More information on causes of death needed to fine-tune ART services in Africa

Knowing the causes of death as well as mortality rates among patients on antiretroviral therapy in Rwanda may help improve service delivery, Innocent Turate and colleagues reported in a study presented at the HIV Implementers’ Meeting in Namibia earlier this month.

Loss of patients to follow-up continues to be a significant problem for treatment programmes in many parts of Africa, but measures to improve patient retention in care require a better understanding of why patients are lost to follow-up, and in particular, the number of deaths and the causes of death among those who start treatment.

A retrospective analysis from 2003 to 2008 was undertaken and included a chart review of 11,785 patients on antiretroviral therapy (ART) in 29 Family Health International (FHI)-supported sites in Rwanda, where Family Health International is involved in the delivery of treatment and care funded by the US PEPFAR programme.

During the period under review 592 patients died (5.8% of all those treated). The cause of death was determined by medical file and chart review and post-mortem interview with relatives. Data collected included date of starting ART, baseline CD4 count, baseline WHO clinical stage, cause of death if known and place of death (home, health centre, hospital).

With increased coverage (from two sites in 2003 to 29 in 2008) the number of deaths appears to have stabilised at around 150 for each of the last three years, whereas the numbers lost to follow-up increased from three in 2003 to 143 in 2008.

The median time between beginning ART and death was two months. Death registration forms with additional information were available for 73.3% (434). Of these, close to half died at home with CD4 counts available for over 90%. Just under 80% of those had CD4 counts below 200 cells/mm³ (the median was 109 cells/mm³) at the time of treatment initation, compared to 52.4% for all ART patients. The vast majority (83%) were identified as having WHO clinical stage 3 or 4 HIV disease at baseline.

Cause of death was determined for 66.4%. The most frequent known cause of death was tuberculosis and non-tuberculosis confirmed pulmonary disease (28%) followed by chronic diarrhoea at 8.5%. Eighty per cent of all deaths with unknown causes occurred at home.

Close to 80% of patients died within the first six months of starting treatment and death was strongly associated with a low CD4 count.

The death rate of 5.8% is similar to rates seen in Europe and North America. In contrast to such high-income countries non-HIV-related causes of death, for example, cardiovascular disease and diabetes, drug-related toxicity and cancers did not appear to be a major cause of death in Rwanda.

Limitations of the study cited by the author include the retrospective nature of a study that relied on a recording and reporting system that lacked codification and classification of causes of death, and the absence of documentation for the high proportion of deaths that occurred in the home.

The authors suggest that in light of the number of deaths due to non-TB confirmed pulmonary disease, and the high proportion of patients with low CD4 counts, an Immune Reconstitution Inflammatory Syndrome (IRIS) diagnosis is relevant and recommend strengthening clinician skills enabling them to identify and manage IRIS as well as other critical conditions in patients with low CD4 counts.

With TB identified as the leading cause of death, maintenance and reinforcement of the integration of TB/HIV activities was also recommended. In addition, strategies are required to enable earlier initiation of ART and intensive follow-up of those with low CD4 counts to help address unknown causes of death as well as improve service delivery (FHI has just implemented a new strategy for the latter called Acuity Case Management).

With increased life expectancy due to ART, assessment of non-AIDS defining diseases is suggested. The authors also recommended that the Ministry of Health consider making the systematic reporting of causes of death standard policy at the health facility level, together with the institution of a national death registry.