EP621 - Local and national trends in emergency presentation of colon and rectal cancer

Purpose/Background: It has been shown that emergent diagnosis and treatment of colorectal cancer is associated with higher rate of postoperative complications and poorer prognosis. We have previously shown that we have high rates of emergent presentation of colon cancer at our institution, and we sought to assess similar outcomes for rectal cancer as well as compare our own results to national data.

Methods/Interventions: Retrospective review was conducted of all patients treated for colon and rectal cancer at a single academic tertiary care facility between 2012 and 2021. Information about the setting of diagnosis, timing of resection, presenting symptoms, and demographics was collected.
Kendall’s tau was used to identify trends over time in five major outcomes: setting of diagnosis, surgery during index encounter, setting of resection, and obstruction and/or perforation at time of surgery. Binomial logistic regression was used to determine associations between said outcomes and various preoperative and demographic factors. Analyses were performed for colon and rectal cancer combined as well as individually.
NSQIP data of patients who underwent colectomy for colon cancer from 2012-2021 was analyzed for emergent surgery, obstruction, and perforation. Similar data was not available for rectal cancer.

Results/Outcomes: A total of 1162 patients were included (733 with colon cancer, 429 with rectal cancer). Overall, 38.7% of patients were diagnosed in the inpatient setting (45.8% colon alone, 26.6% rectal alone). Of oncologic resections performed, 11.1% were urgent/emergent (41.5% colon alone, 3.0% rectal alone). 16.2% of patients were obstructed preoperatively (16.7% colon alone, 13.8% rectal alone) and 5.4% were perforated (6.0% colon alone, 4.4% rectal alone). Over the study period, inpatient diagnosis, urgent/emergent resection, obstruction, and perforation increased over time, with obstruction being statistically significant (p < 0.001).
Overall, Medicaid insurance status increased risk of inpatient diagnosis (OR=3.71), surgery during index encounter (OR=3.15), obstruction (OR=2.33), and perforation (OR=4.85). Being up to date on colonoscopy was protective against inpatient diagnosis (OR=0.54), surgery during index encounter (OR=0.54), and obstruction (OR=0.47). Black patients were more likely to undergo urgent/emergent resection (OR=1.85), patients who reported melena/hematochezia were less likely to undergo urgent/emergent resection, and female patients were less likely to be obstructed (OR=0.60).
NSQIP data from the same time period reports 6.4% of colectomies for colon cancer were emergent, 4.0% of patients were obstructed, and 1.5% of patients were perforated. There was no significant change in any of these measures over time.

Conclusion/Discussion: The rate of emergent presentation and treatment of colon and rectal cancer at our institution is strikingly high, with these outcomes increasing over time. Medicaid insurance status was a significant risk factor while being up to date on screening colonoscopy was protective, suggesting that access to care is the major barrier to timely diagnosis and treatment. When compared to national data, our institution reports a higher percentage of emergent surgery, obstruction, and perforation, which we suspect reflects the large population of underserved patients who seek care at our urban, tertiary facility; this adds important data to the literature, as it may reflect other tertiary centers in underserved areas. We are currently taking steps to improve access to screening.

Table 1
Likelihood of inpatient diagnosis of colorectal cancer. Multivariate binary logistic regression was performed for preoperative variables determined significant by chi-squared test. P-values ≤0.05 were considered significant and are displayed in bold.

Figure 1
Trends over time from 2012-2021 in A) inpatient diagnosis, B) urgent/emergent resection, C) obstruction preoperatively, and D) perforation preoperatively. Trends were calculated using Kendall’s tau.