Impact of Co-Morbidities on Outcome in a Predominantly Hispanic Population of Hodgkin and Non-Hodgkin Lymphoma Patients

BACKGROUND Evidence suggests that co-morbidities at diagnosis can influence treatment decisions and outcomes in lymphoma patients. Considering the bimodal presentation of Hodgkin Lymphoma (HL) and that the majority of Non-Hodgkin lymphoma (NHL) patients are over 65 years of age, it can be especially challenging to manage them as older patients have a higher number of co-morbidities. Studies have shown that comorbidity is associated with an inferior outcome and a lower likelihood of receiving treatment with curative intent. It must also be noted that older adults with significant co-morbidities are often excluded from clinical trials due to co-morbidities and that Hispanics (HI) have been historically underrepresented. There is a need to take a closer look at this precise patient population. The main objective of our study was to determine the common co-morbidities and their impact on outcome in a prevalently Hispanic population with both HL and NHL at the only NCI designated Cancer Center of South Texas. To our knowledge this is the largest cohort of HL patients from a single academic institution that serves primarily Hispanics. METHODS We located and retrospectively analyzed a total of 616 patients with diagnosis of Lymphoma (HL and NHL) by International Classification of Diseases (ICD) codes and identified 477 patients who met criteria for inclusion; the patients all received care at UT Health San Antonio, between 2008-2018. Key variables for each patient included age, gender, race/ ethnicity, comorbidities, and vitality status in 2018. Continuously distributed outcomes were summarized with the mean and standard deviation and categorical outcomes were summarized with frequencies and percentages. The significance of variation in the mean with disease category was assessed with one way ANOVA and the significance of associations between categorical outcomes was assessed with Pearson's Chi Square or Fisher's Exact test as appropriate. Multivariate logistic regression was used to model binary outcomes in terms of covariates and indicators of disease. All statistical testing was two-sided with a significance level of 5%. R1 was used throughout. The study was approved by the local Institutional Review Board. The findings will be available to patients, funders and medical community through traditional publishing and social media. RESULTS We identified 477 patients with HL and NHL, 262 were Hispanic (HI) [55%], 205 non-Hispanic (NH) [43%], 10 not specified [2%];