Introduction

Surgery remains the mainstay of treatment of rectal cancer. Advancement in surgical technique, such as total mesorectal excision (TME) and minimally invasive surgery (MIS), along with multimodal therapies have led to improved surgical and oncologic outcomes [1,2,3]. Rates of sphincter-preserving surgery (SPS) have been proposed to be a quality metric in rectal cancer surgery [4]. Although previous reports have described an increase in the rate of SPS, significant variability and inconsistency exists in its implementation [5,6,7]. Previous reports from US centers have indicated rates of SPS ranging from approximately 50 to 75% [8,9,10]. However, European and Australian centers have reported generally higher and less variable SPS rates [11,12,13]. No study to date has utilized the National Cancer Database (NCDB) to evaluate rates and trends of SPS in the USA. Furthermore, studying the rates and trends of SPS has further value given with the recent development of the National Accreditation Program for Rectal Cancer (NAPRC), and this may serve as a baseline for which future studies can be compared post-implementation of the NAPRC.

We aimed to report on trends and outcomes of SPS from participating hospitals of the NCDB. We hypothesize that rates of abdominoperineal resection (APR), which serve as a proxy for rates of overall non-SPS, should be decreasing over time compared to other resection extents. Furthermore, we hypothesize that rates of SPS may be impacted by patient or facility characteristics as opposed to tumor-related factors.

Materials and methods

This study was deemed institutional review board exempt. The 2014 Participant User File of the National Cancer Database was queried for all patients with a rectal adenocarcinoma from 2008 to 2012. Patients with tumors located in the recto-sigmoid or anus were excluded. Patients undergoing SPS (surgery site codes describing low anterior resection (LAR) or coloanal) were compared to non-SPS (surgery site codes describing abdominoperineal resection). Those patients undergoing local excision, total proctocolectomy, and multivisceral resection (including T4b tumors) were excluded. Any patient with stage IV disease and incomplete staging information were also excluded.

Where appropriate, chi-square or Fisher’s exact tests were used to examine categorical variables. The Cochrane-Armitage trend test was used to evaluate trends of SPS over time. Univariate and multivariable logistic regression models were used to assess factors such as age, sex, race, education level, income level, insurance status, urban/rural location, year of diagnosis, Charlson-Deyo Score, and facility type associated with the likelihood of undergoing SPS.

Univariate and multivariable Cox proportional hazards models were used to assess the differences in mortality. Statistical significance was defined as p < 0.05 for all comparisons. All statistical analyses were completed using SAS 9.4 (SAS Institute Inc., Cary, NC, USA).

The National Cancer Database is a joint project of the Commission on Cancer (CoC) of the American College of Surgeons (ACS) and the American Cancer Society. The CoC’s NCDB and the hospitals participating in the CoC NCDB are the source of the de-identified data used herein; they have not verified and are not responsible for the statistical validity of the data analysis or the conclusions derived by the authors.

Results

Twenty-four thousand eighteen patients were identified, of which 18,452 (76.8%) underwent SPS. The majority of these patients (n = 13,688, 74.2%) were aged < 70. A greater proportion of female patients underwent SPS compared to male patients (78.4% vs. 75.8%, p < 0.001). Demographic and clinical characteristics are outlined in Table 1. Of note, 51.8% of the SPS patients had private insurance and 42.8% had Medicare or Medicaid. This compared to 44.9% in the non-SPS cohort with private insurance and 47.4% with Medicare or Medicaid (p < 0.001). Regarding clinical TNM stage, a greater proportion of Stage II and III patients underwent non-SPS compared to SPS (40.2 vs. 33.2% and 40.2 vs. 34.4%, respectively, both p < 0.001). The majority of patients (54.2%) underwent treatment at a comprehensive community cancer program. More than half of the patients (58.5%, n = 13,342) received neoadjuvant chemoradiotherapy. Of this cohort, 9635 (72.2%) underwent SPS.

Table 1 Basic demographics, clinical characteristics of the study population
Full size table

Univariate and multivariable logistic regression models for assessing factors associated with SPS are reported in Table 2. Univariate analysis shows that patients with age ≥ 70 years, male sex, living in a rural or urban region, and having Medicare, Medicaid, or no insurance were less likely to undergo SPS. Multivariable analysis shows that patients with age ≥ 70 (OR 0.87, 95% CI 0.80–0.95), male sex (OR 0.90, 95% CI 0.84–0.96), having Medicare (OR 0.83, 95% CI 0.76–0.90), Medicaid (OR 0.72, 95% CI 0.63–0.81), and poorly differentiated histology (OR 0.78, 95% CI 0.71–0.85) were less likely to undergo SPS.

Table 2 Univariate and multivariable logistic models assessing factors associated with undergoing SPS
Full size table

Univariate and multivariable Cox proportional hazards models calculated for mortality are demonstrated in Table 3. Notably, on univariate analysis, patients undergoing non-SPS had a higher likelihood of mortality compared to those with SPS (HR 1.38, 95% CI 1.28–1.49). Patients aged ≥ 70 were more likely to die with estimated hazard ratio of 2.53 (95% CI 2.36–2.71) compared to patients aged < 70. Males (HR 1.17, 95% CI 1.09–1.26) were at increased risk of death. There was lower risk of death in patients belonging to metropolitan areas vs. those from urban (HR 1.12, 95% CI 1.03–1.23) areas. Patients having private insurance were less likely to die compared to those having Medicare (HR 2.49, 95% CI 2.31–2.69) or Medicaid (HR 1.95, 95% CI 1.68–2.26), and no insurance (HR 1.71, 95% CI 1.44–2.04). Compared to community cancer programs, academic/research programs (HR 0.66, 95% CI 0.59–0.74) and comprehensive community cancer programs (HR 0.89, 95% CI 0.80–0.99) had a lower likelihood of mortality. Multivariable analysis shows that non-SPS, age ≥ 70 years, male gender, insurance other than private and poorly differentiated grade had higher likelihood of mortality.

Table 3 Univariate and multivariable Cox proportional hazards models (please note—year of diagnosis 2012 was removed due to missing follow-up)
Full size table

There was no statistical difference in the proportion of patients undergoing SPS over the study period (76.0% in 2008 to 77.2% in 2012, p = 0.386). The trends of SPS over time are shown in Table 4.

Table 4 Trends of SPS over time
Full size table

Discussion

This study described the rate, pattern, and associated factors of sphincter-preserving surgery in rectal cancer patients diagnosed from 2008 to 2012 from the National Cancer Database. The study not only identified tumor and patient-related factors but also highlighted some of the disparities in care. Our study identified that older age (≥ 70 years), male sex, having a government insurance, and poorly differentiated grade were factors leading to a lower likelihood of undergoing SPS.

The rate of SPS in the USA has been variable and historically less than that reported in European centers, but has reportedly increased in recent years [11, 13]. A study conducted on 41,631 rectal cancer patients in 2007 reported an increase in the proportion of SPS from 1988 (26.9%) to 2003 (48.3%) [14]. Our study of the NCDB database showed a greater percentage of patients undergoing SPS (76.8%) as treatment for rectal cancer than non-SPS (23.2%). In an Australian study conducted by Marwan and colleagues, similar results were reported, with 76.6% patients having SPS for rectal cancer among whom the majority were aged < 70 years (60%) and were of male sex (64.7%) [12]. Abdelsattar and colleagues studied 329 rectal cancer patients who were treated at ten different hospitals in Michigan and reported the overall rate of SPS to be 72% [6]. Given this report of the NCDB and similar results compared to other reports may indicate that the rates of SPS and non-SPS in the USA have plateaued.

Variability in care in the USA has previously been attributed to colorectal specialization and hospital/surgeon volume [15,16,17,18]. Secondary to these deficiencies and variances in care, the American College of Surgeons in conjunction with the Commission on Cancer have recently developed a National Accreditation Program for Rectal Cancer [19]. It remains to be seen if implementation and adherence to such standards improves overall disparities of care, either by forcing improved training, greater specialization, or regionalization of care by creation of centers of excellence. It is notable, however, that in our study, there was no statistically significant difference in rates of SPS between facility types. While differences may in reality exist, there are limitations in the available data through NCDB with regard to facility type that may make a true determination of difference based on surgery location difficult.

Several previous studies have revealed that rectal cancer patients having private insurance and of greater socioeconomic status are more likely to undergo SPS [8, 14]. These reports are congruent with findings in our study which demonstrated that having private insurance was correlated with an increased probability of undergoing SPS. This may be due to the reason that each of these variables offers greater options in choosing hospital and surgeon compared to government insurances.

The strengths of this study are the large patient cohort (n = 24,018) with comprehensive cancer-related data. The NCDB has been reported to capture approximately 70% of new cancer diagnoses in the USA [20]. Limitations include the retrospective nature and lack of detailed patient/tumor specific factors. Furthermore, NCDB lacks granular data describing exact facility type that the operative intervention took place, thus making it difficult to evaluate trends in SPS by facility type.

In summary, we found that SPS rate is higher than non-SPS among rectal cancer patients. Regional and institutional disparities persist, and the positive impact of the implementation of the NAPRC on this and general outcomes in rectal cancer care is eagerly anticipated.

Contribution of each author/coauthor

Authors AM, FS, and PMK conceptualized the study and discussed with other authors DC, KLM, and DWL. Authors AM, FS, PMK, DC, KLM, and DWL designed and planned the study and developed methodology. Data was analyzed by statistician CND. This was in close discussions and revisions with authors AM and EH. All the authors approved of the final analysis and results. AM and FS wrote the initial draft which was extensively reviewed and edited by all the authors. Final version of the manuscript was approved by all the authors prior to submission of the paper.