Differentiated service delivery (DSD) has highlighted the importance of developing diverse, patient-centered models of care for stable patients on antiretroviral therapy (ART). However, similar innovations are not yet available for PLHIV with advanced disease who are at high risk for poor outcomes and have received little attention beyond recognition of their need for prompt ART initiation.
We analyzed data on PLHIV ≥15 years starting ART with CD4 cell count ≤200 cell/mm3 at PEPFAR-funded ICAP-supported clinics in Ethiopia, Kenya, Mozambique and Tanzania 2004-2015. PLHIV with low CD4 at start of ART were divided into 3 groups: CD4 101-200, CD4 50-100 and CD4<50 cell/mm3 to examine a combined endpoint of attrition (loss to follow up (LTF) and death) following ART start using Kaplan-Meier estimators and log-rank tests. LTF after ART initiation was defined as no visit >6 months; deaths were ascertained from medical records. The combined endpoint was used as deaths are under-recorded. Cox proportional hazards models examined the effect of CD4 at ART start on attrition adjusted for sex, age, country and intrasite clustering.
477,086 PLHIV started ART at 350 clinics, of whom 203,622 (42.7%) had CD4 ≤200 cell/mm3 at ART start. Among these, 60.2% were female, median age was 35.0 years [IQR: 28.9-42.1], and 60.7% were WHO stage III/IV. Among those with CD4 <50 cell/mm3, 70.6% were stage III/IV compared to 52.9% of those with CD4 101-200(p<0.0001). Overall, for patients with CD4 ≤200, attrition at 3, 6 and 12 months was 14.8% (95%CI 14.6-14.9), 19.7% (95%CI 19.6-19.9) and 25.1% (95%CI 24.9-25.3), respectively. Attrition was significantly higher with lower CD4 counts; at 12 months 32.2% (95%CI 31.8-32.6) of those with CD4<50 cell/mm3 at ART start were lost or had died compared to 20.8% (95%CI 20.6-21.1) with CD4 101-200 cell/mm3 (p<0.0001)(Table 1). In multivariable models, PLHIV with CD4<50 cells/mm3 had 40% increased risk of attrition compared to those with CD4 101-200 cell/mm3 (adjusted hazard ratio 1.4, 95%CI 1.3-1.4).
Even after ART initiation, PLHIV with advanced disease had notably inferior outcomes within lower CD4 strata. With the transition of stable patients to less intensive DSD models requiring less provider time, the opportunity exists to focus on optimizing outcomes for PLHIV with advanced HIV disease. Novel DSD models that include biomedical, psychosocial and structural interventions are urgently needed.