CROI: Wasting and anaemia predict higher risk of death when starting ART in low-income countries

This article is more than 19 years old. Click here for more recent articles on this topic

Low haemoglobin levels and wasting caused by AIDS before starting treatment appear to be the strongest predictors of death in people receiving antiretroviral therapy (ART) in Senegal and Haiti, according to research presented at the Thirteenth Conference on Retroviruses and Opportunistic Infections in Denver earlier this month. An international cohort analysis also found that the risk of illness was almost two and a half times higher in the first year of treatment in developing countries.

Background

Despite numerous reports of successful antiretroviral therapy in low-income countries, clinicians and treatment advocates remain concerned by the high death rates being seen in patients who qualify for antiretroviral therapy when compared with higher income countries. In general, clinicians have explained the higher death rate by pointing to the advanced HIV disease of patients in their care, but three studies presented at the Thirteenth Conference on Retroviruses and Opportunistic Infections were able to provide much more precise information about the factors that predict an increased risk of death or illness after starting antiretroviral therapy.

Mortality risks in the Senegal cohort

Eric Delaporte of the University of Montpellier in France reported on predictors of death in patients initiating antiretroviral therapy in the Senegal national cohort. This was a prospective observational cohort that began in 1998, and follow-up was available on 404 patients who had received treatment for a median of four years. Fifty-five per cent had CDC stage C disease (the equivalent of WHO stage 4 disease) at the time they started treatment, and the median CD4 cell count at this time was 128 cells/mm3.

Death was ascertained by hospital chart or post-mortem interview with a relative.

Glossary

wasting

Muscle and fat loss.

 

p-value

The result of a statistical test which tells us whether the results of a study are likely to be due to chance and would not be confirmed if the study was repeated. All p-values are between 0 and 1; the most reliable studies have p-values very close to 0. A p-value of 0.001 means that there is a 1 in 1000 probability that the results are due to chance and do not reflect a real difference. A p-value of 0.05 means there is a 1 in 20 probability that the results are due to chance. When a p-value is 0.05 or below, the result is considered to be ‘statistically significant’. Confidence intervals give similar information to p-values but are easier to interpret. 

haemoglobin (HB)

Red-coloured, oxygen-carrying chemical in red blood cells.

hazard ratio

Comparing one group with another, expresses differences in the risk of something happening. A hazard ratio above 1 means the risk is higher in the group of interest; a hazard ratio below 1 means the risk is lower. Similar to ‘relative risk’.

low income countries

The World Bank classifies countries according to their income: low, lower-middle, upper-middle and high. While the majority of the approximately 30 countries that are ranked as low income are in sub-Saharan Africa, many African countries including Kenya, Nigeria, South Africa and Zambia are in the middle-income brackets. 

The cumulative probability of death was 7.2% at month 6, 11.7% at month 12, 17.4% by 24 months and 23.2% at 60 months, indicating that the risk of death was highest in the first year of treatment. Mortality was reduced from 12.5 deaths per 100 patient years of follow-up to 0.9 deaths per 100 patient years of follow-up.

The study found that a baseline body mass index below 19, a baseline haemoglobin below 10g/dl or a CD4 cell count below 200 were the only significant predictors of an increased risk of death.

Mortality risks in a Haitian community cohort

The first study reported on the GHESKIO cohort – 1,882 patients receiving antiretrovirals at community clinics in Haiti.

The study looked for predictors of early mortality in patients who had started antiretroviral therapy and who had been followed for up to one year.

The study found that only two factors remained significant predictors after adjusting for gender, CD4 cell count, development of TB and age.

These were the presence of AIDS wasting at the time of ART initiation (hazard ratio 2.36, p=0.001), and a haemoglobin below 9.5g/dl at the time of ART initiation (hazard ratio 2.14, p=0.003). Baseline CD4 cell count did not remain significant in the multivariate analysis.

A Kaplan-Meier analysis showed that almost a third of patients with AIDS wasting were likely to have died after twelve months of follow-up, compared to less than one in ten of those without wasting.

Commenting on the findings, a member of the research team from Cornell University in New York said: “These findings are important because they show that with just a haemoglobin measure and a scale we can identify the patients at high risk of dying during the first few months of treatment.”

Morbidity risks in the ART-LINC cohort collaboration

The largest analysis of outcomes presented at the meeting came from the ART-LINC cohort collaboration (Antiretroviral Treatment in Low Income Countries), which combines data on cohorts at 23 centres in developing countries and in Brazil, comprising information on 4540 patients. Dr Paula Braitstein of the University of Berne in Switzerland presented the findings on behalf of the ART-LINC group, and compared them to outcomes in the ART Cohort Collaboration, which links a large number of cohort studies in the developed world.

The ART-LINC analysis looked at the risk of developing opportunistic infections rather than death after starting ART, but its results are a useful complement to the findings on predictors of mortality from Haiti and Senegal.

The study found that the risk of death was approximately four times higher in the first year of antiretroviral treatment in the ART LINC compared with the ART Cohort Collaboration (ART CC). However the rate at which the risk of death declined during the first year of treatment was very similar between the two cohorts.

Unlike the studies looking at predictors of death, the ART-LINC study also found that baseline CD4 count influenced the risk of illness on ART, as did the type of drug regimen use. Patients who took NNRTIs were significantly less likely to develop an opportunistic infection on treatment. However, a previous history of OIs, whether off treatment or on treatment, significantly increased the risk of illness.

ART LINC

ART LINC

ART CC

Patients

4540

12,574

Person years of follow-up

3255

8709

No of clinical events

689

710

Event rate per 1000 years of follow-up

212

82

Most frequent opportunistic infections on ART

  1. TB
  2. KS
  3. Candidiasis
  4. Bacterial pneumonia
  5. Cryptococcus
  1. MAI
  2. KS
  3. CMV
  4. PCP
  5. TB

New HAI

Adjusted hazard ratio

95% CI

New HAI

in first year

post-ART

P value

New HAI

in m 1-3

post-ART

P value

Sex female

0.75

0.001

0.88

0.266

Baseline CD4

25-49

50-99

100-199

200-350

>350

1.0

0.75 (0.59 –0.96)

0.85 (0.59 –1.06)

0.61 (0.49- 0.77)

0.65 (0.50 –0.84)

0.30 (0.19-0.48)

1.0

0.60 (0.42-0.85)

0.72 (0.54-0.97)

0.59 (0.44-0.80)

0.49 (0.33-0.72)

0.21 (0.10-0.46)

Initial ART

NNRTI-based

PI-based

Other triple

1.00

1.36 (0.97-1.91)

2.00 (1.45- 2.77)

OI history (vs none)

Pre-ART

During ART

2.93 (2.07-4.16)

2.59 (2.13-4.15)

References

Braitstein P et al. When are adults in resource-limited settings most likely to experience an HIV-associated illness following HAART initiation and what is it related to? Thirteenth Conference on Retroviruses and Opportunistic Infections, Denver, abstract 67, 2006.

Etard JF, Delaporte E et al. Mortality and causes of death in adults receiving HAART in Senegal: a 7-year cohort study Thirteenth Conference on Retroviruses and Opportunistic Infections, Denver, abstract 556, 2006.

Dillingham R et al. Predictors of early mortality in Haitian patients treated with antiretroviral therapy in a community setting. Thirteenth Conference on Retroviruses and Opportunistic Infections, Denver, abstract 556, 2006.