Secondary Logo

Journal Logo

Epidemiology and Social Science

Anemia Is an Independent Predictor of Mortality and Immunologic Progression of Disease Among Women With HIV in Tanzania

O'Brien, Megan E PhD*; Kupka, Roland PhD*; Msamanga, Gernard I MD, ScD; Saathoff, Elmar PhD*; Hunter, David J ScD*; Fawzi, Wafaie W MD, DrPH*

Author Information
JAIDS Journal of Acquired Immune Deficiency Syndromes: October 1st, 2005 - Volume 40 - Issue 2 - p 219-225
doi: 10.1097/01.qai.0000166374.16222.a2


Anemia is the most frequent hematologic abnormality of HIV disease1-3 and one of the most common manifestations of nutritional deficiency disorders in the world.4 Among HIV-infected individuals from the United States and Europe, anemia has been independently associated with mortality,2,3,5-10 progression to AIDS,11,12 and decreased quality of life.13 In turn, recovery from anemia has been associated with improved survival in HIV disease.3,5 Low hemoglobin levels have been associated with advanced HIV stage and immunosuppression14,15 and the incidence of anemia increases with advancing HIV disease.1,16-18

Potential causes of anemia in the context of HIV disease include HIV infection of hematopoietic stem cells/erythroid progenitors, immune system-mediated hemolysis, aplastic anemia, malignancies, blood loss, bone marrow infections, and deficiency of erythropoietin.19 Deficiencies in the nutrients vitamin B12, folate, and iron also contribute to the development of anemia, though to a much lesser extent among patients in industrialized countries.20 Rather, HIV-associated anemia typically resembles anemia of chronic disease,21 with hypoferremia, adequate iron stores, normal or elevated ferritin, usually moderate, normocytic, and normochromic in appearance, often hyporegenerative and refractive to iron supplementation.19,22 Low serum iron in patients with HIV-related anemia is attributed to altered iron metabolism and not to iron deficiency per se.19

In sub-Saharan Africa, home to 70% of the world's people living with HIV/AIDS,23 the prevalence of anemia is higher than in developed countries,4,18 and anemia is commonly caused by nutrient deficiencies, sickle cell disease, AIDS, malaria, hookworm infection, and other infections.4,24,25 Prevalence of anemia is generally higher in women than men4,18 and among individuals with HIV compared with those who are HIV negative.26,27 Reported prevalences of anemia range from 50%-60% among pregnant African women26,28,29 and are approximately 70%-80% among those with HIV infection.26,30-32

Few studies have examined the association of anemia with mortality and HIV disease progression among women in this region. The purpose of this study was to examine the role of anemia in mortality and immunologic progression among a cohort of women with HIV disease in Dar es Salaam, Tanzania.


Study Population

Between April 1995 and July 1997, 1078 HIV-positive pregnant women between 12-27 weeks' gestation, 98% of whom had World Health Organization (WHO) clinical stage 1 or 2 disease (not progressed to AIDS), were enrolled into the Trial of Vitamin Supplementation Study in Dar es Salaam, Tanzania. Details of the design and methods are described elsewhere,33 but in brief, the trial was a 2-by-2 factorial design in which women received vitamin A, multivitamins including vitamin A, multivitamins excluding vitamin A, or placebo. In accordance with local guidelines for antenatal care, all women received 400 mg of ferrous sulfate (equivalent to 120 mg of ferrous iron) and folate (5 mg) for anemia prophylaxis and weekly doses of 500 mg of chloroquine phosphate (equivalent to 300 mg of chloroquine base) for malaria prophylaxis during pregnancy. At the time of randomization and at monthly visits, women underwent physical examinations and completed interviews about interim medical problems. Blood cell counts including total leukocyte counts, differential leukocyte counts, and absolute T-cell subset counts and blood films were done at baseline, 6 weeks and 30 weeks postpartum, and then every 6 months thereafter. Women were followed up either until they died or were lost to follow-up or until the study closed in August 2003.

Mortality Information

A study nurse visited the home of women who did not attend the clinic or who traveled out of Dar es Salaam and contacted the woman, or neighbors and relatives in the area, to collect information on survival status. For women who died, data to determine the approximate cause of death were collected using verbal autopsy techniques. Nurses reviewed medical records and interviewed relatives to complete structured questionnaires describing the clinical condition of the woman in the days preceding her death. Two independent senior Tanzanian internists coded the verbal autopsy data and met at a later point to resolve discrepancies in their assignments.

Laboratory Methods

Hemoglobin was measured using a CBC5 Coulter counter (Coulter Corp., Miami, FL) or using the cyanmethemoglobin method with a Colorimeter (Corning, Corning, NY). Thin blood films with Leishman stain were examined by trained laboratory technicians, who coded the extent of the following erythrocyte characteristics: anisocytosis, poikilocytosis, hypochromasia, hyperchromasia, microcytosis, macrocytosis, normochromasia, and normocytosis. The prevalence of each characteristic among the sample of cells was coded by the technicians as being either absent, <25%, 25%-50%, 50%-75%, or >75%. CD4 cell counts were enumerated with flow cytometry (FACScount or FACScan systems, Becton-Dickinson, San Jose, CA).

Statistical Methods

Mean hemoglobin levels at baseline were approximately normally distributed (Shapiro-Wilks P value 0.08). To describe differences in hemoglobin levels at the time of enrollment among demographic and clinical groups, we conducted pairwise comparisons using the Tukey honestly significant difference (HSD) method to adjust for multiple comparisons, where appropriate.

Cox proportional hazards regressions were used to examine the relationship of anemia to all-cause mortality and AIDS-related mortality. Observations were censored for participants in whom the survival status was not known, and AIDS-related death was censored for participants in whom the cause of death was not known to be AIDS related. Hemoglobin, CD4 cell count (categories: missing, <100, 100-199, 200-299, 300-399, 400-499, and >499 cells/mL), WHO clinical stage of disease (stage 4 vs. stages 1-3), age (years), pregnancy (yes vs. no), and body mass index (categories: <20, 20-22, 23-25, >25 kg/m2) were included in the models as time-dependent covariates, and treatment arm in the vitamin study was included as a time-fixed covariate. For the Cox proportional hazards regression models, time-dependent covariate values were carried forward for up to 1 year if no new measures were available. After 1 year, the value for the measure was set to missing until the next date it was measured except for CD4 cell count, which was set to a missing indicator category. We used this method to maximize the data available for the model but to avoid allocating person-time to categories of a covariate that were deemed too far in the past to be clinically relevant. The result of this censoring is that for participants with some missing data, the participant is retained in the analysis, but periods of their follow-up were excluded if the covariate information was missing.

Cox proportional hazards regression models were also used to examine the relationship between anemia and HIV disease progression, measured as the time to a 50% drop in CD4 cell count from the baseline value. Hemoglobin, WHO clinical stage of disease, age, pregnancy, and body mass index were included in the models as time-dependent covariates and baseline CD4 cell count category (categories: missing, <350, and ≥350 cells/mL) and treatment arm in the vitamin study were included as a time-fixed covariates. Hemoglobin was categorized at 3 cutoff points for anemia; severe anemia (<8.5 g/dL), moderate anemia (hemoglobin 8.5-10.9 g/dL), and normal (≥11.0 g/dL) to reflect levels of anemia used as criteria for referral to district hospitals in Tanzania.29

Erythrocyte morphology was grouped into categories of hypochromasia and microcytosis, characteristics commonly caused by iron deficiency.34 These categories were hypochromasia with microcytosis, hypochromasia without microcytosis, any other abnormality in morphology, and normal (normocytic and normochromic). The models described here were run again, but with substitution of morphologic categories for hemoglobin.

All analyses were conducted with SAS v.8 (SAS Institute, Cary, NC). Tests of significance were 2-sided with a probability cutoff value of 0.05.

All women provided informed consent for data collection and the study was approved by the ethical review boards of the Harvard School of Public Health, the National AIDS Control Program of the Tanzanian Ministry of Health, and the Muhimbili University College of Health Sciences in Dar es Salaam, Tanzania.


Of the 1078 women enrolled in the trial, 1062 had a hemoglobin value measured at baseline and were included in the cross-sectional analysis. Four women (0.4%) did not have any hemoglobin measured at baseline or during follow-up and were excluded from all analyses. The median follow-up time available for the 1074 women in these analyses was 5.9 years [interquartile range (IQR): 3.8 to 6.7 years].

At enrollment into the trial, 55.1% of women had moderate anemia (hemoglobin 8.5-10.9 g/dL) and 27.7% had severe anemia (hemoglobin <8.5 g/dL) (Table 1). The mean hemoglobin level was 9.4 g/dL (SD: 1.67). Mean hemoglobin levels were significantly higher for patients with greater mid-upper arm circumference, lower gestational age, lower erythrocyte sedimentation rate, greater total lymphocyte count, greater plasma vitamin A levels, erythrocyte characteristics that were not suggestive of iron deficiency, greater mean corpuscular volume, greater packed cell volume, and no evidence of malaria infection.

Factors Associated With Hemoglobin Level at Baseline (n = 1062 women)
(continued) Factors Associated With Hemoglobin Level at Baseline (n = 1062 women)

A total of 343 deaths occurred during the duration of follow-up. Information to determine whether death was AIDS related was lacking for 74 women (21.6%). Deaths of 243 women were deemed AIDS related and were attributed to AIDS (n = 82), anemia (n = 10), stroke (n = 5), meningitis (n = 14), pneumonia (n = 23), diarrhea (n = 21), fever (n = 24), pulmonary tuberculosis (n = 61), and extrapulmonary tuberculosis (n = 3). Fifty-five of the participants who died (16.0%) had no hemoglobin values available within 1 year of death, so person-time for the interval that ended with death was not included in the Cox model, although the periods of follow-up in which hemoglobin information was available were included. This resulted in 288 all-cause deaths and 204 AIDS-related deaths included in the models.

In the Cox proportional hazards models predicting time to all-cause death, low hemoglobin was associated with an increased risk of mortality, independent of CD4 cell count, WHO clinical stage of disease, age, pregnancy, treatment arm in the vitamin study, and body mass index (Table 2). The risk of death was more than doubled for patients with anemia (relative hazard [RH]: 2.06, 95% CI: 1.52 to 2.79 for moderate anemia and RH: 3.19, 95% CI: 2.23 to 4.56 for severe anemia), compared with those without anemia. The risk of AIDS-related mortality was also more than doubled for patients with anemia (RH: 2.21, 95% CI: 1.53 to 3.19 for moderate anemia and RH: 3.47, 95% CI: 2.25 to 5.33 for severe anemia), compared with patients without anemia. In a separate model in which hemoglobin was modeled as a continuous variable (not shown), each 1-g/dL decrease in hemoglobin was associated with a 25% increased risk of death due to any cause (RH: 1.25, 95% CI: 1.17 to 1.33) and a 28% increased risk of AIDS-related death (RH: 1.28, 95% CI: 1.19 to 1.38). In separate models in which erythrocyte morphology was substituted for hemoglobin level, the risks of all-cause and AIDS-related mortality were both significantly increased for patients with hypochromasia (with or without microcytosis), relative to patients without hypochromasia (Table 2).

Association of Anemia With Time to Death (n = 1074 women)

Anemia was also associated with time to immunologic disease progression (Table 3). After adjustment for baseline CD4 cell count, WHO clinical stage of disease, age, pregnancy, treatment arm in the vitamin study, and body mass index, anemia (RH: 1.79, 95% CI: 1.38 to 2.33 for moderate anemia and RH: 1.59, 95% CI: 1.02 to 2.49 for severe anemia, compared with normal) was associated with time to 50% decrease in CD4 count from baseline among the 865 women with CD4 count measured at enrollment and at least once during follow-up. In a separate model in which erythrocyte morphology was substituted for hemoglobin level, presence of hypochromasia (with or without microcytosis) was also associated with time to 50% decline in CD4 count.

Association of Anemia With Time to 50% Decline in CD4 Cell Count From Baseline (n = 865 women)

To compare the findings in this study with those of other published findings, we also ran Cox proportional hazard models for the mortality endpoints that considered the hemoglobin value and the other covariates as time-fixed covariates measured at the time of enrollment in the study. In these models, hemoglobin was not a significant predictor of time to death or AIDS-related death (data not shown).


In this study of HIV-positive women in Tanzania we found that anemia was associated with increased risk of death over the follow-up, independent of CD4 cell count, WHO clinical stage of disease, pregnancy, age, treatment arm in the vitamin study, and body mass index. These findings are similar to those reported from cohorts in the United States and Europe.2,3,5-10 In Haiti, Deschamps et al35 reported an association between anemia at the time of seroconversion and subsequently lowered survival. However, to our knowledge, our study is the first to document such an association among women in sub-Saharan Africa. The majority of studies of anemia and survival among persons with HIV infection have relied on a single measure of hemoglobin at baseline3,8-10 and among patients who had been diagnosed with AIDS at the start of the study.8-10 In contrast, our findings are based on survival models with time-varying measures of hemoglobin and they originated from a cohort of women who were at an early stage of disease when follow-up began. Mocroft et al2 conducted a similar analysis to the one presented here, using serial measures of hemoglobin, but among a predominantly male cohort from EuroSIDA in which the median CD4 cell count (201 cells/mL, IQR: 84 to 323) was lower than in our female cohort (405 cells/mL, IQR: 275 to 543). Nevertheless, in a model adjusted for CD4 cell count, the authors reported a similar effect size for the risk of death for each 1-g/dL decrease in hemoglobin (RH: 1.39) as we found in this analysis (RH: 1.25, results not shown).

Anemia was also associated with AIDS-related death, after similar adjustment. Compared with normal hemoglobin (≥11.0 g/dL), the risk of AIDS-related death was approximately 2-fold for patients with moderate anemia and 3-fold for those with severe anemia. For 10 of the 243 patients whose death was deemed AIDS related, anemia was reported to be the cause of death; censoring these 10 patients did not change the results. To our knowledge, no other studies have examined the association between anemia with AIDS-related death.

In adjusted models, anemia was also associated with time to immunologic disease progression, measured as a 50% decrease in CD4 cell counts from baseline. Among patients receiving zidovudine therapy, low baseline levels of hemoglobin or hematocrit were related to accelerated progression to AIDS,12,36 but to our knowledge, no published studies have examined the association between hemoglobin and decreases in CD4 cell counts.

At the baseline visit, hypochromasia and microcytosis were identified in the blood smears of 45% of the women, and 86% had mean corpuscular volume values ≤80 fL, suggesting that iron deficiency was an important contributing factor to the high rates of anemia we observed.20 These results are in accordance with other findings from sub-Saharan Africa.24,25,37,38

In this study, erythrocyte morphology characteristics indicative of iron deficiency were associated with all-cause and AIDS-related death as well as declines in CD4 cell count. To our knowledge, no other published studies have examined the association between iron deficiency and HIV disease progression, but poor nutrition has been linked with accelerated progression of HIV disease.39

We used hemoglobin as a measure of anemia and examined erythrocyte characteristics (hypochromasia and microcytosis) as markers of iron deficiency. However, it is important to note that both measures are limited. Hemoglobin is the most widely used screening measure of anemia as a proxy for iron deficiency, because it simple, quick, inexpensive, and a better marker than hematocrit.40-42 Because it captures the more advanced stage of iron deficiency, hemoglobin is a marker of nutritionally significant iron deficiency,43 but it may not detect early stages of iron deficiency.40-42 Erythrocyte characteristics are fairly specific indicators of iron deficiency, but they are not very sensitive and may be more likely to underestimate the occurrence of iron deficiency.24 Other methods to measure iron deficiency anemia include bone marrow biopsy, ferritin, erythrocyte protoporphyrin, transferrin saturation, and serum iron. Studies that incorporate these methods are needed to further elucidate the relationship between iron deficiency and HIV progression.

Women enrolled in this cohort were pregnant at the time of enrollment, so their hemoglobin levels at the baseline visit may have been lowered as a result of pregnancy. Seventy-eight percent of participants were at least 16 weeks pregnant at the time of enrollment, and the mean hemoglobin for these women was significantly lower (9.3 g/dL) than that for women who were <16 weeks pregnant (9.8 g/dL). Given that 32% of the women had at least 1 additional pregnancy during the follow-up period, pregnancy-related anemia might have been an important contributing factor in the high rates of anemia is this cohort.

Contrary to several other studies, we did not detect a significant association between the baseline hemoglobin value and time to death or disease progression. There are several possible explanations for this lack of association. One possibility is that only 1 measure taken during pregnancy may not be an accurate predictor of long-term risk of disease progression. The women in this study were clinically asymptomatic for HIV disease (98% with WHO stage 1 or 2 disease), with a median CD4 cell count of 405 cells/mL (IQR: 275 to 543) at the time of enrollment, and it is plausible that a single measure of hemoglobin is more strongly associated with disease progression in more advanced stages of disease than was present in our cohort.

Evidence that iron overload enhances progression of HIV disease44,45 has generated concern that iron supplementation may be harmful for HIV-infected individuals.46,47 However, these data arise from studies of patients in industrialized countries who have low risk of iron deficiency; furthermore, iron status was not positively related to markers of HIV disease severity in a cross-sectional study conducted among pregnant women in Malawi.48 Iron supplementation is still regarded as a safe and effective intervention in developing countries, though further studies are needed to clarify the effectiveness of such supplementation interventions. Screening for anemia is relatively inexpensive and hence it should be conducted more extensively, particularly among pregnant women and persons with HIV.

In developed countries, HIV-associated anemia is commonly treated with recombinant human erythropoietin, which has been associated with improved survival5 and quality of life.13,49 However, at its current price, erythropoietin is not an option for widespread use in resource-limited settings such as Tanzania.50 Transfusion for treatment of anemia is associated with an increased risk of death among patients with HIV5 and is discouraged in Tanzania except in cases of very severe anemia, due to the risk of transmission of other infections as well as other predominant HIV subtypes. Therefore, efforts to correct anemia among women with HIV in Tanzania, as in other resource-limited settings, should rely more on micronutrient supplementation programs, treatment of infections associated with anemia, such as malaria and hookworm, and improving access to diverse diets rich in iron, such as meat, fish, fowl, poultry, organs from cattle, legumes, and leafy green vegetables. Iron and folate supplementation is a standard part of antenatal care in Tanzania, but these findings suggest that this effort may not be sufficient and that steps to reduce anemia among women with HIV are warranted beyond the antenatal period as well.

At the time of this study, antiretrovirals were not available for treatment of HIV in Tanzania. Presently, the government of Tanzania is rolling out a national antiretroviral treatment program. Because treatment with highly active antiretroviral therapy (HAART) has been associated with reductions in anemia among patients in developed countries,7,51-55 monitoring of the antiretroviral program should include assessments of anemia to determine whether HAART has similar effects in Tanzania and to evaluate the impact of zidovudine-containing regimens, which are part of the national plan for HIV care and treatment and have been associated with development of anemia.1,20,56


The authors thank the women who participated in this study and the field teams who made the study possible.


1. Levine AM, Berhane K, Masri-Lavine L, et al. Prevalence and correlates of anemia in a large cohort of HIV-infected women: Women's Interagency HIV Study. J Acquir Immune Defic Syndr. 2001;26:28-35.
2. Mocroft A, Kirk O, Barton S, et al. Anaemia is an independent predictive marker for clinical prognosis in HIV-infected patients from across Europe. AIDS. 1999;13:943-950.
3. Sullivan P, Hanson D, Chu S, et al. Epidemiology of anemia in human immunodeficiency virus (HIV)-infected persons: results from the Multistate Adult and Adolescent Spectrum of HIV Disease Surveillance Project. Blood. 1998;91:301-308.
4. World Health Organization. The Prevalence of Anaemia in Women: A Tabulation of Available Information. Document WHO/MCH/MSM/92.2. Geneva: World Health Organization; 1992.
5. Moore RD, Keruly JC, Chaisson RE. Anemia and survival in HIV infection. J Acquir Immune Defic Syndr. 1998;19:29-33.
6. Moore RD. Human immunodeficiency virus infection, anemia, and survival. Clin Infect Dis. 1999;29:44-49.
7. Lundgren JD, Mocroft A. Anemia and survival in human immunodeficiency virus. Clin Infect Dis. 2003;37(Suppl 4):S297-S303.
8. Saah AJ, Hoover DR, He Y, et al. Factors influencing survival after AIDS: report from the Multicenter AIDS Cohort Study (MACS). J Acquir Immune Defic Syndr. 1994;7:287-295.
9. Creagh-Kirk T, Doi P, Andrews E, et al. Survival experience among patients with AIDS receiving zidovudine: follow-up of patients in a compassionate plea program. JAMA. 1988;260:3009-3015.
10. Swanson CE, Cooper DA. Factors influencing outcome of treatment with zidovudine of patients with AIDS in Australia. The Australian Zidovudine Study Group. AIDS. 1990;4:749-757.
11. Morfeldt-Manson L, Bottiger B, Nilsson B, et al. Clinical signs and laboratory markers in predicting progression to AIDS in HIV-1 infected patients. Scand J Infect Dis. 1991;23:443-449.
12. Moore RD, Creagh-Kirk T, Keruly J, et al. Long-term safety and efficacy of zidovudine in patients with advanced human immunodeficiency virus disease. Zidovudine Epidemiology Study Group. Arch Intern Med. 1991;151:981-986.
13. Revicki DA, Brown RE, Henry DH, et al. Recombinant human erythropoietin and health-related quality of life of AIDS patients with anemia. J Acquir Immune Defic Syndr. 1994;7:474-484.
14. Kreuzer KA, Rockstroh JK, Jelkmann W, et al. Inadequate erythropoietin response to anaemia in HIV patients: relationship to serum levels of tumour necrosis factor-alpha, interleukin-6 and their soluble receptors. Br J Haematol. 1997;96:235-239.
15. Kaslow RA, Phair JP, Friedman HB, et al. Infection with the human immunodeficiency virus: clinical manifestations and their relationship to immune deficiency. A report from the Multicenter AIDS Cohort Study. Ann Intern Med. 1987;107:474-480.
16. Mir N, Costello C, Luckit J, et al. HIV-disease and bone marrow changes: a study of 60 cases. Eur J Haematol. 1989;42:339-343.
17. Zon LI, Arkin C, Groopman JE. Haematologic manifestations of the human immune deficiency virus (HIV). Br J Haematol. 1987;66:251-256.
18. Semba RD, Gray GE. Pathogenesis of anemia during human immunodeficiency virus infection. J Investig Med. 2001;49:225-239.
19. Kreuzer K, Rockstroh J. Pathogenesis and pathophysiology of anemia in HIV infection. Ann Hematol. 1997;75:179-187.
20. Moyle G. Anaemia in persons with HIV infection: prognostic marker and contributor to morbidity. AIDS Rev. 2002;4:13-20.
21. Volberding PA, Baker KR, Levine AM. Human immunodeficiency virus hematology. Hematology (Am Soc Hematol Educ Program). 2003:294-313.
22. Means RT Jr, Krantz SB. Progress in understanding the pathogenesis of the anemia of chronic disease. Blood. 1992;80:1639-1647.
23. UNAIDS. 2004 Report on the Global AIDS Epidemic. Geneva: UNAIDS; 2004.
24. Massawe SN, Urassa EN, Mmari M, et al. The complexity of pregnancy anemia in Dar-es-Salaam. Gynecol Obstet Invest. 1999;47:76-82.
25. Hinderaker SG, Olsen BE, Lie RT, et al. Anemia in pregnancy in rural Tanzania: associations with micronutrients status and infections. Eur J Clin Nutr. 2002;56:192-199.
26. Semba RD, Kumwenda N, Hoover DR, et al. Assessment of iron status using plasma transferrin receptor in pregnant women with and without human immunodeficiency virus infection in Malawi. Eur J Clin Nutr. 2000;54:872-877.
27. van den Broek NR, White SA, Neilson JP. The relationship between asymptomatic human immunodeficiency virus infection and the prevalence and severity of anemia in pregnant Malawian women. Am J Trop Med Hyg. 1998;59:1004-1007.
28. Massawe S, Urassa E, Lindmark G, et al. Anaemia in pregnancy: a major health problem with implications for maternal health care. Afr J Health Sci. 1996;3:126-132.
29. Massawe SN, Urassa EN, Lindmark G, et al. Effectiveness of primary level antenatal care in decreasing anemia at term in Tanzania. Acta Obstet Gynecol Scand. 1999;78:573-579.
30. Antelman G, Msamanga GI, Spiegelman D, et al. Nutritional factors and infectious disease contribute to anemia among pregnant women with human immunodeficiency virus in Tanzania. J Nutr. 2000;130:1950-1957.
31. Meda N, Dao B, Ouangre A. HIV, maternal anemia and perinatal intervention using zidovudine. DITRAME Study Group (ANRS 049 Clinical Trial). Int J Gynaecol Obstet. 1998;61:65-66.
32. Ramon R, Sawadogo D, Koko FS, et al. Haematological characteristics and HIV status of pregnant women in Abidjan, Cote d'Ivoire, 1995-1996. Trans R Soc Trop Med Hyg. 1999;93:419-422.
33. Fawzi WW, Msamanga GI, Spiegelman D, et al. Rationale and design of the Tanzania Vitamin and HIV Infection Trial. Control Clin Trials. 1999;20:75-90.
34. Dacie J, Lewis S. Practical Haematology. 7th ed. Edinburgh: Churchill Livingstone; 1991.
35. Deschamps MM, Fitzgerald DW, Pape JW, et al. HIV infection in Haiti: natural history and disease progression. AIDS. 2000;14:2515-2521.
36. Graham N, Piantadosi S, Park L, et al. CD4+ lymphocyte response to zidovudine as a predictor of AIDS-free time and survival time. J Acquir Immune Defic Syndr. 1993;6:1258-1266.
37. van den Broek NR, Letsky EA, White SA, et al. Iron status in pregnant women: which measurements are valid? Br J Haematol. 1998;103:817-824.
38. Fleming AF. The aetiology of severe anaemia in pregnancy in Ndola, Zambia. Ann Trop Med Parasitol. 1989;83:37-49.
39. Tang AM, Graham NM, Saah AJ. Effects of micronutrient intake on survival in human immunodeficiency virus type 1 infection. Am J Epidemiol. 1996;143:1244-1256.
40. Binkin NJ, Yip R. When is anemia screening of value in detecting iron deficiency? In: Hercberg S, Galan P, Dupin H, eds. Recent Knowledge on Iron and Folate Deficiencies in the World. Paris: Colloq INSERM; 1990:197:137-46.
41. US Department of Health and Human Services. Screening for iron deficiency anemia including iron prophylaxis. In: Guide to Clinical Preventive Services. 2nd ed. Washington, DC: Office of Disease Prevention and Health Promotion; 1996.
42. Centers for Disease Control and Prevention. Recommendations to prevent and control iron deficiency in the United States. MMWR Morb Mortal Wkly Rep. 1998;47(RR-3):1-29.
43. Cook J. The nutritional assessment of iron status. Arch Latinoam Nutr. 1999;49:11S-14S.
44. Jacobus DP. Randomization to iron supplementation of patients with advanced human immunodeficiency virus disease: an inadvertent but controlled study with results important for patient care. J Infect Dis. 1996;173:1044-1045.
45. Salmon-Ceron D, Fontbonne A, Saba J, et al. Lower survival in AIDS patients receiving dapsone compared with aerosolized pentamidine for secondary prophylaxis of Pneumocystis carinii pneumonia. Study Group. J Infect Dis. 1995;172:656-664.
46. Boelaert JR, Weinberg GA, Weinberg ED. Altered iron metabolism in HIV infection: mechanisms, possible consequences, and proposals for management. Infect Agents Dis. 1996;5:36-46.
47. Savarino A, Pescarmona GP, Boelaert JR. Iron metabolism and HIV infection: reciprocal interactions with potentially harmful consequences? Cell Biochem Funct. 1999;17:279-287.
48. Semba RD, Taha TE, Kumwenda N, et al. Iron status and indicators of human immunodeficiency virus disease severity among pregnant women in Malawi. Clin Infect Dis. 2001;32:1496-1499.
49. Phair JP, Abels RI, McNeill MV, et al. Recombinant human erythropoietin treatment: investigational new drug protocol for the anemia of the acquired immunodeficiency syndrome: overall results. Arch Intern Med. 1993;153:2669-2675.
50. Bartlett J, Gallant J. 2003 Medical Management of HIV Infection. Baltimore: Johns Hopkins University, Division of Infectious Diseases and AIDS Service; 2003.
51. Moore R, Forney D. Anemia in HIV-infected patients receiving highly active antiretroviral therapy. J Acquir Immune Defic Syndr. 2002;29:54-57.
52. Semba RD, Shah N, Vlahov D. Improvement of anemia among HIV-infected injection drug users receiving highly active antiretroviral therapy. J Acquir Immune Defic Syndr. 2001;26:315-319.
53. Sarcletti M, Quirchmair G, Weiss G, et al. Increase of haemoglobin levels by anti-retroviral therapy is associated with a decrease in immune activation. Eur J Haematol. 2003;70:17-25.
54. Huang SS, Barbour JD, Deeks SG, et al. Reversal of human immunodeficiency virus type 1-associated hematosuppression by effective antiretroviral therapy. Clin Infect Dis. 2000;30:504-510.
55. Servais J, Nkoghe D, Schmit J, et al. HIV-associated hematologic disorders are correlated with plasma viral load and improve under highly active antiretroviral therapy. J Acquir Immune Defic Syndr. 2001;28:221-225.
56. Semba RD, Shah N, Klein RS, et al. Prevalence and cumulative incidence of and risk factors for anemia in a multicenter cohort study of human immunodeficiency virus-infected and -uninfected women. Clin Infect Dis. 2002;34:260-266.

epidemiology; mortality; HIV; anemia; Tanzania; CD4 lymphocyte count; female

© 2005 Lippincott Williams & Wilkins, Inc.