Differences in Virological and Immunological Risk Factors for Non-Hodgkin and Hodgkin Lymphoma

Abstract

Background: Non-Hodgkin lymphoma (NHL) and Hodgkin lymphoma (HL) are increased in populations with immune dysfunction, including people living with HIV; however, there is little evidence for to what degree immunological and virological factors differently affect NHL and HL risk.

Methods: Data from the Data Collection on Adverse events of Anti-HIV Drugs Study cohort were analyzed to identify independent risk factors for NHL and HL using hazard ratios (HRs), focusing on current and cumulative area under the curve (AUC) measures of immunological and virological status. Variables with different associations with NHL and HL were identified using marginal Cox models. All statistical tests were two-sided.

Results: Among 41 420 people followed for 337 020 person-years, 392 developed NHL (incidence rate = 1.17/1000 person-years of follow-up [PYFU], 95% confidence interval [CI] = 1.06 to 1.30) and 149 developed HL (incidence rate = 0.44/1000 PYFU, 95% CI = 0.38 to 0.52). Higher risk of both NHL and HL was associated with lower current CD4 cell count (adjusted HR [aHR] of NHL for CD4 <100 vs > 599 cells/mm3 = 8.08, 95% CI = 5.63 to 11.61; HL = 4.58, 95% CI = 2.22 to 9.45), whereas higher current HIV viral load (aHR of NHL for HIV-VL >1000 vs < 50 copies/mL = 1.97, 95% CI = 1.50 to 2.59) and higher AUC of HIV-VL (aHR of NHL for highest vs lowest quintile = 2.91, 95% CI = 1.92 to 4.41) were associated with NHL only. Both current and AUC of HIV-VL were factors that had different associations with NHL and HL, where the hazard ratio for NHL was progressively higher than for HL with increasing HIV-VL category. Lower current CD4 cell count had a strong but similar association with both NHL and HL.

Conclusions: CD4 depletion increased risk of both types of lymphomas while current and accumulated HIV-VL was associated with NHL only. This suggests that NHL development is related to both CD4 cell depletion and added immune dysfunction derived from ongoing HIV replication. This latter factor was not associated with HL risk.

Figure 1.

Crude incidence of non-Hodgkin lymphoma and Hodgkin lymphoma by calendar time. CI = confidence interval; HL = Hodgkin lymphoma; NHL = non-Hodgkin lymphoma; PYFU = person-years of follow-up.

Figure 2.

Adjusted ratio of the hazard ratios of non-Hodgkin lymphoma (NHL) and Hodgkin lymphoma (HL) for each considered risk factor. P value of the Wald chi-square statistic with respect to a chi-square distribution with one degree of freedom. All statistical tests of significance were two-sided. Models were adjusted for age, sex, ethnicity, mode of HIV acquisition, smoking status, body mass index, baseline year, cumulative time on combination antiretroviral treatment, hepatitis C virus and hepatitis B virus status, prior AIDS-defining cancer (ADC), prior AIDS events (excluding ADC), prior cardiovascular disease diagnosis, hypertension, diabetes, current CD4, current HIV viral load, area under the curve of HIV-viral load. Risk factors for which the 95% confidence interval for the ratio of the hazard ratios of NHL to HL does not cross 1 (and a corresponding P value < .05) indicate that there is evidence of a different association of the risk factor with HL and NHL. AUC = area under the curve; cART = combined antiretroviral therapy; CI = confidence interval; HIV-VL = HIV viral load; HL = Hodgkin lymphoma; HR = hazard ratios; NHL = non-Hodgkin lymphoma; RHR = ratio of the hazard ratios.