Vitamin A deficiency and other factors associated with severe - Nature

European Journal of Clinical Nutrition (2009) 63, 1130–1135

& 2009 Macmillan Publishers Limited All rights reserved 0954-3007/09 $32.00 www.nature.com/ejcn

ORIGINAL ARTICLE

Vitamin A deficiency and other factors associated with severe tuberculosis in Timor and Rote Islands, East Nusa Tenggara Province, Indonesia TA Pakasi1,2, E Karyadi1,7, Y Wibowo1, Y Simanjuntak1, NMD Suratih3, M Salean4, N Darmawidjaja5, JWM van der Meer6, K van der Velden7 and WMV Dolmans6 1

SEAMEO TROPMED RCCN and formerly joint research with, University of Indonesia, Jakarta, Indonesia; 2Department of Community Medicine, Faculty of Medicine, University of Indonesia, Jakarta, Indonesia; 3Department of General Medicine, WZ Johannes General Hospital, Kupang, Indonesia; 4Department of Radiology, WZ Johannes General Hospital, Kupang, Indonesia; 5Kefamenanu General Hospital, North Middle Timor District (TTU), Indonesia; 6Department of Internal Medicine, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands and 7Department of Public Health, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands

Background: Plasma zinc and vitamin A concentrations have been reported to be low in tuberculosis (TB) patients in some studies, although it is not clear whether this constitutes a risk for a more severe clinical presentation among TB patients. The acute phase reaction may also deplete zinc and vitamin A in the plasma. Therefore, we further studied these associations. Methods: We carried out a cross-sectional study among newly diagnosed sputum smear-positive TB patients in East Nusa Tenggara. The patients were categorized as either mild TB when Karnofsky Score (KS) X80 or severe TB (KS o80). Body mass index (BMI), mid upper arm circumference (MUAC), chest radiograph, and the results of hemoglobin, erythrocyte sedimentation rate, albumin, C-reactive protein (CRP), zinc and vitamin A in plasma were correlated with TB category. Results: A total of 300 TB patients participated in the study (63% male and 37% female), and were categorized as mild TB (53%) or severe TB (47%). Vitamin A, hemoglobin and plasma albumin were significantly lower, and CRP was significantly higher, in severe TB than in mild TB, and the active lesion area on the chest radiograph was greater among severe TB patients. In a multiple regression analysis, after adjustment for CRP, low vitamin A (b ¼ 3.2, 95%CI (confidence interval) 1.6–4.9, P ¼ 0.000) but not zinc, correlated with the severity of TB. MUAC was better than BMI as a predictor of TB severity (b ¼ 1.3, 95%CI 0.6–6.2, P ¼ 0.000). Conclusions: Severe TB was associated with vitamin A deficiency. MUAC can be applied as a measure of TB severity.

European Journal of Clinical Nutrition (2009) 63, 1130–1135; doi:10.1038/ejcn.2009.25; published online 27 May 2009 Keywords: vitamin A; zinc; CRP; MUAC; severity; tuberculosis

Introduction Malnutrition is a major public health problem in all developing countries including Indonesia. It increases susceptibility to and severity of an infection (Muller and Correspondence: Dr TA Pakasi, Department of Community Medicine, Faculty of Medicine, School of Medicine, University of Indonesia, JI. Pegangsaan Timur no. 16, Jakarta Pusast, Jakarta 10430, Indonesia. E-mail: [email protected] 7 Current address: Micronutrient Initiative, Komp. Bappenas A-1 Jl. Siaga raya Pejaten Barat, Jakarta 12510, Indonesia. Received 4 June 2008; revised 13 January 2009; accepted 2 March 2009; published online 27 May 2009

Krawinkel, 2005) and tuberculosis (TB) is a prominent example (Cegielski and McMurray, 2004). Malnutrition as reflected by a low body mass index (BMI) is common among TB patients (Metcalfe, 2005) and it increases mortality from TB (Zachariah et al., 2002). Thus, malnutrition seems to be an important determinant of clinical outcome of TB. In addition to BMI, other nutritional indicators, such as mid upper arm circumference (MUAC) and percentage of fat, are also associated with malnutrition. Apart from macronutrient deficiency, deficiency of micronutrients has been shown to be an important determinant of TB outcome. Two important micronutrients that have been studied recently in Jakarta are zinc and vitamin A. In this

Vitamin A deficiency and risk for severe tuberculosis TA Pakasi et al

1131 case–control study, it was found that zinc and vitamin A deficiency were more prevalent among sputum smearpositive (SS þ ) TB patients than in matched healthy controls, thus contributing to the susceptibility for TB (Karyadi et al., 2000). This conclusion is supported by in vitro and animal studies that showed the role of vitamin A (Villamor and Fawzi, 2005) and zinc in cell-mediated immunity (Prasad, 1998), the important host defence mechanism against TB. Although micronutrients such as zinc and vitamin A may affect the immune status, the mechanism by which these micronutrients influence the outcome of TB is still unclear. The effects of zinc and vitamin A on the immune response lead to the question whether deficiency of these micronutrients poses a risk for severe TB. However, it is not easy to prove this notion, as the acute phase response in inflammation causes depletion of plasma zinc (Walravens, 1979) and vitamin A in the form of plasma retinol (Paracha et al., 2000). In this study, we investigated whether severity of TB was associated with low plasma zinc and retinol, adjusted for C-reactive protein (CRP) level and its clinical determinants. The results may justify interventions with zinc and vitamin A on top of TB treatment.

Participants and methods Study design, location and participants We carried out a cross-sectional study in four districts, covering 21 primary health centers (Puskesmas) and three district hospitals in Timor and Rote Islands, East Nusa Tenggara (NTT) Province, Indonesia. Participants were newly diagnosed SS þ pulmonary TB outpatients, who presented at the provincial hospitals or primary health centers. The inclusion criteria were new SS þ pulmonary TB, age 15–55 years and not pregnant or lactating. Exclusion criteria were diabetes mellitus, heart disease, and accident or trauma or a surgical procedure within 1 month before inclusion into the study. HIV testing was not carried out. All eligible patients were informed about the study and signed an informed consent form. Ethical clearance was provided by the Board of Research Ethics from the Faculty of Medicine, University of Indonesia, Jakarta.

Data collection Personal interview and clinical examination. Interviews using structured questionnaires were used to collect the sociodemographic background and data on medical history. Subsequently, the patients were thoroughly examined by medical doctors trained for the research purpose at the district hospitals (WZ Johannes, Kefamenanu and Ba’a). On the basis of this information the Karnofsky score (KS) was calculated, ranging from 0 (dead) to 100 (normal). Diagnosis of TB was on the basis of three times sputum smear examinations (SS þ ) in accordance with the National TB

Program guidelines from the Ministry of Health of the Republic of Indonesia. Chest X-ray examination. Chest radiographs were made of all patients after the diagnosis of TB and were evaluated by judging the visible active lesion area in both the lungs. In patients with cavities, the total area of exposed cavity wall was calculated from the radius of visible cavities (pr2) as described earlier (Karyadi et al., 2000). Anthropometric measurements. Body weight was assessed using an electronic platform model weighing scale (770 alpha; SECA, Hamburg, Germany) to the nearest 0.1 kg. Height was recorded to the nearest 0.1 cm using a microtoise, BMI was calculated using these data (kg/m2). Malnutrition was defined according to WHO (World Health Organization) criteria (available at: http://www.who.int/bmi/index.jsp): severe malnutrition (BMI o16 kg/m2), moderate malnutrition (BMI 16–16.9 kg/m2), mild malnutrition (BMI 17–17.49 kg/m2), normal (BMI 18.5–24.9 kg/m2) and overweight/obesity (BMIX25 kg/m2). MUAC was measured with a flexible in-elastic plastic measurement tape. Skin fold thickness was assessed at four sites: biceps, triceps, subscapular and supra-iliac regions, using a slim guide skin fold callipers (Creative health product, Plymouth, Michigan, 48170), recorded to the nearest 0.2 mm. Each site was measured twice on the left side of the body. The percentage of total body fat was on the basis of anthropometric data, and calculated using the Durnin and Womersley equations (Durnin and Wormesley, 1974). Blood collection and analyses. Fasting blood samples were collected between 0800 hours and 1000 hours in the district hospital. Approximately, 5 ml of the whole blood was taken by venipuncture into 5 ml vacutainers (Greiner Bio-One, Germany) containing EDTA. Hemoglobin, erythrocyte sedimentation rate and serum albumin were determined at the district hospital laboratory on the same day. Plasma was separated after centrifugation of blood at 750 g for 10 min at room temperature, and then stored at 20 1C until analysis of CRP, retinol and zinc. Specimens were transported at a temperature of 160 1C using dry ice, in cryoboxes to avoid direct sunlight. The cut-off points for normal hemoglobin were 412 and 413 g/100ml for women and men, respectively. Erythrocyte sedimentation rate was assessed using Westengren technique with a normal value o20 mm/ h. Serum albumin was measured using a spectrophotometer (Microlab 300, Merck, Germany) with a normal range 35–52 g/l. CRP was measured using ELISA (Enzyme-Linked Immuno-Sorbent Assay) method with a normal value o10 mg/l. Plasma retinol concentration was measured by HPLC (High Performance Liquid Chromatography), values o0.70 mmol/l were regarded as indicating vitamin A deficiency. Plasma zinc was analyzed using Atomic Absorption Spectrometry with normal values p10.7 mmol/l (Dijkhuizen et al., 2004). European Journal of Clinical Nutrition


1132 Food intake questionnaire. Food intake was assessed on the basis of 24-h recall and semi-quantitative Food Frequency Questionnaire to estimate the intake of energy, protein, fat, carbohydrate, vitamin A, zinc and iron. The 24-h recall was conducted on the weekdays. Each 24-h recall was conducted using a standardized four-stage protocol (Gibson, 1993). Part of the training session consisted of determining the differences between the amount estimated by each trainee and the actual weight of the food. An acceptable training level was considered when the average difference between the trainee’s estimate and the actual food weight was p5 g. We used Nutrisurvey 2005 r SEAMEO TROPMED RCCN, University of Indonesia, to convert the weight of the food into total calories, grams of fat, protein and carbohydrate, and zinc and vitamin A intake. Laboratory definitions. Vitamin A deficiency was defined as plasma retinol concentration below 0.70 mmol/l (De Pee and Dary, 2002). Zinc deficiency was defined as plasma Zn concentration 10.7 mmol/l or below (Karyadi et al., 2000).

Statistic analysis One sample Kolmogorov–Smirnov test was applied to examine normality of the data. For not-normally distributed variables, we examined differences of medians between groups using Mann–Whitney U-test for significance, and independent student-t test to test differences between means for normally distributed variables (Petrie and Sabin, 2000) We measured predictors using stepwise linear regression analysis (Hosmer and Lemeshow, 2000).

Results Patients were recruited between January 2004 and June 2005. Of the 497 patients initially enrolled, 348 fulfilled the entry criteria. Reasons for exclusion of the 149 were age over 55 (83) or below 15 years (5), relapse (23), lactating mothers (11), extra-pulmonary TB (9), co-morbidity with diabetes mellitus (5), hepatitis (5) and other causes (8). A further 48 patients did not participate because of: lost to follow-up (11), living in very remote area and out of the coverage area of the primary health centers (4) and refusing to participate (33). Therefore, data are presented of 300 patients. On the basis of KS, 140 (47%) out of these were classified as severe TB and 160 (53%) as mild TB. Most of the patients had chronic cough (98.3%) and weight loss (90.7%); other frequent symptoms were anorexia (65.3%), shortness of breath (45.3%) and hemoptysis (67%). Table 1 shows the clinical, laboratory and radiographic characteristics of the patients. Younger patients were found significantly more often within the mild TB group. No difference was found for gender. It is interesting that people indigenous to Timor and Rote had more severe TB than the other ethnic groups. There was a tendency for higher degree European Journal of Clinical Nutrition

of acid-fast bacilli positivity of the sputum smears in patients with severe TB (P ¼ 0.056). Most patients showed low BMI, the mean BMI was 16 Kg/m2, sub-categorized as severe malnutrition (43.7%), moderate malnutrition (16.3%), mild malnutrition (23%) and normal BMI (17%). Median MUAC in the 300 patients was 21.2 cm (IQR (interquartile range) 19.8–23.3 cm). Mild TB patients had significantly higher mean BMIs and MUACs than those with severe TB (P ¼ 0.000). The area with active lesions on chest radiographs was significantly greater in severe TB (P ¼ 0.000), but no significant differences were found for the presence of cavities and the cavity areas. Among the patients with a normal CRP, zinc deficiency occurred in 30.2% and vitamin A deficiency in 19%. When comparing mild and severe TB, there was no significant difference with regard to the occurrence of zinc deficiency, but vitamin A deficiency was significantly more frequent among patients with severe TB (P ¼ 0.000). We found significant correlations between BMI, MUAC, KS and plasma retinol, regardless of whether CRP was elevated or not. Spearman correlation coefficient between BMI and retinol was 0.3 (P ¼ 0.000), and between BMI and KS was 0.5 (P ¼ 0.000). Similarly, the correlation coefficient between MUAC and retinol was 0.3, and between MUAC and KS was 0.5. In stepwise linear regression analysis, age, ethnicity, MUAC, CRP and plasma retinol were significant predictors of severity of TB (Table 2). BMI, gender, presence of cavities and zinc status were not the significant predictors. The 24-h food recall of all patients showed lower intake than the recommended daily allowance for Indonesia (Table 3). The energy intake of this study group reached roughly one-third of the recommended amount, and the protein intake reached only approximately half of the recommendation. Regarding zinc and vitamin A intake, our study population consumed approximately half of the recommended daily allowances. Significant differences in energy intake (P ¼ 0.02), protein intake (P ¼ 0.007) and zinc intake (P ¼ 0.048) were found between mild TB and severe TB (Table 3).

Discussion In this study we showed that plasma concentrations of retinol, but not zinc, were associated with the severity of TB, on the basis of KS, after adjustment for CRP concentration in plasma. Similar findings were found for age, ethnicity and MUAC. Our findings are in line with the study in Malawi that found vitamin A deficiency, and not zinc deficiency, to be associated with wasting among TB patients, regardless of the HIV serostatus (van Lettow et al., 2004). In that study, vitamin A deficiency might have been influenced by the inflammatory process, but inflammation markers such as CRP were not measured. Thus, vitamin A deficiency might have been overestimated. The inflammatory process downregulates protein production by the liver, such as albumin, retinol-binding protein and other proteins for transporting


1133 Table 1 Characteristics and clinical presentation of the tuberculosis patients Severe TB (KS p80) n ¼ 160

All TB patients n ¼ 300

Variables Age (years)a

28 (22–39)

31.5 (24–41.8)

Mild TB (KS 480) n ¼ 140

P-value

26 (22–38.8)

0.004**

Gender (%) Male Female

189 (63) 111 (37)

89 (63.6) 51 (36.4)

99 (61.9) 61 (38.1)

0.811

Ethnic group (%) Indigenous Timor/Rote Other

225 (75) 75 (25)

120 (85.7) 20 (14.3)

105 (65.6) 55 (34.4)

AFB positivity grade (%) Grade 1 Grade 2 Grade 3

89 (29.7) 91 (30.3) 120 (40)

38 (27.1) 36 (25.7) 66 (47.1)

51 (31.9) 55 (34.4) 54 (33.8)

Body mass index (Kg/m2) b Severe malnutrition, n (%) Moderate malnutrition, n (%) Mild malnutrition, n (%) Normal, n (%)

16±2.3 131 (43.7) 49 (16.3) 69 (23) 51 (17)

15.5±1.8 93 (66.4) 17 (12.1) 20 (14.3) 10 (7.1)

17.4±2.2 38 (23.8) 32 (20.0) 49.9 (30.6) 41 (25.6)

0.000***

MUAC (cm)a Percentage body fat (%)a

21.2 (19.8–23.3) 10.2 (7.3–14.5)

20 (18.1–22) 9.5 (6.7–13.1)

22.7 (21–24.7) 10.9 (7.7–15.5)

0.000** 0.001**

120.5 (60.8–197) 85(38.1) 10.7 (4.9–18.5) 11.3±2.5 68(42.3–98) 3.9±0.8 27.7 (13.4–42.4) 11.9±2.6 107 (35.7) 0.7 (0.5–1.1) 149 (49.7)

174 (100–239.4) 55 (39.3) 12.6 (7.1–24.5) 10.8±25 65(40–68) 3.8±0.8 34.5 (19.3–46.8) 11.9±2.9 49 (35) 0.56 (0.39–0.88) 90 (64.3)

92.3 (44.4–158.6) 59 (36.9) 9.6 (4.4–15.9) 11.8±2.5 70 (45–110) 4.1±0.8 24.3 (8.5–38.1) 11.8±2.2 58 (36.3) 0.82 (0.58–1.22) 59 (36.9)

0.000** 0.718 0.374 0.001*** 0.193 0.002*** 0.000** 0.876 0.904 0.000** 0.000*

0.000*

0.056

Chest radiographic changes Active lesion area (cm)a Cavity (%) Cavity area (cm)a Hemoglobin (g/100 ml)b ESR (mm/h)a Plasma albumin (g/100 ml)b CRP (mg/l)a Zinc in plasma (mmol/l)b Zinc deficiency, n (%) Retinol in plasma (mmol/l)a Vitamin A deficiency, n (%)

Abbreviations: AFB, acid-fast bacilli; CRP, C-reactive protein; ESR, erythrocyte sedimentation rate; IQR, interquartile range; KS, Karnofsky Score; MUAC, mid upper arm circumference; TB, tuberculosis. Differences were tested for significance using w2-test for proportions*, Mann–Whitney U-test** for medians, and t-test*** for means. a Data were expressed as median (IQR) b Data were expressed as mean±s.d.

Table 2 Stepwise linear regression coefficients for Karnofsky score Variable Age Non-indigenousa BMI MUAC CRP Retinol

b

95% CI

P-value

0.2 2.8 0.033 1.3 0.06 3.2

0.3–0.1 0.6–5.1 0.9–0.8 0.6–2.0 0.09–0.03 1.6–4.9

0.000 0.015 0.941 0.000 0.005 0.000

Abbreviations: BMI, body mass index; CI, confidence interval; CRP, C-reactive protein; MUAC, mid upper arm circumference. The bold characters indicates significant P-values. a Compared with indigenous Timorese/Rotenese patients.

minerals, thus causing depletion of such vitamins and minerals. Whether vitamin A deficiency can be diagnosed in the presence of inflammation has been the subject of

debate for a long time. Thurnham et al., (2003) suggested to use inflammation markers, that is, CRP and a1-acid glycoprotein, to classify patients as vitamin A-deficient. However, categorizing patients on the basis of CRP concentrations does not completely solve the problem because an existing vitamin A deficiency may become worse because of the inflammation. CRP correlates negatively with plasma retinol. In addition, it makes a difference whether one deals with overt vitamin A deficiency (such as night blindness) or with subclinical vitamin A deficiency. A study among children with night blindness found that a classification on the basis of CRP did not give the expected prevalence of vitamin A deficiency, but instead led to sampling bias (Maqsood et al., 2004). The participants in this study, like those in the study by Thurnham et al. had no symptoms of vitamin A deficiency, therefore we only dealt with subclinical vitamin A deficiency. Therefore, we measured the European Journal of Clinical Nutrition


1134 Table 3 Average daily food intake of TB patients compared with the Indonesian recommended daily allowancesa Variables

Energy (kcal)b Protein (g)b Vitamin A (mg) Zinc (mg)b

All TB patients

922.4 26.4 282 3.3

(682.8–1121.0) (17–40.2) (105–576.5) (2.2–5.2)

Mild TB

972 29.2 299.7 3.5

(722.5–1245.4) (20.1–40.7) (117.6–556) (2.4–5.5)

Severe TB

850.6 23.1 258 3.1

(624.3–1186) (14.5–37.2) (62.2–598.6) (2.1–4.9)

P-value

0.022 0.007 0.463 0.048

Recommended daily allowance Men

Women

2550 60 600 12.1

1900 50 500 9.3

Abbreviation: TB, tuberculosis. Data were expressed as medians; between brackets P25 and P75 are given. a Recommended daily allowance for men and women aged 19–29 years was used according to the mean age of the patients in the study (28 years). b Significant differences were found using Mann–Whitney U-test.

association of low plasma retinol with TB severity using stepwise linear regression analysis after adjustment for CRP. Active lesions on the chest X-ray in severe TB were more pronounced than in mild TB. Extensive active lesions are associated with a higher degree of inflammation in the lung (as reflected by higher CRP), which in turn may reduce appetite and results in decreased food intake. Both groups showed low intake of energy, protein, zinc and vitamin A compared with the standard recommended daily allowance for Indonesia (Table 3). Thus, the low retinol concentrations in this study may have been a direct consequence of poor food intake or, indirectly, a consequence of the chronic inflammation. Zinc concentrations were not different between severe and mild TB. Both groups showed mean concentrations above the cut-off point for zinc deficiency. This finding differs from that in other studies, reporting that plasma zinc concentrations declined during infections as a result of the acute phase response (Brown, 1998). Another explanation would be that plasma zinc may be transiently low, particularly during the acute phase of the inflammatory process, and that the homeostatic mechanisms restore the zinc concentration in plasma, as the plasma zinc represents only 0.2% of total body zinc (Brown, 1998). We found that BMI, percentage of the body fat and MUAC were significantly lower among severe TB patients, although in linear regression analysis only MUAC was a predictor of severity of TB. MUAC and BMI are commonly used as indicators of malnutrition. However, MUAC reflects a more severe condition, as it also reflects muscle wasting. For screening in children, MUAC is an important indicator predicting mortality (Briend et al., 1987), but it is not commonly used for monitoring and follow-up of patients (Alam et al., 1989). There is an indirect correlation between MUAC and vitamin A deficiency. MUAC merely reflects long-term energy and protein deficiency and this may be also associated with vitamin A deficiency. More importantly, however, vitamin A needs fat for its absorption from the gastro-intestinal tract, and retinol-binding protein for transportation. Thus, low MUAC in TB patients may be used as a European Journal of Clinical Nutrition

better predictor of poor nutrition, and might indicate vitamin A deficiency; it is more reliable than BMI. In general, our patients presented with advanced TB, as most of them reported chronic cough, hemoptysis, anorexia and loss of weight. Delayed presentation and treatment likely caused the patients to lose appetite and become severely malnourished (indicated by low MUAC). Low intake of vitamin A and the disease itself caused low plasma retinol concentrations, and decrease of immune response. This in turn resulted in more severe TB. To summarize, our findings show that the clinical severity of TB is associated with low plasma retinol concentrations, but not with changes in plasma zinc. Low retinol may be the consequence of both inflammation and low intake. MUAC is better than BMI as a predictor for severity of TB. These findings may contribute to the justification for nutritional and particularly vitamin A supplementation among TB patients.

Acknowledgements This study was carried out with financial support from the Canadian International Development Agency (CIDA) through World Vision International, Indonesia: Food Integrated to Hinder Tuberculosis Project (the WVI—FIGHT Project) in East Nusa Tenggara. We would like to thank the provincial and district health offices in NTT, as well as the Puskesmas and the three district hospitals for collaboration in this study. We send our gratitude to our research team in Kupang, administrative and laboratory staffs in SEAMEO TROPMED RCCN, University of Indonesia, Jakarta. In this study, TAP carried out the field work as the coordinator, data management and analysis, and drafted the paper with the help of two master students, YW and YS. NMDS, MS and ND carried out the data collection and analysis of the participants. EK designed the study, participated in the coordinating and supervision of the data collection, and helped drafting the paper. JvdM, WD and KvdV helped in the interpretation of the data, and the drafting of the paper. All authors have read and approved the final paper.


1135

References Alam N, Wojtyniak B, Rahaman MM (1989). Anthropometric indicators and risk of death. Am J Clin Nutr 49, 884–888. Briend A, Wojtyniak B, Rowland MGM (1987). Arm circumference and other factors in children at high risk of death in rural Bangladeshi. The Lancet 1987, 725–728. Brown KH (1998). Effect of infections on plasma zinc concentration and implications for zinc status assessment in low-income countries. Am J Clin Nutr 68, 425S–429S. Cegielski JP, McMurray DN (2004). The relationship between malnutrition and tuberculosis: evidence from studies in humans and experimental animals. Int J Tuberc Lung Dis 8, 286–298. De Pee S, Dary O (2002). Biochemical indicators of vitamin A deficiency: serum retinol and serum retinol binding protein. J Nutr 132, 2895S–2901S. Dijkhuizen MA, Wieringa FG, West CE, Muhilal (2004). Zinc plus (-carotene supplementation of pregnant women is superior to (-carotene supplementation alone in improving vitamin A status in both mothers and infants. Am J Clin Nutr 80, 1299–1307. Durnin JW, Wormesley J (1974). Body fat assessment from total body density and its estimation from skinfold thickness: Measurement on 481 men and women aged from 16 to 72 year. Br J Nutr 32, 77–97. Gibson RS (1993). Nutritional Assessment. Oxford University Press: New York, NY. Hosmer DW, Lemeshow S (2000). Applied Logistic Regression, 2nd edn. John Wiley&Sons Inc: New York. Karyadi E, Schultink W, Nelwan RHH, Gross R, Amin Z, Dolmans WMV et al. (2000). Poor micronutrient status of active pulmonary tuberculosis patients in Indonesia. J of Nutr 130, 2953–2958.

Maqsood M, Dancheck B, Gamble MV, Palafox NA, Ricks MO, Briand K et al. (2004). Vitamin A deficiency and inflammatory markers among pre-school children in the Republic of the Marshall Islands. Nutr J 3, 1–6. Metcalfe N (2005). A study of tuberculosis, malnutrition and gender in Sri Lanka. Trans R Soc Trop Med Hyg 99, 115–119. Muller O, Krawinkel M (2005). Malnutrition and health in development countries. CMAJ 173, 279–286. Paracha PI, Jamil A, Northrop-Clewes CA, Thurnham DI (2000). Interpretation of vitamin A status in apparently healthy Pakistani children by using markers of subclinical infection. Am J Clin Nutr 72, 1164–1169. Petrie A, Sabin C (2000). Basic technique for analyzing data. In: Petrie A, Sbin C (eds). Medical statistics At A Glance, 2nd edn. Blackwell Publiching: Massachusetts. Prasad AS (1998). Zinc and Immunity. Mol Cell Biochem 188, 63–69. Thurnham DI, McCabe GP, Northrop-Clewes CA, Nestel P (2003). Effects of subclinical infectiion on plasma retinol concentrations and assessment of prevalence of vitamin A deficiency: meta analysis. Lancet 362, 2052–2058. Van Lettow M, Harries AD, Kumwenda JJ, Zijlstra EE, Clark TD, Taha TE et al. (2004). Micronutrient malnutrition and wasting in adults with pulmonary tuberculosis with and without HIV co-infection in Malawi. BMC Infect Dis 4, 61. Villamor E, Fawzi WW (2005). Effects of Vitamin A supplementation on immune responses and correlation with clinical outcomes. Clin Microbiol Rev 18, 446–464. Walravens PA (1979). Zinc metabolism and its implications in clinical medicine. West J Med 130, 133–142. Zachariah R, Spielmann MP, Harries AD, Salaniponi FM (2002). Moderate to severe malnutrition in patients with tuberculosis is a risk factor with early death. Trans R Soc Trop Med Hyg 96, 291–294.

European Journal of Clinical Nutrition