Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J Public Health Policy. Author manuscript; available in PMC 2010 September 1.
Published in final edited form as:
PMCID: PMC2764270

Factors associated with serum retinol, alpha-tocopherol, carotenoids, and selenium in Hispanics with problems of HIV, chronic hepatitis C and drug use


The effects of hepatitis and drug use on nutritional problems in HIV infection have rarely been examined despite the importance of drug use in the global HIV pandemic. We examined the effects of HIV, hepatitis C and drug use on serum micronutrients in 300 USA Hispanic adults. Chronic hepatitis C infection was associated with lower serum retinol (−8.2 ug/dl, p<0.0001), alpha-tocopherol (−0.10 ln ug/dl, p=0.024), and carotenoids (−19.8 ug/dl, p<0.0001). HIV infection was associated with lower selenium (−6.1 ug/l, p=0.028). Elevated trigycerides in HIV infection were associated with higher serum retinol and alpha-tocopherol. Drug use was not independently associated with micronutrient alterations. We conclude that hepatitis C is an important determinant of low serum micronutrients, and should be considered in any nutritional assessment of HIV infected populations. Since the safety of micronutrient supplementation is not established, policy for appropriate HIV clinical care should distinguish between populations with and without hepatitis co-infection.

Keywords: HIV, drug use, hepatitis C, serum micronutrients


In many parts of the world, intravenous drug use (IDU) is an important driver of the human immunodeficiency virus (HIV) epidemic, and the co-occurrence of hepatitis infection is common in these populations. For example, 37% of HIV infected USA veterans have detectable antibodies to hepatitis C (HCV).1 In Eastern Europe, where IDU is the predominant mode of HIV transmission2, HCV or hepatitis B (HBV) co-infection rates are over 30% among high-risk populations, including female sex workers 3,4. In Iran, where IDU accounts for two-thirds of HIV infections, HCV co-infection rates are estimated to be 74%.5 Thus, in some areas of the world, the majority of HIV infected persons may be co-infected with HCV.

Nutritional and metabolic abnormalities have been frequently reported in both HIV and HCV infections. Serum micronutrient deficiencies were commonly observed in HIV infected persons prior to the availability of effective antiretroviral therapy (ART)614. Low serum micronutrients have also been reported in persons with chronic hepatitis infection.1517 In HIV infection, abnormal serum lipid profiles are seen, and have been related to HIV replication, low CD4 cell counts18, and ART19, while low cholesterol is found in chronic HCV infection 20. The metabolism of lipids and micronutrients, such as vitamin A, vitamin E and carotenoids, are intimately linked.2123 Alterations in lipid levels resulting from either HIV or chronic HCV infection can impact serum micronutrient levels. In HIV/HCV co-infected persons, both HIV and HCV infections have the potential to influence micronutrient and lipid levels.

The role of HCV infection as a potential determinant of serum micronutrient levels in HIV infected drug-using populations is not well described as few studies have evaluated HCV infection. Based on the published literature, we suspected that hepatitis infection and drug use might confound the effect of HIV infection on serum micronutrient status. Therefore, our study was designed to separate the effects of hepatitis infection and drug use from the effects of HIV on serum micronutrient levels. We hypothesized that, because of its impact on metabolism, chronic HCV infection would be a significant risk factor for low serum micronutrients in persons with HIV infection. We examined serum vitamin A (retinol), vitamin E (alpha-tocopherol), carotenoids, and selenium because of their importance in immune function. We also measured serum zinc and will present those results separately.

We have chosen to publish our data in a policy journal because the results of our study suggest that, in addressing the problem of micronutrient deficiencies in persons living with HIV/AIDS, we need to make a distinction between those who are infected with HIV alone and those who are co-infected with HIV and HCV. We present detailed analyses to support our conclusion that chronic HCV infection has a profound influence on serum micronutrient status, and believe our findings have important implications for policies on the nutritional assessment of persons living with HIV/AIDS and their appropriate clinical care, including recommendations for micronutrient supplementation. While the potential benefits of micronutrient supplementation in HIV infection are best evaluated by randomized trials, a thorough understanding of the determinants of serum micronutrient deficiencies is important to ensure that trials are targeted appropriately. The benefits micronutrient supplementation may differ among subgroups, including those who are or are not co-infected with HCV.

Materials and Methods


Between August 2002 and December 2006, we recruited to the present study all qualified persons from a cohort of Hispanic adults participating in a study of nutrition and drug use. In this population of Hispanics, drug use is an important risk factor for HIV infection.24 We recruited the cohort by street out-reach into one of four groups: HIV infected drug users, HIV-non-infected drug users, HIV infected persons who do not use drugs, and healthy Hispanic adults, who denied drug use. The exclusion criteria for the cohort were: pregnant at recruitment, non-HIV associated malignancies, refusal to consent to the release of medical records, a history of gender re-assignment, and renal dialysis. The Institutional Review Board of the Tufts Medical Center approved the study and participants gave written informed consent.


We asked the participants to come to the clinic in a fasting state, and the staff noted whether the participant reported eating within five hours of the visit. No participants were febrile. Using standardized interviews, Hispanic study staff collected information on medical history, including all current medications, alcohol and drug use. The protocol defined current drug use as reported use of opiates, cocaine or amphetamines within the last six months. We did not consider marijuana a drug of abuse. Past drug users were classified as non-drug users. The statistical programmer coded self-reported alcohol use in several ways as: current use of any alcohol (vs. no use); use of alcohol at an average frequency of at least once per day (vs. < once per day); or use of alcohol at an average frequency of at least three times per day (vs. < three times per day). We assessed smoking as current (vs. never or past). The interviewers measured body mass index (BMI) from height and weight,25 and body composition by bioelectric impedance analysis (RJL Systems Inc., Clinton, MI) using the Lukaski equations.2628

Micronutrient analyses

Clinic staff drew blood in trace element-free conditions, and stored it at −70 ° C until processing. The laboratory staff measured serum vitamin A (retinol) and vitamin E (alpha-tocopherol) by reverse-phase high performance liquid chromatography (ESA EZChrom Elite Chromatography Data System, ESA, Chelmsford, MA)29; carotenoids by spectrophotometry30; and selenium by graphite furnace atomic absorption spectrophotometry (AAanalyst 800, Perkin Elmer, Shelton, CT). We defined low values as: retinol <30 ug/dl; alpha-tocopherol < 500 ug/dl; carotenoids <29.5 ug/dl, which is the 5th percentile for the US general population level of serum carotenoids from the NHANES III survey 1988–1994; and selenium <85 ug/l.10

Other blood assays

The laboratory of the Massachusetts Department of Public Health conducted the hepatitis tests. We defined chronic viral hepatitis as a positive serum test for hepatitis B virus surface antigen (HBsAg Enzyme Immunoassay 3.0, Biorad Laboratories, Redmond, WA, USA) or hepatitis C RNA (AMPLICOR Hepatitis C Virus test, Version 2.0 Roche Molecular Systems Inc., Branchburg, NJ, USA). Since HBV infection was uncommon (<2%), we excluded data from six participants with HBV infection to focus on HCV (total N=300). The Tufts Medical Center laboratories conducted the following assays: HIV status (Genetic Systems HIV1/HIV2 Plus O EIA, Biorad Laboratories, Redmond, WA); high-sensitivity c-reactive protein (UniCel® DxC 800 Synchron® Clinical System and CRP High Sensitivity Reagent (Beckman Coulter, Fullerton, CA, USA); CD4 lymphocyte counts using a specific monoclonal antibody and fluorescence-activated cell-sorting analysis and alanine amino transferase (ALT), total cholesterol, and trigycerides (Synchron® Clinical System, Beckman Coulter, Fullerton, CA, USA). The virology laboratory measured HIV RNA (Roche Amplicor Monitor reverse transcriptase polymerase chain reaction, Roche Molecular Systems, Somerville NJ, USA). We set undetectable viral load (< 400 copies/ml) to a value of 200 copies/ml.

Assessment of dietary intake and supplement use

We measured nutrient intake with a food frequency questionnaire (FFQ) previously validated in this population.31,32 The nutrition staff calculated the nutrient profiles with the Nutrient Data System, version 4.06 (NDS, University of Minnesota, Minneapolis, MN). The interviewers asked study participants about the use of vitamin and food supplements.

Statistical Analyses

In tables 1 and and22 we used the F test, chi square test or Fisher’s exact test where appropriate to test the between-group differences at alpha = 0.05). To give a “cleaner” comparison group in the analyses shown in Table 2, we excluded data from 45 HIV/HCV uninfected drug users. Serum vitamin E was transformed using the natural logarithm. We treated the following variables as potential predictors of serum micronutrients: HIV infection, HCV infection, current drug use, dietary intake, vitamin supplement use, age, gender, alcohol use, smoking, BMI, cholesterol, and c-reactive protein. Since we had an a priori interest in all the variables we report the full models in Table 3 to allow readers to compare the predictors of each micronutrient. We examined triglycerides as a predictor of serum micronutrients in a subset (N=255) with available triglyceride data. There were too few data to assess men and women separately. We conducted the multivariate modeling using the general linear models procedure in SAS version 9.0 (SAS Institute, Cary, NC), and evaluated influential observations using Cook’s distance.

Table 1
Participant characteristics in relation to infection with HIV and/or chronic hepatitis C (HCV) infection, mean (SE)
Table 2
Micronutrient status in relation to infection with HIV and/or chronic hepatitis C (HCV)
Table 3
Results of multivariate analyses of predictors of serum micronutrients


Table 1 shows the characteristics of the 300 study participants, stratified by infection status. There were significant between-group differences in age, smoking, alcohol use, ALT, BMI and cholesterol. Supplement use was more common among the HIV infected groups. Selenium intake from food was lowest in the HIV infected group.

Among all participants with HIV infection, the unadjusted prevalence of low micronutrients was 13% (18/141) for vitamin A, 11% (15/141) for vitamin E, 10% (14/141) for carotenoids, and 18% (26/141) for selenium. In contrast, when we stratified the results by infection status, the prevalence of low values among participants with HIV mono-infection did not differ remarkably from healthy controls, while low values were common in those with HCV infection (Table 2).

The factors associated with the individual serum micronutrients in multivariate analyses are shown in Table 3. Infection with HIV was associated with a higher serum vitamin A and vitamin E, but lower selenium compared to uninfected controls. Chronic HCV infection was associated with lower serum vitamin A, vitamin E and carotenoids. There was no evidence of an interaction effect between HIV and HCV for any micronutrient. Alcohol (> one serving/day) was associated with higher vitamin A. Supplement use, but not food, was associated with higher serum vitamin E, while serum carotenoids were influenced by dietary carotenoids, but not by supplements. Accounting for fasting state did not alter the conclusions. Treatment with ART, CD4 cell count, and HIV viral load were not predictors of micronutrient status in HIV infected participants.

When we replaced the term for cholesterol in the models shown in Table 3 with a term for triglycerides, triglycerides were significantly associated with vitamin A (6.8 ug/dl, SE=1.4, p<0.0001), and HIV infection was no longer associated with higher vitamin A (3.0 ug/dl, SE= 1.9, p=0.11). Similarly, triglycerides were significantly associated with vitamin E (0.0017 ln ug/dl, SE=0.00016, p<0.0001), but the association of HIV infection with higher vitamin E was no longer significant in the model that included a term for triglycerides (0.052 ln ug/dl, SE=0.046, p=0.26). Carotenoids and selenium were not significantly associated with triglycerides.


Our study demonstrates that chronic HCV co-infection is an important determinant of lower serum vitamin A, vitamin E and carotenoids, independently of acute phase reaction, in persons with clinically stable HIV infection. Though it has long been recognized that hepatic disease leads to low serum vitamin A,1517,33 this effect has not been well described in the context of HIV infection.

Studies of HIV infected persons with a low representation of drug users (<10%) have shown a low prevalence of serum micronutrient deficiencies.1114 Studies with a greater representation of drug users (>20%) consistently demonstrate deficiencies, with rates of vitamin A deficiency varying from 15–80%.610 Baum and Shor-Posner compared micronutrient deficiencies in HIV infected drug users and homosexual men. Deficiencies in vitamin A, vitamin E and selenium were more common in drug users, at 55%, 47% and 7%, respectively, compared to homosexual men in who the corresponding rates of deficiency were 10%, 18% and 11%.34 Our data demonstrate that the high rate of micronutrient deficiencies seen in drug users is associated with chronic HCV infection rather than with dietary deficiencies, drug use, or HIV infection.

The depressive effect of HCV infection on serum micronutrients is of concern because a block in micronutrient metabolism in the liver may lead to functional deficiencies. Carney and Russell found in patients with hepatic and gastrointestinal disease that 71% of patients with serum vitamin A < 30 ug/dl had night blindness, a level of deficiency consistent with immune dysfunction.35 Overall, forty one (14%) of our participants had serum vitamin A below 30 ug/dl. Low serum vitamin A has been associated with a five to six-fold increased risk of mortality in HIV infected drug users.6 However, it is not clear that micronutrient supplementation would result in an appropriate serum response in the presence of chronic HCV infection, or whether hepatic disease might block micronutrient metabolism. In the presence of liver damage, supplementation may even be harmful.33 There is a need to clarify the safety and efficacy of micronutrient supplementation in the presence of chronic HCV infection.

When we included known predictors of serum micronutrients in multivariate models, HIV was associated with higher serum vitamin A and vitamin E. Tang et al.36 found that HIV infected drug users had significantly higher vitamin E compared to HIV-uninfected drug users, an effect the investigators attributed to protease inhibitor therapy. Protease inhibitor therapy is associated with elevated trigycerides.19 In our study, higher vitamin E in HIV infection was explained by higher triglycerides, but not by ART. Since the metabolism of vitamin E, vitamin A and lipids are interrelated,2022 lipid elevations induced by ART may elevate serum micronutrient levels.

Among the micronutrients examined here, we found HIV infection to be independently associated only with lower selenium. Baum and colleagues showed that selenium <85 ug/l was associated with a high risk of mortality in HIV patients.10 A randomized controlled trial of selenium supplementation found that HIV patients who had a positive serum response to selenium supplementation had a better controlled HIV viral load and improved CD4 cell counts compared to placebo-treated patients. 37 These data suggest a clinically important relationship of selenium to HIV disease progression that warrants further investigation, including studies to establish whether selenium can be of benefit to HIV/HCV co-infected persons.

In this study, drug use was not a predictor of alterations in serum micronutrients once HCV infection was accounted for. Alcohol use was associated with significantly higher serum vitamin A. Alcohol has been associated with higher vitamin E and selenium in a previous population-based study 38 and, thus, is not always associated with poor micronutrient status.

We recruited participants from the community, and therefore our estimates are likely a good reflection of the target population. Since the diet of Hispanics differs from that of other racial/ethnic groups, the estimates of specific nutrient deficiencies may not be generalizable. However, the results pertaining to the effects of chronic HCV infection on serum micronutrients should be widely generalizable. In the USA, ART has been available for over a decade, and the food supply is increasingly fortified with micronutrients. For these reasons we have not compared our study to those conducted in areas of the world where access to ART is limited and malnutrition is common.

Weaknesses of our study include the lack of information on the stage of liver disease. We may have misclassified self-reported lifestyle habits, which would underestimate associations. Because we used a food frequency questionnaire to measure dietary intake, we were unable to quantify specific micronutrient intakes.

In conclusion, our results demonstrate that HCV co-infection is an important determinant of micronutrient deficiencies in the context of HIV infection. Our data further suggest that the presence of HCV co-infection may confound the nutritional assessment of HIV infected individuals. The extent to which this has affected our general understanding of the problem of micronutrient deficiencies in HIV infection is unknown.

Our results have implications for policies related to the nutritional requirements of persons living with HIV/AIDS. Recommendations for nutritional assessment should include an evaluation of HCV infection status in areas of the world where HIV/HCV co-infection is common. Dietary supplementation should be approached with caution until there is a better understanding of how HCV infection alters the metabolism of micronutrients, and how these alterations impact the progression of HIV and HCV infections.


We thank the Massachusetts Department of Public Health for conducting the hepatitis serology. This study was supported by the National Institutes of Health (DA 11598, DA14501, DK4 5734-07, P308142853, P30 DA013868, M01-RR00054).


The authors have no conflicts of interest to declare.


1. Backus LI, Boothroyd D, Deyton LR. HIV, hepatitis C and HIV/hepatitis C virus co-infection in vulnerable populations. AIDS. 2005;19 (suppl 3):S13–9. [PubMed]
2. UNAIDS. AIDS Epidemic Update/Eastern Europe and Central Asia. 2006. [Accessed July 7, 2008]. Available at:
3. Lioznov D, Nikolaenko S, Sabadash N, Antonova T, Belayeva T. Prevalence of HIV, hepatitis B and C viruses in high risk population in St. Petersburg, Russia. Poster Exhibition: The 3rd IAS Conference on HIV Pathogenesis and Treatment: Abstract no. TUPe1.1C12.
4. Uusküla A, Fischer K, Raudne R, Kilgi H, Krylov R, Salminen M, et al. A study on HIV and hepatitis C virus among commercial sex workers in Tallinn. Sex Transm Infect. 2008;84:189–91. [PubMed]
5. Alavi SM, Etemadi A. HIV/HBV, HIV/HCV and HIV/HTLV-1 co infection among injecting drug user patients hospitalized at the infectious disease ward of a training hospital in Iran. Pak J Med Sci. 2007;23:510–3.
6. Semba RD, Graham NMH, Caiaffa WT, Margolick JB, Clement L, Vlahov D. Increased mortality associated with vitamin A deficiency during human immunodeficiency virus type I infection. Arch Intern Med. 1993;153:2149–54. [PubMed]
7. Karter DL, Karter AJ, Yarrish R, Patterson C, Kass PH, Nord J, et al. Vitamin A deficiency in non-vitamin-supplemented patients with AIDS: a cross-sectional study. J Acquir Immune Defic Syndr Hum Retrovirol. 1995;8:199–203. [PubMed]
8. Baum MK, Shor-Posner G, Zhang G, Lai H, Quesada JA, Campa A, et al. HIV-1 infection in women is associated with severe nutritional deficiencies. J Acquir Immune Defic Syndr Hum Retrovirol. 1997;16:272–8. [PubMed]
9. Greenberg BL, Semba RD, Vink PE, Farley JJ, Sivapalasingam M, Steketee RW, et al. Vitamin A deficiency and maternal-infant transmission of HIV in two metropolitan areas in the United States. AIDS. 1997;11:325–32. [PubMed]
10. Baum MK, Shor-Posner G, Lai S, Zhang G, Lai H, Fletcher MA, et al. High risk of HIV-related mortality is associated with selenium deficiency. J Acquir Immune Defic Syndr Hum Retro. 1997;15:370–4. [PubMed]
11. Coodley GO, Coodely MK, Nelson HD, Loveless MO. Micronutrient concentrations in the HIV wasting syndrome. AIDS. 1993;7:1595–1600. [PubMed]
12. Lacey CJ, Murphy ME, Sanderson MJ, Montiero EF, Vail A, Schorah CJ. Antioxidant micronutrients and HIV infection. Int J STD AIDS. 1996;7:485–9. [PubMed]
13. Beach RS, Mantero-Atienza E, Shor-Posner G, Javier JJ, Szapocznik J, Morgan R, et al. Specific nutrient abnormalities in asymptomatic HIV-1 infection. AIDS. 1992;6:701–8. [PubMed]
14. Skurnick JH, Bogden JD, Baker H, Kemp FW, Sheffet A, Quattrone G, Louria DB. Micronutrient profiles in HIV-1 infected heterosexual adults. J Acquir Immune Defic Syndr Hum Retro. 1996;12:75–83. [PubMed]
15. Harris AD, Moore T. Vitamin A in infective hepatitis. Br Med J. 1947;1(4503):553–9. [PMC free article] [PubMed]
16. Smith FR, Goodman DS. The effects of diseases of the liver, thyroid and kidneys on the transport of vitamin A in human plasma. J Clin Invest. 1971;50:2426–36. [PMC free article] [PubMed]
17. Rocchi E, Casalgrandi G, Ronzoni A, Rosa MC, Cioni G, Marazzi A, et al. Antioxidant liposoluble vitamins and carotenoids in chronic hepatitis. Eur J Intern Med. 2001;12:116–21. [PubMed]
18. El-Sadr WM, Mullin CM, Carr A, Gibert C, Rappoport C, Visnegarwala F, et al. Effects of HIV disease on lipid, glucose and insulin levels: results from a large anti-retroviral-naïve cohort. HIV Med. 2005;6:114–21. [PubMed]
19. Carr A, Samaras K, Thorisdottir A, Kaufmann GR, Chisholm DJ, Cooper DA. Diagnosis, prediction, and natural course of HIV-1 protease-inhibitor-associated lipodistrophy, hyperlipidemia, and diabetes mellitus: a cohort study. Lancet. 1999;353:2093–9. [PubMed]
20. Maggi G, Bottelli R, Gola D, Perricone G, Posca M, Zavaglia C, et al. Serum cholesterol and chronic hepatitis C. Ital J Gastroenterol. 1996;28:436–40. [PubMed]
21. Infante M, Patorino U, Chiesa G, Bera E, Pisani P, Valente M, et al. Laboratory evaluation during high-dose vitamin A administration: a randomized study on lung cancer patients after surgical resection. J Cancer Res Clin Oncol. 1991;117:156–62. [PubMed]
22. Traber MG. Vitamin E regulation. Curr Opin Gastroenterol. 2005;21:223–7. [PubMed]
23. Cartmel B, Dziura J, Cullen MR, Vegso S, Omenn GS, Goodman GE, et al. Changes in cholesterol and triglyceride concentrations in the Vanguard population of the carotene and retinol efficacy trial (CARET) Eur J Clin Nutr. 2005;59:1173–80. [PubMed]
24. Massachusetts Department of Public Health. Who is most at risk of HIV infection? Massachusetts HIV/AIDS Data Fact Sheet [Online] Dec2005. [Accessed June 2006]. Available from:
25. Lohman TG, Roche AF, Martorell R, editors. Anthropometric standardization reference manual. Champaign (IL): Human Kinetics Books; 1988.
26. Lukaski HC, Bolonchuk WW. Theory and validation of the tetrapolar bioelectrical impedance method to assess human body composition. In: Ellis KJ, Yasamura S, Morgan WD, editors. In Vivo Body Composition Studies. London, UK: The Institute of Physical Sciences in Medicine; 1987. pp. 410–4.
27. Lukaski HC, Johnson PE, Bolonchuk WW, Lykken GI. Assessment of fat-free mass using bioelectrical impedance measurements of the human body. Am J Clin Nutr. 1985;41:810–7. [PubMed]
28. Forrester JE, Sheehan HMB, Joffe TH. A validation study of body composition by bioelectrical impedance analysis in HIV-positive and HIV-negative Hispanic men and women. J Am Diet Assoc. 2008;108:534–8. [PubMed]
29. Bieri JG, Tolliver TJ, Catignani GL. Simultaneous determination of alpha-tocopherol and retinol in plasma or red cells by high pressure liquid chromatography. Am J Clin Nutr. 1979;32:2143–9. [PubMed]
30. Neeld JB, Jr, Pearson WN. Micro- and macro-methods for the determination of serum vitamin A using trifluoroacetic acid. J Nutr. 1963;79:454–62. [PubMed]
31. Sahni S, Forrester JE, Tucker KL. Assessing dietary intake of drug abusing Hispanic adults with and without human immunodeficiency virus infection. J Am Diet Assoc. 2007;107:968–76. [PubMed]
32. Tucker KL, Bianchi LA, Maras J, Bermudez OI. Adaptation of a food frequency questionnaire to assess the diets of Puerto Ricans and non-Hispanic adults. Am J Epidemiol. 1998;148:507–18. [PubMed]
33. Mernitz H, Wang X-D. The bioconversion of carotenoids into vitamin A: implications for cancer prevention. In: Loessing IT, editor. Vitamin A: New Research. Hauppauge, NY: Nova Science Publishers; 2007. pp. 1–19.
34. Baum MK, Shor-Posner G. Micronutrient status in relationship to mortality in HIV-1. Nutr Rev. 1998;56:S135–9. [PubMed]
35. Carney EA, Russell RM. Correlation of dark adaptation test results with serum vitamin A levels in diseased adults. J Nutr. 1980;110:552–7. [PubMed]
36. Tang AM, Smit E, Semba R, Shah N, Lyles CM, Li D, et al. Improved antioxidant status among HIV infected injecting drug users on potent antiretroviral therapy. J Acquir Immune Defic Syndr. 2000;23:321–6. [PubMed]
37. Hurwitz BE, Klaus JR, Llabre M, Gonzalez A, Lawrence PJ, Maher KJ, et al. Suppression of human immunodeficiency virus type 1 viral load with selenium supplementation: a randomized controlled trial. Arch Intern Med. 2007;167:148–54. [PubMed]
38. Galan P, Viteri FE, Bertrais S, Czernichow S, Faure H, Arnaud J, et al. Serum concentrations of β-carotene, vitamins C and E, and zinc and selenium are influenced by sex, age, diet, smoking status, alcohol consumption and corpulence in a general French adult population. Eur J Clin Nutr. 2005;59:1181–90. [PubMed]