Body mass index (BMI) independently predicts mortality in studies of HIV infected patients initiating antiretroviral therapy (ART). We hypothesized that poorer nutritional status would be associated with smaller gains in CD4 count in Rwandan women initiating ART.
Methods and Findings
The Rwandan Women's Interassociation Study and Assessment, enrolled 710 ART-naïve HIV-positive and 226 HIV-negative women in 2005 with follow-up every 6 months. The outcome assessed in this study was change in CD4 count at 6, 12, and 24 months after ART initiation. Nutritional status measures taken prior to ART initiation were BMI; height adjusted fat free mass (FFMI); height adjusted fat mass (FMI), and sum of skinfold measurements. 475 women initiated ART. Mean (within 6 months) pre-ART CD4 count was 216 cells/µL. Prior to ART initiation, the mean (±SD) BMI was 21.6 (±3.78) kg/m2 (18.3% malnourished with BMI<18.5); and among women for whom the following were measured, mean FFMI was 17.10 (±1.76) kg/m2; FMI 4.7 (±3.5) kg/m2 and sum of skinfold measurements 4.9 (±2.7) cm. FFMI was significantly associated with a smaller change in CD4 count at 6 months in univariate analysis (−6.7 cells/uL per kg/m2, p = 0.03) only. In multivariate analysis after adjustment for covariates, no nutritional variable was associated with change in CD4 count at any follow up visit.
In this cohort of African women initiating ART, no measure of malnutrition prior to ART was consistently associated with change in CD4 count at 6, 12, and 24 months of follow up, suggesting that poorer pre-treatment nutritional status does not prevent an excellent response to ART.
Citation: Kiefer E, Hoover DR, Shi Q, Dusingize J-C, Cohen M, Mutimura E, et al. (2011) Association of Pre-Treatment Nutritional Status with Change in CD4 Count after Antiretroviral Therapy at 6, 12, and 24 Months in Rwandan Women. PLoS ONE 6(12): e29625. doi:10.1371/journal.pone.0029625
Editor: Landon Myer, University of Cape Town, South Africa
Received: June 30, 2011; Accepted: December 1, 2011; Published: December 28, 2011
Copyright: © 2011 Kiefer et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study was supported by supplements from the National Institute of Allergy and Infectious Diseases and the National Cancer Institute to the Bronx/Manhattan Women's Interagency HIV Study (WIHS) funded by the National Institute of Allergy and Infectious (UO1-AI-35004). This work was also supported in part by the AIDS International Training and Research Program (Fogarty International Center, NIH D43-TW001403) and the Center for AIDS Research at the Albert Einstein College of Medicine and Montefiore Medical Center funded by the National Institutes of Health (NIH AI-51519) and by the National Institute of Diabetes and Digestive and Kidney Disease (DK54615). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. No additional external funding was received for this study.
Competing interests: The authors have declared that no competing interests exist.
HIV-infection and poor nutritional status are prevalent in Africa. In Rwanda, malnutrition defined by the World Health Organization (WHO) definition of body mass index (BMI) < 18.5 kg/m2,  is substantially greater than in developed countries ; in our preliminary studies of Rwandan HIV-positive women in 2005 approximately 19% were malnourished .
Low body mass index (BMI) and serum albumin have been shown to independently predict mortality in HIV infected persons initiating ART in several United States and African studies , , , , , ; and also may be markers of advanced HIV disease , . In an Ethiopian study, weight loss in the first month after antiretroviral treatment (ART) initiation was associated with shorter survival . In the Tufts Nutrition for Healthy Living Cohort as little as 5% weight loss over 6 months or 3% from baseline, was associated with increased mortality in HIV infected persons, even after adjustment for ART use . Malnutrition can lead to increased susceptibility to infection through suppression of immune defense mechanisms, and HIV infection can cause malnourishment through increased concomitant opportunistic infections, malabsorption, and anorexia. However, dietary supplementation with food, which could potentially increase weight and improve nutritional status, has failed to consistently show reductions in HIV mortality in both high- and low-resource settings , . It is thus unclear whether nutritional measures such as BMI are simply markers of more progressive HIV disease and thus impending mortality, or are functioning independently as causal factors for illness or death; these questions have important programmatic and policy implications for food supplementation for HIV infected persons.
If CD4 cell response to ART were lower in malnourished compared to well-nourished individuals, this might imply a causal role of malnutrition in HIV mortality. However, few studies have investigated the association of malnutrition and CD4 response. Those that did used only BMI or body weight as a marker of malnutrition, which will not separately capture the different fat and non-fat (i.e. protein) aspects of malnutrition.. In the Singapore HIV Observational Cohort, there was no association of BMI measured 2 months before or after ART initiation with the magnitude of CD4 increase at 6 and 12 months after ART initiation, although lower BMI was significantly associated with reduced survival . Among HIV infected patients initiating ART in Cote d'Ivoire there was no difference in the proportion of people who gained 50 or more CD4 cells/uL 6 months after ART initiation among BMI categories >25, 18.5–25, and <18.5 kg/m2 .
BMI may provide an incomplete picture of pre-ART nutritional status in HIV infected individuals. BMI is an easily measured and well-validated indicator of malnutrition in populations without HIV, but it may not capture the changes in body composition that occur specifically in HIV infection because it does not distinguish fat from protein. Starvation and cachexia are two pathologically different forms of malnutrition that may occur in HIV-infected populations; whereas starvation involves losses in body fat, cachexia results in mainly a loss of lean body mass such as skeletal muscle . Further, in cachexia, changes in inflammation and the acute phase response leads to an elevation in resting energy expenditure and hypermetabolism, which in turn may affect response to ART . Therefore, height-normalized measures of fat-free mass and fat mass may provide important information about specific fat and protein malnutrition within these different compartments in persons with HIV infection . Skinfold measurements may also capture malnutrition by providing an estimate of subcutaneous fat. These measures have been validated as nutritional markers in non-African populations with HIV . In this cohort we previously showed correlation between these measures (r2 = .48), and several small studies have shown high correlation (r2 = .61 to .87) between fat mass and fat free mass between calculated and anthropometric measures of body fat, although among Africans of unknown HIV serostatus , , .
We therefore examined the association of nutritional status prior to ART initiation, using height-normalized measures of body mass, fat-free mass, fat mass, and skin fold thickness, with the change in CD4 count at an average of 6, 12, and 24 months after ART initiation in a cohort of Rwandan women with a high prevalence of malnutrition at study entry.
The objectives of this study were to examine the association of nutritional status prior to ART initiation, using height-normalized measures of body mass, fat-free mass, fat mass, and skin fold thickness, with the change in CD4 count at an average of 6, 12, and 24 months after ART initiation in a cohort of Rwandan women with a high prevalence of malnutrition at study entry. We hypothesized that poorer nutritional status would be associated with smaller gains in CD4 count in Rwandan women initiating ART.
The Rwanda Women's Interassociation Study and Assessment (RWISA) is an observational prospective cohort of 710 ART-naïve (at enrollment) HIV-infected and 226 HIV-uninfected Rwandan women enrolled in 2005 . Briefly, participants met the following inclusion criteria: age 25 years or older at study entry, willingness to give informed consent, presence in Rwanda during 1994. Participants were excluded if there was a prior history of receiving antiretroviral treatment other than single-dose nevirapine to prevent mother-to-child transmission of HIV. At study entry and at six-month interval visits, participants provided historical information including socio-demographics, medical history and symptoms, anthropometric measurements and blood specimens.
At each follow up visit, participants were asked whether they had initiated ART. Women reporting ART initiation provided written documentation of the exact date of ART initiation and medication regimen. From 2005 through 2007, following WHO and Rwandan guidelines, women in this cohort were eligible for ART if they had: WHO Stage IV disease, irrespective of CD4 cell count; WHO Stage III disease with CD4 <350 cells/µL; or CD4 <200/µL regardless of clinical stage. In 2008 Rwandan guidelines expanded the CD4 criterion to include ART initiation in all patients with CD4<350 cells/µl.
Included in this analysis are all HIV-positive RWISA participants who initiated ART after study entry, had a CD4 count obtained within one year prior to ART initiation (pre-ART CD4 count), nutritional measures at the pre-ART CD4 visit, and at least 6 months of follow-up with a CD4 measurement.
Body composition and anthropometric measurements
Height and weight were measured while the participant was wearing light clothing and no shoes. Body impedance analysis (BIA) was performed twice using a standard tetrapolar electrode placement on the hand and foot with resistance and reactance recorded. Skin-fold measurements were obtained by study nurses who were trained on techniques for standardized anthropometric measurement, using correct anatomic location of skin folds.
CD4 counts were determined with a FACS counter (Becton and Dickinson, Immunocytometry Systems, San Jose, CA, USA).
The main outcome of interest was change in CD4 count from the pre-ART to the 6, 12, and 24 month follow-up visits. Pre-ART values were defined as those measured at the study visit which fell between 1 day and 12 months prior to the exact date of ART initiation. If there were multiple visits in this time window, we chose the date closest to the exact date of ART initiation. The 6-, 12- and 24-month follow up visits were defined as the study visits which fell between 3 and 9 months, 9 and 15 months, and 21 and 27 months respectively after the exact date of ART initiation. Because some participants did not complete their follow-up at exact 6 month intervals, it was possible to have more than one visit in the follow up range, in which case the visit closest to the 6, 12, or 24 month follow up date was used.
The change in CD4 was calculated as the absolute CD4 count at the follow up visit minus the pre-ART CD4 count. Since the pre-ART CD4 count visit was on average 91 days (3 months) prior to the exact date of ART initiation, the change in CD4 at 6 months reflects on average a change over 9 months (with on average 6 of these months on ART). The change in CD4 at 12 months reflects on average, a change over 15 months (with on average 12 months on ART), and the change in CD4 at 24 months reflects on average a change over 27 months (with on average 24 months on ART).
Nutritional predictor variables
Four nutritional indicators were used, all from the pre-ART visit: body mass index (BMI); fat free mass index (FFMI); fat mass index (FMI); and the sum of skinfold measurement at the mid triceps, front thigh, and sub-scapular regions.
BMI, obtained from standing height and weight measurements, was calculated as weight divided by height-squared (kg/m2). We used WHO-established BMI cutoffs for nutritional status: malnourished (BMI < 18.5 kg/m2), normal (BMI 18.5–25 kg/m2), and overweight (BMI≥25 kg/m2) .
FFMI was obtained from BIA as described above. Resistance and reactance were entered into standard formulae to calculate fat free mass in kg; these formulae have been previously validated in a multi-ethnic HIV and non-HIV population in the United States, and have been used in several African-studies , , , , . For <5% of the participants the calculated FFM exceeded weight and was thus set to equal the weight. We then standardized the calculated fat free mass by dividing it by height in meters, squared, to obtain the FFMI (kg/m2). The FMI was calculated as weight in kg minus fat free mass in kg, and was standardized for height by dividing by height-squared (kg/m2).
Skinfolds measurements were taken twice by the same person at each site of mid triceps, front thigh, and sub-scapular region, and the average of the two measurements was used for analysis. If the two measurements differed by ≥2.0 mm, a third measure was taken and the closest two were averaged. The sum of the skinfolds measurement was the sum of the triceps, thigh and sub-scapular measurements in centimeters.
We included the following covariates obtained at the pre-ART visit because of their potential to confound the relationship between malnutrition and change in CD4 count: age, income, education, CD4 count, and self-reported prior occurrence of a Stage 4 WHO AIDS defining illness (ADI). Smoking was not included in the analysis as less than 3% of the women smoked.
Age was included as age per 5 years. Income was categorized as <10,000 Rwandan francs (FRW) per month (in 2005, this was equivalent to < $17), 10,000 to 35,000 FRW per month, and greater than 35,000 FRW per month. Education was categorized as none, some primary school, completed primary school, or some secondary school or higher. Pre-ART CD4 count was included in the analysis per 100 cells/µL increment. We determined the presence of Stage 4 WHO illness from participant self-report at all visits prior to and including the pre-ART study visit.
Each participant provided written informed consent after viewing a video demonstrating study procedures, and discussing the study with research personnel. Study protocols were approved by the Rwanda National Ethics Committee and the Institutional Review Board of Montefiore Medical Center.
The primary outcomes were changes in CD4 count (as a continuous variable), from the pre ART visit at 6, 12, and 24 months of follow up, as defined above. Univariate linear regression analysis was performed for the change in CD4 count at the 6-, 12-, and 24-months post ART. All nutritional measures were analyzed as continuous variables. Other covariates were analyzed as described above.
For the multivariate analysis, we created four multivariate linear regression models each containing one of the nutritional variables alone in the model (BMI, FFMI, FMI or skinfold sum) with the covariates listed above. This was performed 3 times using change from pre-ART visit to 6, 12, and 24 months post-ART as the outcome. We used backwards selection analysis with a p-value of p = 0.1 to stay. Variables that did not reach statistical significance were removed from the model until we obtained a final model. SAS software, version 9.1.3 (Cary, North Carolina) was used for the analysis.
Table 1 displays demographic and clinical characteristics of the 537 included women. The median preART CD4 count was 200 (interquartile range 142 to 279) cells/µL, The mean time from pre-ART measurement to the 6 month follow up visit was 244 days (8.0 months) with a median of 210 days (6.9 months); to the 12 month follow up visit, 425 days (14.0 months) with a median of 390 days (12.8 months); and to the 24 month follow up visit, 777 days (25.6 months) median of 771 days (25.3 months).
Prior to ART initiation, the mean (±Standard Deviation) BMI was 21.6 (±3.8) kg/m2, with 18.3% of the women classified as malnourished (BMI<18.5). The mean pre-ART FFMI and FMI were 17.1 (±1.8) kg/m2 and 4.7 (±3.5) kg/m2 respectively. The sum of these means (17.1 kg/m2 + 4.7 kg/m2) does not equal the mean BMI (21.6 kg/m2), as some people with BMI were missing BIA measures and thus did not have a calculated FMI and FFMI. The mean pre-ART sum of the three skinfolds measurements was 4.9 (±02.7) cm.
The mean changes in CD4 count from pre ART initiation at 6, 12, and 24 months post initiation were 71 (±107), 89 (±109) and 153 (±135) cells/µL, respectively. In univariate linear models (Table 2), the only significant association (at p < 0.05) between any nutritional measure and pre-post ART change in CD4 count occurred at 6 months of follow up. A higher FFMI was associated with a smaller increase in CD4 count at 6 months (−7 cells/µL per kg/m2, p = 0.03). For example, those with a FFMI of 17 kg/m2 gained on average 7 fewer CD4 cells/µL from the pre-ART visit to the 6 month post-ART visit compared to those with a FFMI of 16 kg/m2. BMI, FMI and skinfold measurements were not associated with change in CD4 count at 6 months. The changes in CD4 count at 12 and 24 months were not significantly associated with any of the pre-ART nutritional measurements in univariate analysis.
A higher pre-ART CD4 count was strongly and significantly associated in univariate analysis with smaller CD4 increases at all timepoints (p<0.0001 for all) perhaps in part reflecting regression to the mean . Of the demographic variables, at 6 months of follow up the only statistical association was with educational attainment: those having secondary school or higher compared to no education had a smaller increase in CD4 count (p = 0.001). At 12 and 24 months of follow up, an income of 10,000 to 35,000 FRW was associated with a smaller increase in CD4 count (p = 0.03 and p = 0.02, respectively).
In the four backwards selection models (BMI, FFMI, FMI, and sum of the skinfolds), no nutritional variable met criteria to stay in the adjusted models and thus none was independently associated with a change in CD4 cell count from pre-ART to 6, 12, or 24 months post-ART after adjustment for covariates (data not shown). In particular FFMI, which was associated with change in CD4 count at 6 months in univariate analysis, was no longer independently associated with this outcome in multivariate analysis. In all multivariate models preART CD4 remained independently inversely associated with change in CD4 at all time points (range from −13 and −36 cells/µl per 100 cells/µl increment; p = .01 to p<0.0001).
We did not find any association between BMI, FFMI, FMI or skinfolds measurements and the change in CD4 count from the pre-ART visit to 6, 12, or 24 months after ART initiation. In our study, we examined not only BMI, but a panel of anthropometric measures through the use of FFMI, FMI, and skinfolds to better capture pre-ART nutritional status in terms of loss of fat or protein within these body compartments. We found that none of these measures predicted change in CD4 count at any time point after initiation of ART, especially after adjustment for other known predictors of CD4 response such as pre-ART CD4 count, HIV viral load, or history of WHO Stage 4 AIDS defining illness. Thus, having a low BMI, low fat mass, low fat free (protein) mass or thin skinfolds did not preclude a robust response to ART as measured by the change in CD4 count. These multiple measures of either fat or protein loss, such as experienced in starvation or cachexia, were not associated with the ability to gain CD4 cells at several post-ART timepoints. In fact, the only statistically significant association we observed in univariate analysis (FFMI with CD4 change at 6 months) was in the opposite direction, suggesting a less robust response with higher FFMI. These anthropometric findings are similar to the findings from by Paton, Koethe and Toure, who examined only pre-treatment BMI and various responses to CD4 count and found no associations. To our knowledge, this is the first study to examine several anthropometric pre-ART markers of nutritional status and their associations with changes in CD4 count, particularly in African subjects.
It is possible that our measures may not have accurately captured nutritional status prior to treatment. However, our recent study of this cohort has shown that serum albumin levels, traditionally thought of as a marker of nutritional status, also did not correlate well with FMI, FFMI, BMI, or skinfolds . Serum albumin was lowest in the lowest CD4 count strata, and thus was more likely a marker of disease severity than of nutritional status.
Some limitations to this study should be noted. Although the exact dates of ART initiation were known, the time from the pre-ART nutritional measurements to follow up visit varied. However, our results show that the time from the pre-ART measurements to follow up visits was generally within 1–2 months of the goal follow up at 6, 12, and 24 months. Another limitation is that the resistance and reactance measurements, obtained from BIA measurements, were used to calculate fat mass and fat free mass in equations derived from a white and African-American population which may not completely reflect the relationships in Rwanda , although these equations have been used in several published studies in Africa , , , . We also note inter-operator variability when obtaining skinfold measurements, although we took care with multiple measurements for accuracy by our trained nurses. We were unable to define fully from participant self-report all WHO Stage 4 AIDs defining illnesses prior to initation of ART, as some of these illnesses were not specifically inquired about; thus we may not have accurately captured all confounding illness. Finally, caution must be made in extrapolating these data to populations with different levels of malnutrition.
In summary, we found that pre-ART nutritional status, as measured by BMI, FFMI, FMI and skinfolds did not predict the pre to post ART initiations changes in CD4 count at 6, 12, and 24 months after ART initiation in a cohort of HIV infected Rwandan women. Thus having poor values on these anthropometric measures may not preclude an excellent response to ART as measured by CD4 count. Further studies are warranted, using additional measures of nutritional status and with longer follow-up time, to assess further the impact of malnutrition on surrogate markers and clinical outcomes including mortality in Africans initiating ART.
Conceived and designed the experiments: EK DRH QS KA. Performed the experiments: EK DRH QS KA. Analyzed the data: EK DRH QS J-CD KA. Contributed reagents/materials/analysis tools: EK DRH QS J-CD MC EM KA. Wrote the paper: EK DRH QS J-CD MC EM KA.
- 1. WHO (1995) Physical Status: The Use and Interpretation of Anthropometry; Report of a WHO Expert Committee. Geneva: WHO.
- 2. (2009) The State of Food Insecurity in the World 2009. Rome: Food and Agriculture Organization of the United Nations.
- 3. Mutimura E, Anastos K, Zheng L, Cohen M, Binagwaho A, et al. (2010) Effect of HIV infection on body composition and fat distribution in Rwandan women. Journal of the International Association of Physicians in AIDS Care 9: 173–178.
- 4. Feldman JG, Burns DN, Gange SJ, Bacchetti P, Cohen M, et al. (2000) Serum albumin as a predictor of survival in HIV-infected women in the Women's Interagency HIV study. AIDS 14: 863–870.
- 5. Feldman JG, Gange SJ, Bacchetti P, Cohen M, Young M, et al. (2003) Serum albumin is a powerful predictor of survival among HIV-1-infected women. J Acquir Immune Defic Syndr 33: 66–73.
- 6. Jerene D, Endale A, Hailu Y, Lindtjorn B (2006) Predictors of early death in a cohort of Ethiopian patients treated with HAART. BMC Infect Dis 6: 136.
- 7. Johannessen A, Naman E, Ngowi BJ, Sandvik L, Matee MI, et al. (2008) Predictors of mortality in HIV-infected patients starting antiretroviral therapy in a rural hospital in Tanzania. BMC Infect Dis 8: 52.
- 8. Tang AM, Forrester J, Spiegelman D, Knox TA, Tchetgen E, et al. (2002) Weight loss and survival in HIV-positive patients in the era of highly active antiretroviral therapy. J Acquir Immune Defic Syndr 31: 230–236.
- 9. Zachariah R, Fitzgerald M, Massaquoi M, Pasulani O, Arnould L, et al. (2006) Risk factors for high early mortality in patients on antiretroviral treatment in a rural district of Malawi. AIDS 20: 2355–2360.
- 10. Graham SM, Baeten JM, Richardson BA, Wener MH, Lavreys L, et al. (2007) A decrease in albumin in early HIV type 1 infection predicts subsequent disease progression. AIDS Res Hum Retroviruses 23: 1197–1200.
- 11. Irlam JH, Visser ME, Rollins N, Siegfried N (2005) Micronutrient supplementation in children and adults with HIV infection. Cochrane Database Syst Rev CD003650.
- 12. Mahlungulu S, Grobler LA, Visser ME, Volmink J (2007) Nutritional interventions for reducing morbidity and mortality in people with HIV. Cochrane Database Syst Rev CD004536.
- 13. Paton NI, Sangeetha S, Earnest A, Bellamy R (2006) The impact of malnutrition on survival and the CD4 count response in HIV-infected patients starting antiretroviral therapy. HIV Med 7: 323–330.
- 14. Kotler DP (2000) Nutritional alterations associated with HIV infection. Journal of acquired immune deficiency syndromes 25: Suppl 1S81–87.
- 15. Kotler D, Heymsfield SB (1998) HIV infection: a model chronic illness for studying wasting diseases. The American journal of clinical nutrition 68: 519–520.
- 16. VanItallie TB, Yang MU, Heymsfield SB, Funk RC, Boileau RA (1990) Height-normalized indices of the body's fat-free mass and fat mass: potentially useful indicators of nutritional status. The American journal of clinical nutrition 52: 953–959.
- 17. Kotler DP, Burastero S, Wang J, Pierson RN Jr (1996) Prediction of body cell mass, fat-free mass, and total body water with bioelectrical impedance analysis: effects of race, sex, and disease. Am J Clin Nutr 64: 489S–497S.
- 18. Luke A, Durazo-Arvizu R, Rotimi C, Prewitt TE, Forrester T, et al. (1997) Relation between body mass index and body fat in black population samples from Nigeria, Jamaica, and the United States. Am J Epidemiol 145: 620–628.
- 19. Dioum A, Gartner A, Cisse AS, Delpeuch F, Maire B, et al. (2005) Validity of impedance-based equations for the prediction of total body water as measured by deuterium dilution in African women. Am J Clin Nutr 81: 597–604.
- 20. Mutimura E, Anastos K, Zheng L, Cohen M, Binagwaho A, et al. (2010) Effect of HIV infection on body composition and fat distribution in Rwandan women. J Int Assoc Physicians AIDS Care (Chic) 9: 173–178.
- 21. Cohen MH, Fabri M, Cai X, Shi Q, Hoover DR, et al. (2009) Prevalence and predictors of posttraumatic stress disorder and depression in HIV-infected and at-risk Rwandan women. Journal of women's health 18: 1783–1791.
- 22. Mupere E, Zalwango S, Chiunda A, Okwera A, Mugerwa R, et al. (2010) Body composition among HIV-seropositive and HIV-seronegative adult patients with pulmonary tuberculosis in Uganda. Ann Epidemiol 20: 210–216.
- 23. Shah S, Whalen C, Kotler DP, Mayanja H, Namale A, et al. (2001) Severity of human immunodeficiency virus infection is associated with decreased phase angle, fat mass and body cell mass in adults with pulmonary tuberculosis infection in Uganda. J Nutr 131: 2843–2847.
- 24. Van Lettow M, Kumwenda JJ, Harries AD, Whalen CC, Taha TE, et al. (2004) Malnutrition and the severity of lung disease in adults with pulmonary tuberculosis in Malawi. Int J Tuberc Lung Dis 8: 211–217.
- 25. Villamor E, Saathoff E, Mugusi F, Bosch RJ, Urassa W, et al. (2006) Wasting and body composition of adults with pulmonary tuberculosis in relation to HIV-1 coinfection, socioeconomic status, and severity of tuberculosis. Eur J Clin Nutr 60: 163–171.
- 26. Bland JM, Altman DG (1994) Regression towards the mean. BMJ 308: 1499.
- 27. Dusingize JC, Hoover D, Shi Q, Kiefer E, Anastos K (2011) Association of serum albumin with markers of nutritional status among HIV-Infected and uninfected Rwandan women. 2011 Clinical and Translational Research and Education Meeting. Washington, DC.