Characteristics and mortality risk among esophageal cancer patients with varied HIV status seeking care in western Kenya: A 10-year analysis.

Abstract

e16090 Background: Esophageal cancer (EC) is among the top five prevalent cancers in Africa, including at Jaramogi Oginga Odinga Training and Referral Hospital (JOOTRH). EC is associated with poor prognosis and data to inform improvements are limited. Furthermore, the characteristics and impact of EC in high HIV prevalence region has not been studied. This study explores EC characteristics and evaluates the influence of HIV status on survival and mortality risk. Methods: This retrospective study included all patients diagnosed with EC and receiving care at JOOTRH in Kisumu, Kenya between January 2013 and December 2024. We extracted demographic and clinical data from 4,004 patient medical records. We summarized data using descriptive statistics and Kaplan-Meir curves and Cox proportional Hazard model to assess survival and identify factors associated with death. Results: Among 464 clients, 29% (n = 136) were HIV+, 49% (n = 227) HIV- and 22% (n = 101) had unknown HIV status. 50% (n = 231) of clients were female and the median age of the sample was 60 years.[IQR:50 – 70]. Median age differed by HIV status [HIV+: 53; HIV-:64; HIV unknown: 62; p < 0.01]. 17% (n = 80) reported ever smoking cigarettes while 19% (n = 88) reported ever using alcohol. Treatment coverage was 38% (n = 175) with no significant difference across HIV status [HIV+: 45%; HIV-: 37%; HIV unknown: 29%; p = 0.1]. 11% (n = 49) of patients received surgical treatment alone, 17% (n = 78) received chemotherapy alone, and 8% (n = 39) received a combination of chemotherapy, surgery, or radiation. The median time to treatment from diagnosis was 46 days [IQR:23 - 112] and was similar across HIV status. The clinical outcome of most EC patients was unknown due to loss to follow-up (LTFU) (57%, n = 250,); 22%% (n = 98) were confirmed alive, and 22% (n = 99) dead. The highest proportion of deaths were among those with unknown HIV status, 28% (n = 27) versus 22% (n = 29) and 20% (n = 43) among HIV+ and HIV-, respectively. Median survival time in the sample ranged between 2 and 3 months and was highest among those HIV negative (median: 3 [IQR: 2 – 5]). Survival did not differ by HIV status. Having insurance cover resulted in a ~15% decrease in hazard of mortality when compared to those paying cash (Hazard ratio: 0.85 [95% CI: 0.68 – 1.07, p = 0.02]), and receiving treatment conferred 47% decrease in risk of death (Hazard ratio: 0.52 [95% CI: 0.42 – 0.67, p = 0.0005]). Conclusions: Although almost a third of EC patients were HIV+, we found no difference in survival by HIV status among EC patients. Deaths were highest among clients with unknown HIV status, who formed 22% of the sample. Majority of the patients had insurance cover and being insured was significantly associated with survival. Further research identifying ways to increase insurance coverage and to target care to those most at risk of poor outcomes is needed.