Changes in Clinical Context for Kaposi's Sarcoma and Non-Hodgkin Lymphoma Among People With HIV Infection in the United States

Abstract

Purpose: The biology of HIV-associated cancers may differ depending on immunologic and virologic context during development. Therefore, an understanding of the burden of Kaposi's sarcoma (KS) and non-Hodgkin lymphoma (NHL) relative to antiretroviral therapy (ART), virologic suppression, and CD4 count is important.

Patients and methods: KS and NHL diagnoses during 1996 to 2011 were identified among patients with HIV infection in eight clinical cohorts in the United States. Among patients in routine HIV clinical care, the proportion of cases in categories of ART use, HIV RNA, and CD4 count at diagnosis were described across calendar time. Person-time and incidence rates were calculated for each category.

Results: We identified 466 patients with KS and 258 with NHL. In recent years, KS was more frequently diagnosed after ART initiation (55% in 1996 to 2001 v 76% in 2007 to 2011; P-trend = .02). The proportion of patients with NHL who received ART was higher but stable over time (83% overall; P-trend = .81). An increasing proportion of KS and NHL occurred at higher CD4 counts (P < .05 for KS and NHL) and with undetectable HIV RNA (P < .05 for KS and NHL). In recent years, more person-time was contributed by patients who received ART, had high CD4 counts and had undetectable HIV RNA, whereas incidence rates in these same categories remained stable or declined.

Conclusion: Over time, KS and NHL occurred at higher CD4 counts and lower HIV RNA values, and KS occurred more frequently after ART initiation. These changes were driven by an increasing proportion of patients with HIV who received effective ART, had higher CD4 counts, and had suppressed HIV RNA and not by increases in cancer risk within these subgroups. An improved understanding of HIV-associated cancer pathogenesis and outcomes in the context of successful ART is therefore important.

Fig 1.

Antiretroviral therapy (ART), immune, and virologic status of patients with HIV in routine clinical care who were diagnosed with Kaposi's sarcoma and non-Hodgkin lymphoma. (A) and (B) The proportion of Kaposi's sarcoma and non-Hodgkin lymphoma patient cases that occurred at various times relative to ART initiation over three calendar periods. (C) and (D) The distribution of CD4 counts at diagnosis of Kaposi's sarcoma and non-Hodgkin lymphoma across time, including a LOWESS curve to illustrate the smoothed trend over time. (E) and (F) The distribution of HIV RNA values on the log₁₀ scale at diagnosis of Kaposi's sarcoma and non-Hodgkin lymphoma across time, including a LOWESS curve. ART start was defined as the first date at which there was concurrent use of at least three different antiretroviral medications. CD4 counts at cancer diagnosis were defined as the most recent values within 6 months before cancer diagnosis. HIV RNA measurements at cancer diagnosis were defined as the most recent values 3 to 9 months before cancer diagnosis. HIV RNA values below the detection limit were randomly assigned a value between the detection limit and zero. Although cancers were ascertained starting in 1996, the first patient with non-Hodgkin lymphoma in the Centers for AIDS Research Network of Integrated Clinical Systems with an HIV RNA measurement 3 to 9 months before diagnosis was observed in 1998.

Fig 2.

Proportion of person-time contributed to each antiretroviral therapy (ART), CD4 count, and HIV RNA category by calendar period among patients with HIV in routine clinical care. Person-time categories are based on follow-up of the entire Centers for AIDS Research Network of Integrated Clinical Systems HIV population in routine clinical care and were time updated to account for changes in (A) ART status, (B) CD4 count, and (C) HIV RNA values.

Fig A1.

Antiretroviral therapy (ART), immune, and virologic status of patients with HIV in routine clinical care who were diagnosed with diffuse large B-cell lymphoma (DLBCL). (A) The proportion of DLBCL patient cases that occurred at various times relative to ART initiation over three calendar periods. (B) The distribution of CD4 counts at diagnosis of DLBCL across time, including a LOWESS curve to illustrate the smoothed trend over time. (C) The distribution of HIV RNA values on the log₁₀ scale at diagnosis of DLBCL across time, including a LOWESS curve. ART start was defined as the first date at which there was concurrent use of at least three different antiretroviral medications. CD4 counts at cancer diagnosis were defined as the most recent values within 6 months before cancer diagnosis. HIV RNA measurements at cancer diagnosis were defined as the most recent values 3 to 9 months before cancer diagnosis. HIV RNA values below the detection limit were randomly assigned a value between the detection limit and zero. Although cancers were ascertained starting in 1996, the first patient with DLBCL in the Centers for AIDS Research Network of Integrated Clinical System with an HIV RNA measurement 3 to 9 months before diagnosis was observed in 1998.