%), followed by hysterectomy type and household income status in the second layer.
Among the 17 utilization patterns, six had an ovarian conservation rate of more than 80% (range 80.0%-94.4%; Table 3). Patients aged 40 years or younger with second or third quartile household income who had vaginal hysterectomy had the highest rate of ovarian conservation of 94.4%. There were seven patterns that had ovarian conservation rates less than 50% (range 17.2%-48.0%). Patients older than 40 years who had high comorbidity indices and underwent total abdominal hysterectomy had the lowest ovarian conservation rate of 17.2%. Collectively, the absolute percentage rate difference between the highest and lowest groups was 77.2% (range 17.2%-94.4%).

| DISCUSS ION
The key results of the current study include the following: first, ovarian conservation in those with carcinoma in situ of the uterine cervix was relatively common overall, but the utilization of ovarian conservation dropped sharply in patients in their early forties. Second, there was substantial variability in ovarian conservation based on patient, hospital, and surgical factors. As data specific to cervical carcinoma in situ were previously not available, these results add important information to the literature.
An inverse relationship between patient age and ovarian con-  Nevertheless, multiple studies have reported that ovarian conservation for early cervical cancer is not associated with worse oncologic outcomes but rather is associated with decreased all-cause mortality including cardiovascular death in young patients. 18,24 The association of early oophorectomy with increased mortal- This study also found substantial variability in the utilization of ovarian conservation for cervical carcinoma in situ. The heterogeneous patterns in the utilization of ovarian conservation, ranging from less than 20% to more than 90% depending on patient, hospital, and surgical factors, seen in this study may reflect a lack of consensus among providers as to who is a candidate for ovarian F I G U R E 2 A classification-tree model for ovarian conservation at hysterectomy. Metadata for the results are displayed in Table 3. The values with higher than the cohort-level average are shown in red. *including unknown. Abbreviations: abdominal, abdominal hysterectomy; CCI, Charlson Comorbidity Index; hyst, hysterectomy mode; LAVH, laparoscopy-assisted vaginal hysterectomy; MW, Midwest; NE, Northeast; NOS, not otherwise specified; Obes, obesity; Ov con, ovarian conservation; S, South; TLH, total laparoscopic hysterectomy; vaginal, vaginal hysterectomy; W, West.
conservation in those undergoing surgical treatment of cervical carcinoma in situ. Notably, a wide range in the utilization of ovarian conservation at hysterectomy was also reported in benign gynecologic disease, and the investigators called for stronger evidence-based guidelines on surgical practice. 25 The current clinical practice guidelines provided by the National Comprehensive Cancer Network do not specify management recommendations for cervical carcinoma in situ. 26 As noted earlier, the SGO/ASCCP statement does not clearly provide a concrete approach or schema for oophorectomy at hysterectomy for cervical carcinoma in situ. 2 Developing an algorithm to assess appropriate candidates for ovarian conservation for those with premalignant disease would therefore be beneficial. Based on the best practice estimates, a proposed hypothetical algorithm is shown in the supplementary material ( Figure S1) and merits further development. This may be used to triage ovarian conservation candidates in a two-step schema similar to those with endometrial hyperplasia. 27 There are several limitations in this study. First, there is unmea- Second, the accuracy of data was not assessable because actual medical record review was not performed. Third, long-term morbidity after discharge from admission at time of hysterectomy was un-  Figure 2. In each pattern, frequency rate per study population and ovarian conservation rate in the pattern are computed.
Third, as above, there is an opportunity to develop a clinical practice consensus and/or guidelines regarding ovarian conservation at hysterectomy for those with carcinoma in situ of the uterine cervix.
In light of mounting data on the protective cardiovascular effects of ovarian hormones in premenopausal patients in the recent years, it may be the time to revisit this objective. [4][5][6] In conclusion, rapid decrease in the utilization of ovarian conservation at time of hysterectomy for cervical carcinoma in situ began 10 years earlier than the average age of spontaneous menopause and the rate of ovarian conservation remained unchanged over time, which may suggest that both surgeons and patients are yet comfortable with ovarian preservation in cervical carcinoma in situ after age 40 years. As above, long-term benefits of ovarian conservation cannot be overstated for premenopausal women and more studies specific to cervical carcinoma in situ are warranted.

CO N FLI C T O F I NTE R E S T
LDR is a consultant for Quantgene. All other authors have no conflicts of interest to declare.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are openly available