University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Excessive early mortality in the first year of treatment in HIV type 1-infected patients initiating antiretroviral therapy in resource-limited settings
Global Health Sciences Literature Digest
Published August 07, 2008
Journal Article

Marazzi MC, Liotta G, Germano P, Guidotti G, Altan AD, Ceffa S, et al. Excessive early mortality in the first year of treatment in HIV type 1-infected patients initiating antiretroviral therapy in resource-limited settings. AIDS Res Hum Retrovirus 2008 March 26; Epub.


To evaluate risk factors for early mortality in a cohort of patients initiating highly active antiretroviral therapy (HAART)

Study Design

A retrospective cohort study of patients initiating antiretroviral treatment (ART) in the Drug Resource Enhancement against AIDS and Malnutrition (DREAM) program, developed by an Italian faith-based charity for the provision and dissemination of the best available practices in HIV medicine in resource-limited settings


Patients accessing health care at the DREAM clinics of Benfica, Machava, Polana Caniço, Chokwe, Beira, Quelimane, Mangunde (Mozambique), Lilongwe and Blantyre (Malawi), and Arusha and Iringa (Tanzania)


Subjects were 3,456 patients (2,078 women and 1,378 men) receiving care at DREAM clinics from January 2003 to December 2005 who met the following criteria: 1) Age greater than 14 years. 2) No pregnancy during the study period. 3) Availability of pre-HAART CD4 cell count and HIV-1 virus load results. 4) At least 180 days of follow-up since HAART initiation for patients on ongoing treatment. 5) Inclusion of all patients with less than 180 days of follow-up since HAART initiation who ceased medical care because of death or loss to follow-up. 6) Indication for starting HAART: clinical or laboratory criteria for initiation of ART, a WHO clinical stage of 3 or 4, and a CD4 cell count of <200 cells/mm3 or between 200 and 350 cells/mm3 if virus load at baseline was >55,000 copies/ml. HAART was indicated according to WHO international guidelines.(1,2)


Data were retrospective clinical records accessed by DREAM software with follow-up censored at 1 year after start of HAART (up to 30 June 2006) or at last visit for patients who died or were lost to follow-up. Censoring time was 365 days for more than half of the patients. Treatment regimens of the retrospectively included patients were based on triple ART containing two nucleoside reverse transcriptase inhibitors (zidovudine and lamivudine or stavudine and lamivudine) and nevirapine. In case of allergy or intolerance to nevirapine, patients were switched to nelfinavir. All centers consistently followed treatment protocols developed by the DREAM program. Patients were actively traced by peer-to-peer educators, who also were patients on treatment, to increase and facilitate adherence to ART and nutritional supplementation. Food packages were delivered periodically to patients with a body mass index (BMI) value of less than 18. Cotrimoxazole use: All patients with CD4 cell counts of <200cells/mm were given cotrimoxazole. Screening for coinfections: Screening for comorbidities at baseline was not routinely performed at the clinics unless patients were symptomatic. If patients presented with symptoms of tuberculosis, screening evaluations included sputum smears and chest X-rays. Laboratory evaluations: Liver function tests and hemoglobin levels at baseline; start of HAART; 4, 8, and 12 weeks post-initiation of treatment and every 3 months thereafter; CD4 cell subsets every 3 months; and virus loads twice a year. The laboratory procedures among the locations were standardized and of high quality.(3) The risk for every single variable was estimated stratified for the dichotomized level of a CD4 cell count (>200 or <200) and pooled using the Mantel-Haentzel procedure. The hazard ratio for time-to-event outcomes, such as death, was estimated using the Cox proportional hazard regression mode. All covariates were included in the model as categorical except for age, which was considered a numeric variable.

Primary Outcomes

Mortality in the first year after starting HAART


The median age of patients was 36.5 years (34.5 for women and 39.4 for men). Mean pre-HAART values were hemoglobin, 10.5 g/dl (interquartile range [IQR] 9.0-11.9); BMI, 20.6 (18.4-23.1), CD4 cell count, 166 cells/µl (80-257); and viral load, 4.75 copieslog/ml plasma (IQR: 4.11-5.17). Seventy-two percent of patients (2483) were in WHO clinical stages 3 or 4. Before starting HAART, 219 of 3137 patients (7%) had a viral load of <400 copies/ml and 1305 of 3434 patients (38%) had a CD4 cell count of >200/µl. Viral load and CD4 cell count values between 6 months and a year before the beginning of treatment were available for approximately 60% of patients. After 6 months of treatment, the percentage of patients with a viral load of <400 copies/ml and with a CD4 cell count of >200/µl rose, respectively, to 77.1% (1628 of 2112; CL 95%: 75.2-78.8) and 78.9% (1666 of 2112; CL 95%: 77.1-80.5). The mean CD4 cell count value increased from 243 (SD±338) before starting HAART to 375 (SD±227) after 6 months of HAART. The viral load mean logarithmic value decreased from 5.02 (SD±5.26) before starting HAART to 4.01 (SD±4.73) after 6 months of HAART. HAART was well tolerated overall, with no significant toxicities or adverse events.

In the first year of HAART, 260 deaths were recorded (97 per 1,000 person/years). The mortality rate observed in the cohort was four times greater than that observed among patients in high-income countries. The mortality rate peaked in the first 3 months of treatment, during which more than half the deaths observed occurred (137/260; 53%; 164.3 per 1,000 person/years; CI 95%=140.8-191.2). In the last quarter of the first year of treatment, the mortality rate dropped to 13.4 per 1,000 person/years (21/260: 8.1%; CL 95%=9.1-19.8). When the mortality rate was stratified according to CD4 cell values, it was 10% in subjects with <200 cells/µl, 4.8% in subjects with >200 and <350 cells/µl, 4.4% in subjects with >350 and <500 cells/µl, and 2.8% in subjects with >500 cells/µl. Forty-one (1.2%) cases were lost to follow-up. The cause of death could be ascertained in only 47% of patients, given that the vast majority of deaths occurred at home. In nearly half of the patients who died, however, the major cause of death was malaria, anemia, or an association of the two, and the next greatest cause was tuberculosis. Ninety-three percent of the patients who died had CD4 cell counts of <200 cells/µl at treatment initiation, suggesting that advanced HIV disease played the major role.


The authors surmise that the excess mortality in this study population of patients in sub-Saharan Africa could be explained only partially by the lower CD4 counts observed. They say that other comorbidities, such as WHO clinical stage, malnutrition, and anemia, play a critical role. The authors suspect that in many cases, HAART was started too late in a population already highly afflicted by malnutrition, anemia, and an overwhelming burden of disease. To reduce mortality in resource-limited settings, the authors conclude, it is necessary to have a more aggressive approach to treatment and to initiate it earlier. The role of specific opportunistic infections in the study could not be ascertained clearly because screening for coinfections was not routinely done in the DREAM setting unless patients were symptomatic. In addition, the authors were able to ascertain the cause of death in only 47% of patients. In nearly half of the patients who died, however, the major cause of death was malaria, anemia, or an association of the two, with the next greatest cause of death being TB. This finding indicates that in many instances patients died due to a prevailing illness.

Quality Rating

This retrospectively analyzed observational cohort study was of a fair quality, receiving 6 points of the 9 possible in the Newcastle-Ottawa quality assessment scale for cohort studies. It might have received more points, but the authors did not study a cohort or comparison group being treated in a "high-income country," giving no details for the assertion that mortality was four time higher in their cohort than in patients initiating HAART in settings that were not resource limited.

In Context

(Reviewer comment): This study's findings are comparable to those of other investigators who have also observed excessive mortality in patients initiating HAART in resource-limited countries as compared to countries that are not resource limited. (4,5,6) These differences cannot be explained by lower adherence to therapy, considering that results in terms of reduction of viral load and increase in CD4 cell count were comparable to those attained in developed countries. Even when this study's data are compared to those of mid-developed countries like Brazil,(7) one observes a higher early mortality rate post-HAART in the African setting. The elevated prevalence of infections, such as, for example, malaria, or pathological states such as malnutrition, could explain this observation and could justify a more aggressive treatment approach. Avoiding a decline in CD4 cell count below a threshold of 200/µl and considering the initiation of HAART in patients with a BMI level of <18 and/or with a hemoglobin level of <8, even in the presence of CD4 counts between 200 and 350/µl, could represent an effective response for reduction of mortality. Delay in receiving treatment may be the crucial factor in these higher mortality rates.

Programmatic Implications

Patients with HIV in resource-limited settings still should begin HAART in accordance with WHO treatment guidelines. If they meet the WHO criteria for beginning HAART, they should not delay in doing so. In addition, it is critical that treatment protocols for anemia and malnutrition be developed for HIV-infected patients in this setting.


  1. OARAC-DHHS Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines of Antiretroviral Agents in HIV-1-Infected Adults and Adolescents.Accessed July 8, 2006.
  2. WHO Scaling up antiretroviral therapy in resource-limited settings: Treatment guidelines for a public health approach. Accessed July 8, 2006.
  3. Marazzi MC, Guidotti G, Liotta G, and Palombi L: DREAM: An integrated faith-based initiative to treat HIV/AIDS in Mozambique.
  4. Lawn SD, Myer L, Bekker LG, and Wood R: Tuberculosis-associated immune reconstitution disease: Incidence, risk factors and impact in an antiretroviral treatment service in South Africa. AIDS 2007;1:335-341.
  5. Sobhani R, Basavaraj A, Gupta A, et al.: Mortality and clinical characteristics of hospitalized adult patients with HIV in Pune, India. Indian J Med Res 2007;126:116-121.
  6. Losina E, Yazdanpanah Y, Deuffic-Burban S, et al.: The independent effect of highly active antiretroviral therapy on severe opportunistic disease incidence and mortality in HIV-infected adults in Côte d'Ivoire. Antivir Ther 2007;12(4):543-551.
  7. Ivers LC, Kendrick D, and Doucette K: Efficacy of antiretroviral therapy programs in resource-poor settings: A meta-analysis of the published literature. Clin Infect Dis 2005;41:217-224.