These results indicate a difference in readmission rates by HIV status. The current study demonstrated that in BL hospitalizations, comorbid HIV was associated with a significantly higher unadjusted and adjusted 30-day readmission rate and a significantly higher unadjusted 90-day readmission rate (p = 0.021, p = 0.026, p = 0.031, respectively). Hospitalizations of patients carrying an HIV and BL diagnosis had 43% higher odds of readmission within 30 days after discharge, after adjusting for common comorbidities (p = 0.026). However, there was no significant difference in 90-day readmission rate in HIV-positive BL patients after adjusting for comorbidities (p = 0.053). Of note, the unadjusted 30-day and 90-day readmission rates were statically significant, indicating a greater overall risk of readmission for PWH and BL.

One possible explanation for the increased readmission rate is the increased toxicity of the chemotherapy regimen in PWH. In our study, the most common reason for 30-day and 90-day readmission was chemotherapy at 54.96% and 61.46%, respectively. A 2017 study by Xiao, et al. [16], demonstrates more severe toxicity to the CODOX-M/IVAC ± R chemotherapy regimen was experienced by HIV-positive patients due to lower CD4 + counts and lower Karnofsky performance status. Difficulty in tolerating chemotherapy could require more supportive measures during treatment and therefore higher readmission rates. Another possible explanation for this increased readmission rate is the higher rate of infection during treatment. In 2021, Wang, et al. [17] demonstrated almost 60% of PWH and lymphoma experienced infections during chemotherapy. Patients with a higher number of chemotherapy cycles, grade 4 decrease in neutrophil counts, and less than 6-month duration of ART at time of diagnosis were all independent infection risk factors [17]. The combination of myelosuppression and already reduced CD4 count creates a uniquely higher risk of infection development over the course of treatment for this patient population [17]. However, our study found that rate of sepsis-related readmissions was only 4.23%for 30-day readmissions and 3.87% for90-day readmissions. We were not able to determine if other infections impacted readmissions for other hematological and immunological diagnoses due to administrative coding limitations (MS-DRG and ICD coding). Notably, readmissions for other hematological and immunological diagnoses were more common than sepsis-specific readmissions (30-day readmission: major hematological and immunological diagnosis 9.11% vs. sepsis-specific 4.23%l; 90-day readmission: major hematological and immunological diagnosis 9.12% vs. sepsis 3.87%).

BL survival outcomes increased from 10 to 40% in the pre-ART era to 70–80% in the current era, indicating advancements in antiretroviral therapy and HIV treatment may have contributed to increased survival in HIV-associated lymphomas [11]. However, a recent National Cancer Database study demonstrated lymphoma patients with HIV continue to experience worse survival even following implementation of ART, leading us to stratify readmission rates during different time periods of ART implementation in our study [6]. Our data demonstrates universal ART and use of integrase inhibitor based regimens were not associated with any significant differences in readmission rates for those with HIV-associated BL for all-cause 30-day and 90-day readmission. Our findings further support previous data indicating outcomes among HIV-associated BL patients no longer associate with HIV specific features but are more likely to be associated with cancer and treatment specific features [8, 9].

Notably, our results showed between-HIV status disparities in socioeconomic factors, particularly income quartiles. Of HIV-associated BL patients, 38.2% were in the lowest income quartile compared to 23.6% of HIV-negative BL patients (p < 0.001). Previous studies also demonstrated people living with material deprivation (i.e., those in lower income quartiles) are more likely to participate in risky behavior leading to higher chances of HIV infection [18]. Economic issues leading to homelessness, lack of necessities, and sex exchange often result in nonadherence [18].

The current study evaluated data from 2010 to 2020 and importantly, the COVID-19 pandemic began impacting healthcare in the United States in January of 2020. We were unable to assess the impact of COVID-19 on readmissions during 2020; an ICD-10 diagnosis code for COVID-19 was not available until October 1, 2020. We acknowledge that COVID-19 may have impacted readmission outcomes during 2020 (e.g., readmissions due to COVID-19 infections, shortage of hospital beds available, etc.); however, 90% of our data is pre-COVD-19 (2010–2019). Future research should consider the effect of COVID-19 on hospital admissions and readmissions for the subpopulation of HIV-associated lymphoma patients.

Lastly, the NRD is an administrative database and provides no HIV or BL specific information, particularly treatment status or CD4 count. As such, our study has limitations, particularly regarding lack of CD4 count, AIDS status, opportunistic infections, antiretroviral treatment status and COVID-19 status and vaccination. Relatedly, the NRD only captures diagnoses and procedures identified by ICD-codes thereby limiting the specificity of diagnostic and procedural information. Additionally, the NRD is limited to the specificity of the standardized Healthcare Cost and Utilization Project (HCUP) variables (race, income quartile, expected primary payer); these variables may have less detailed racial and ethnic information compared to patient data. Despite these limitations, we identified 510 HIV-associated BL patients in a national sample. To the best of our knowledge, no previous studies have examined readmission rates and characteristics of HIV-associated BL hospitalizations.