Prospective Evaluation of the Impact of ART on the Incidence of KS—IeDEA Consortium: East Africa
Jeffrey Martin*1, M Wenger1, N Busakhala2, N Buziba2, M Bwana3, W Muyindike3, E Amerson1, C Yiannoutsos4, B Musick4, K Wools-Kaloustian4, and East Africa Intl Epidemiologic Databases to Evaluate AIDS (IeDEA) Consortium
1Univ of California, San Francisco, US; 2Moi Univ, Eldoret, Kenya; 3Mbarara Univ of Sci and Tech, Uganda; and 4Indiana Univ, Indianapolis, US
Background: Prior to the rollout of potent ART, Kaposi’s sarcoma (KS) was the most commonly reported malignancy in many resource-limited settings, such as most of sub-Saharan Africa. In resource-replete settings, the advent of ART—and the availability of established research networks—resulted in documentation of a marked decrease in KS incidence. In contrast, in resource-limiting settings such as Africa, the heretofore lack of epidemiologic infrastructure has limited our knowledge about the effect of ART on KS incidence.
Methods: We prospectively followed HIV+ adults, without prior KS, attending 26 HIV clinics at the Mbarara (Uganda) or AMPATH (Kenya) sites of East Africa IeDEA. Time 0 was October 2008, when we introduced skin punch biopsy for KS diagnosis. Patients were followed until incident KS, death, loss-to-follow-up, or administrative closure. Once started on ART, patients were considered on ART irrespective of adherence. Incident KS was defined as any new occurrence of KS 30 days after clinic enrollment that was documented either pathologically or by clinical diagnosis in the absence of pathology.
Results: We followed 98,024 HIV+ adults: 31% men, 66% ever on ART, and median values at study enrollment of 35 years old (IQR 29 to 43) and 277 CD4+ T cells/mm3 (IQR 137 to 453). Patients were followed collectively for 144,182 person-years (median 1.8 years/patient) for 499 incident KS diagnoses, 43% of which were pathologically confirmed. KS incidence during non-ART use was 1876 cases/100,000 person-years in Uganda and 596 in Kenya; incidence during ART use was 201/100,000 person-years in Uganda and 270 in Kenya. After adjustment for age and gender, ART users had a substantially reduced rate of KS compared to non-users: 88% reduction in Uganda (p <0.001), and 50% in Kenya (p <0.001). Further adjustment by CD4+ count showed a persistent ART effect, suggesting ART benefits for KS prevention above and beyond CD4 restoration. Among ART users achieving a CD4+ count of >350 cells/mm3, KS incidence declined to 18/100,000 person-years in Uganda and 93/100,000 person-years in Kenya.
Conclusions: In a prospective study in East Africa, KS incidence was very high in untreated HIV+ adults but has been substantially reduced by ART—similar to that observed in resource-replete settings. Despite ART, absolute rates of KS remained considerable until a CD4+ count of >350 was achieved, suggesting the need for earlier ART initiation.