Optimalisation of HIV care in West Africa

Thesis summary

This thesis focuses on opportunities to improve HIV care in West Africa. Although the roll-out of antiretroviral therapy (ART) in sub-Saharan Africa has been highly successful HIV care remains sub-optimal and tools for its betterment are either now, or will soon be, available. Key constraints specific to West Africa that need to be addressed are the presence of HIV-2 and the limited resources available for care. Most of the research presented in this thesis took place at the Medical Research Council’s (MRC’s) Gambia Unit, specifically the Genito-Urinary Medicine (GUM) Clinic where from October 2004 to June 2010 ART was available as part of a national program supported by the Global Fund for AIDS, TB, and Malaria (GF-ATM). The Gambia has a population of approximately 1.8 million inhabitants. Over 90% of the population is Muslim, and most are engaged in subsistence agriculture. During the time of the studies the HIV prevalence was approximately 2%, including 0.5% HIV-2 infected.

The MRC’s GUM clinic served as a tertiary referral center for HIV care for nearly 25 years. In 1998 it became a platform for clinical and epidemiological research. Following the establishment of the national HIV treatment program in 2004 ART became available. Prior to ART initiation patients were seen quarterly. After starting ART they were seen every two weeks, thereafter monthly, and eventually again quarterly. CD4 counts and HIV viral loads were tested routinely following clinic protocols. Additional visits, treatments, and diagnostics were carried out according to the judgment of the treating physician. Although HIV-infected children were also seen in the GUM clinic, only the data from adults were used for the studies upon which this thesis is based.

Main results

Mortality reduction pre-ART

A well-considered haste at every step prior to the initiation of ART is the key to reducing HIV mortality. This can be addressed by adapting care protocols to hasten provision of opportunistic infections (OI) prophylaxis, including isoniazid preventive therapy, and ART. Loss to follow-up needs to be, and can be, reduced. Several recent developments in HIV clinical care are ready for implementation, such as the GeneXpert™ test to diagnose tuberculosis, and the cryptococcal antigen tests to screen for cyptococcosis. Additional technological developments are in the pipeline whose rapid adoption will further reduce HIV mortality, including point of care testing for cryptococcal antigen, CD4 counts, viral load, and ARV resistance mutations.

Predicting the immune reconstitution inflammatory syndrome (IRIS)

Unlike untreated HIV infection, IRIS is rarely fatal. Nonetheless it is a common cause of morbidity in people initiating ART at low CD4 counts. Diagnosis is currently clinical, and it is likely that IRIS represents diverse pathogenic 167 processes. Early diagnosis and treatment of IRIS is impeded by a lack of clear understanding of the immunological processes involved. We assessed a panel of inflammatory markers, IL-2, IL-6, IFN-γ, TNF-α, MIP-1β, IL-1, IL-12, IL-13 and IL-10, and T-regulatory cell recovery, for their ability to predict IRIS. None of these parameters predicted IRIS, however.

HIV-2 ART recommendations

Antiretroviral combinations consisting of three nucleoside analogs, and those using unboosted protease inhibitors, previously common in the treatment of HIV-2 in low-income settings, performed poorly. We demonstrated that virologic suppression on a par with that achieved in HIV-1 was possible using ritonavir-boosted lopinavir in first line ART for individuals with HIV-2 infection. There is a clear need for effective ARVs from “new” classes for the treatment of HIV-2. When the integrase inhibitor raltegravir, first licensed for use in HIV-1 infection, was used in HIV-2 this yielded mixed, largely negative results despite in vitro evidence of efficacy. A case series presented in this thesis demonstrates that if it is used in conjunction with an optimized background regimen of sufficient potency, raltegravir can generate a durable viral suppression despite its low threshold for developing resistance. Whether its proper place in the current armamentarium is in first- or second-line therapy remains an open question.

Literacy, education, and adherence

We found lower levels of formal education and literacy to be associated with treatment failure mediated through poorer ART adherence. This suggests that literacy and/or education should be assessed prior to ART initiation; identifying those at risk for poor adherence allows them to be targeted for greater adherence support.

Predicting viremia with self-reported adherence and keeping to appointments

Suboptimal antiretroviral adherence typically precedes HIV viremia and subsequent immunologic decline. We demonstrated that self-reported adherence was modestly effective at predicting viremia. Missed appointments were shown to predict viremia across a wide range of days by which they were missed, while not adding predictive power to self-reported adherence in multiple regression analysis. This suggests that in our setting missing appointments is a characteristic of non-adherent patients, not the cause of their missed doses.

Depression screening

In a first study we screened 44 individuals not yet on ART using the CES-D10 score for depression and the IES-R score for post-traumatic stress disorder (PTSD). With over 40% of the subjects testing positive for either condition, this study suggests a serious burden of disease. Depression was associated with low CD4 counts and not having an independent income. The wellestablished relationship between depression and poor ART adherence argues for incorporating depression screening into the process of preparing individuals 168 for ART. In a second study we performed depression and PTSD screening among 252 people on ART. This study revealed a strikingly lower prevalence of depression compared to the study of people not yet on ART. This is in contrast to the results of PTSD screening. We interpret this as a response to ART through reduction in symptoms and short-term risks of illness and death. It is however also possible that the lower prevalence of depression was the result of higher mortality among depressed patients rather than recovery from their depression.


Confirming and staging an individual’s HIV infection in a timely manner, including early CD4 testing, is the first step towards improving care. Screening for illness and risk factors for a poor treatment response is a key element of pre-ART care, and can be roughly grouped into three categories. The first is screening for the classic OIs such as tuberculosis and cryptococcal meningitis, which are a substantial source of morbidity and mortality both directly and through IRIS. The second category of screening includes those conditions that negatively affect ART adherence such as financial and social barriers to treatment, illiteracy, substance abuse, and depression. The third category of screening focuses on those conditions that promote HIV transmission. Addressing transmission control through sexual transmitted infection (STI) and pregnancy screening, safe sex counseling, condom provision, and so forth are vital elements of HIV care, although not addressed in this thesis. A general lesson I draw from my research and experiences during my tenure at the GUM clinic is the value of investing in high quality patient care. Reliable diagnostics, effective counseling, and good care and follow-up are expensive and time consuming. I assert however that the short-term costs are largely outweighed by the long term benefits.