Abstract
Vitamin A supplementation starting at 6 months of age is an important child survival intervention; however, not much is known about the association between vitamin A status before 6 months and mortality among children born to HIV-infected women. Plasma concentrations of vitamins A and B-12 were available at 6 weeks of age (n = 576 and 529, respectively) for children born to HIV-infected women and they were followed up for morbidity and survival status until 24 months after birth. Children in the highest quartile of vitamin A had a 49% lower risk of death by 24 months of age compared to the lowest quartile (HR: 0.51, 95% CI: 0.29–0.90; P-value for trend = 0.01). Higher vitamin A levels were protective in the sub-groups of HIV-infected and un-infected children but this was statistically significant only in the HIV-uninfected subgroup. Higher vitamin A concentrations in plasma are protective against mortality in children born to HIV-infected women.
Introduction
Globally, close to 140 million children of preschool age suffer from vitamin A deficiency and 43.2 million of these are in sub-Saharan Africa [1]. Supplementation with vitamin A in children aged 6 months and above is an important child survival intervention. However, the benefits of supplementation before 6 months of age are unclear, especially among children born to HIV-infected women. Trials that assessed the efficacy of vitamin A supplementation in the neonatal period showed mixed results [2–5], while supplementation between the ages of 1 and 5 months was not beneficial [6–9].
Children in developing countries are at risk of developing deficiencies of multiple micronutrients [10]. Approximately 30% children aged 5–14 years in a study from Kenya had severe deficiency of vitamin B-12 while 11% of children under 7 years were B-12 deficient in Venezuela [11, 12]. HIV-infected women have a higher risk of micronutrient deficiencies including vitamin B-12 and the risk may extend to children born to these women [13]. Low serum vitamin B-12 levels were associated with low CD4 counts and progression to AIDS and death in adults with HIV [14, 15]. However, the associations between vitamin B-12 levels in children and child morbidity and mortality have not been studied.
This prospective study among HIV-infected and uninfected children born to HIV-infected women in Tanzania examined the association between plasma concentrations of vitamins A and B-12 at 6 weeks and 6 months of age and mortality as well as respiratory and diarrheal morbidity through 24 months of age.
Methods
This prospective study was conducted among children born to HIV-infected women in Dar es Salaam, Tanzania, within a randomized trial setting. Details of the trial have been published elsewhere [16]. HIV-infected pregnant women were enrolled between 12 and 27 weeks of gestation and randomized in a double-blind factorial design to one of four arms: vitamin A (30 mg of beta carotene plus 5000 IU of preformed vitamin A); multivitamins excluding vitamin A (20 mg of vitamin B-1, 20 mg of vitamin B-2, 25 mg of vitamin B-6, 100 mg of Niacin, 50 µg of vitamin B-12, 500 mg of vitamin C, 30 mg of vitamin E and 0.8 mg of folic acid); multivitamins including vitamin A (same doses as above); or placebo. All women received daily iron and folic acid and weekly chloroquine phosphate. Women in the vitamin A arms received 200 000 IU of vitamin A at delivery while others received placebo.
Infant birth weight was measured immediately after birth and low birth weight was defined as birth weight <2500 g. Children were followed from birth through monthly visits to the study clinic and home visits were made in case of a missed clinic visit. Child mortality was based on maternal report of the child's death. During clinic visits starting at 6 weeks of age, children underwent a complete physical examination. Respiratory infection was defined based on the presence of cough along with rapid respiratory rate (breaths per minute ≥50 for infants, ≥40 for children older than 1 year). Diarrhea was defined based on maternal report of three or more watery stools in a 24 h period during the past month. As per national guidelines, all children in the study received 6-monthly doses of oral vitamin A (100 000 IU at 6 months and 200 000 IU at the age of 1 year and thereafter). Mothers provided a blood sample for assessment of absolute counts of CD4 T-cells at enrolment (using the FACScan and FACScount system, Beckton-Dickinson, San Jose, CA). Children were provided a blood sample at birth, at 6 weeks and every 3 months thereafter for assessment of HIV infection, and at birth and every 6 months thereafter for assessment of absolute CD4 T-cell counts. HIV-1 infection was diagnosed in the children on the basis of a positive polymerase chain reaction (PCR) before 18 months of age (using the Amplicor HIV-1 detection kit, Roche Diagnostics, Branchburg, NJ) or a positive enzyme linked immunosorbent assay (ELISA) confirmed by a western blot, at or after 18 months of age. Child CD4 counts were measured as in the mothers. Blood draws for measurement of serum concentrations of vitamin A and vitamin B-12 were scheduled to occur at 6 weeks and 6 months of age. Plasma vitamin A was measured by high-performance liquid chromatography using the Shimadzu system [17] and vitamin B-12 was analyzed by a competitive magnetic separation assay on the Technico Immuno-1 analyzer (Bayer, Tarryton, NY, USA).
Of the 1078 HIV-infected women enrolled in the trial 984 children were born alive including 939 singletons. Plasma vitamin A concentrations were available for 576 children at 6 weeks and for 425 children at 6 months while vitamin B-12 concentrations were available for 529 children at 6 weeks and 345 children at 6 months. Baseline characteristics were compared across quartiles of vitamin levels using the Chi-square test for binary variables and Kruskal–Wallis test for continuous variables.
We examined the association between plasma concentrations of vitamin A and vitamin B-12 and child mortality through 24 months of age using Cox proportional hazards models (Proc Phreg in SAS) and a counting process data structure was used to stratify analyses by HIV-infected and un-infected person-time till 24 months of age [18, 19]. For the ‘HIV infected by 24 months’ endpoint, the study population consisted of a subgroup of these children not known to be HIV-infected at 6 weeks. A composite endpoint ‘HIV positive or dead by 24 months’ was defined among these children not infected at 6 weeks, depending on whether the child survived up to 24 months of age and remained free of HIV infection.
Generalized estimating equations (GEE) (Proc Genmod in SAS) were used to examine the associations between time-varying vitamin A and B-12 concentrations and morbidity [20]. The missing indicator method was used for covariates with missing data [21]. The study protocol was approved by the relevant ethics committees in Tanzania and Boston.
Results
Plasma vitamin A levels were available for 576 children at 6 weeks (median = 6.1 weeks; 25th, 75th percentile = 5.9, 7.1) and for 425 children at 6 months (median = 5.9 months; 25th, 75th percentile = 5.9, 6.1). Among these children, 110 died by the age of 24 months and the median time to death was 9 months (25th percentile, 75th percentile = 5, 15). At 6 weeks, 89% of the children had plasma vitamin A concentration <0.70 µmol/l and 22% had plasma vitamin A concentration <0.35 µmol/l while at 6 months, 83% and 20% of the children had levels below these thresholds respectively, which are commonly used for defining deficiency and severe deficiency [22]. Plasma vitamin B-12 levels were available for 529 children at 6 weeks (median = 6.1 weeks; 25th, 75th percentile = 5.9, 7.1) and for 345 children at 6 months (median = 6.0 months; 25th, 75th percentile = 5.9, 6.1). There were 100 deaths in this group by the age of 24 months and the median time to death was 9 months (25th percentile, 75th percentile = 5, 15). The prevalence of vitamin B-12 deficiency as defined by a plasma level of <148 pmol/l [23] was 8% at 6 weeks and 7% at 6 months.
The baseline characteristics of the children across quartiles of vitamin A and vitamin B-12 are shown in Table 1. In multivariate models children in the highest quartile of vitamin A had a 49% lower risk of mortality (HR = 0.51, 95% CI = 0.29–0.90) (Table 2) as compared to those in the lowest quartile. There was no significant association between plasma vitamin A concentration and HIV transmission (Table 2) or with symptoms indicative of respiratory infection or diarrhea (Table 3). In the univariate model, higher concentration of plasma vitamin B-12 was associated with decreased risk of mortality (HR comparing top to bottom quartile = 0.48). However, after adjusting for time-varying HIV status, plasma vitamin B12 concentration showed no significant association with mortality (HR comparing top to bottom quartile = 0.98, P-value test for trend = 0.86) (Table 4). There was also no association with HIV transmission (P-value test for trend = 0.37), HIV-free survival (P-value test for trend = 0.42) (Table 4) or with symptoms indicative of respiratory infection or diarrhea (Table 5).
Table 1.
Baseline characteristics of children by quartiles of plasma vitamin A and vitamin B-12 at 6 weeks
| Quartiles of plasma vitamin A concentrations: median (25th, 75th percentile) (µmol/l) | |||||
|---|---|---|---|---|---|
| Characteristics | 1 0.30 (0.25, 0.34) | 2 0.41 (0.38, 0.43) | 3 0.50 (0.48, 0.53) | 4 0.68 (0.61, 0.77) | P-valuea |
| Albumin (g/l)b | 30 (29, 42) | 32 (30, 41) | 31 (29, 40) | 32 (31, 39) | <0.0001 |
| Vitamin B-12 (pmol/l)b | 319 (216, 467) | 341 (247, 462) | 273 (202, 381) | 283 (202, 389) | 0.0008 |
| CD 4 cell count (cells/mm3)b | 1136 (833, 1478) | 1297 (967, 1694) | 1171 (918, 1509) | 1291 (905, 1630) | 0.09 |
| HIV infected by 6 weeks (%) | 33 | 17 | 28 | 22 | 0.32 |
| Maternal CD4 count during pregnancy <350 cells/mm3 (%) | 28 | 21 | 28 | 23 | 0.73 |
| Low birth weight (%) | 41 | 25 | 18 | 16 | 0.09 |
| Mother has no formal education (%) | 29 | 23 | 31 | 17 | 0.68 |
| Maternal trial regimen | |||||
| Multivitamins (%) | 31 | 25 | 24 | 20 | 0.001 |
| Vitamin A (%) | 21 | 19 | 26 | 34 | <0.0001 |
| Quartiles of plasma vitamin B-12 concentrations: median (25th, 75th percentile) (pmol/l) | |||||
| Characteristics | 1 165 (143, 196) | 2 251 (232, 277) | 3 364 (333, 391) | 4 542 (465, 671) | P-valuea |
| Albumin (g/l)b | 31 (29, 41) | 31 (29, 39) | 32 (30, 41) | 32 (30, 42) | 0.04 |
| Vitamin A (µmol/l)b | 0.49 (0.36, 0.59) | 0.48 (0.40, 0.58) | 0.46 (0.37, 0.59) | 0.41 (0.32, 0.51) | 0.0009 |
| CD 4 cell count (cells/mm3)b | 1244 (975, 1570) | 1194 (953, 1648) | 1157 (784, 1589) | 1263 (876, 1694) | 0.57 |
| HIV infected by 6 weeks (%) | 38 | 26 | 19 | 18 | 0.01 |
| Maternal CD4 count during pregnancy <350 cells/mm3 (%) | 30 | 24 | 25 | 21 | 0.16 |
| Low birth weight (%) | 35 | 30 | 15 | 20 | 0.22 |
| Mother has no formal education (%) | 30 | 36 | 18 | 15 | 0.23 |
| Mternal trial regimen | |||||
| Multivitamins (%) | 10 | 18 | 31 | 41 | <0.0001 |
| Vitamin A (%) | 26 | 27 | 23 | 24 | 0.51 |
aBased on Kruskal–Wallis test for continuous variables and Chi-square test for binary variables.
bMedian (25th, 75th percentile).
Table 2.
Association of plasma vitamin A in children with mortality and HIV transmission
| Endpoint | Quartiles of vitamin A | Child-months at risk | No. of outcomes | Hazard ratioa (univariate) | Hazard ratiob (adjusted) | 95% CIb (adjusted) | P-valuec | |
|---|---|---|---|---|---|---|---|---|
| Death by 24 months | All children | 1 | 2731 | 45 | 1.00 | 1.00 | ||
| 2 | 2500 | 24 | 0.59 | 0.75 | 0.45–1.24 | |||
| 3 | 2841 | 22 | 0.48 | 0.59 | 0.34–1.00 | 0.01 | ||
| 4 | 2910 | 19 | 0.40 | 0.51 | 0.29–0.90 | |||
| HIV infected | 1 | 808 | 36 | 1.00 | 1.00 | |||
| 2 | 489 | 16 | 0.75 | 0.73 | 0.40–1.35 | |||
| 3 | 509 | 17 | 0.73 | 0.64 | 0.34–1.18 | 0.18 | ||
| 4 | 531 | 17 | 0.71 | 0.69 | 0.37–1.28 | |||
| HIV un-infected | 1 | 1923 | 9 | 1.00 | 1.00 | |||
| 2 | 2011 | 8 | 0.85 | 0.93 | 0.35–2.47 | |||
| 3 | 2332 | 5 | 0.46 | 0.43 | 0.14–1.30 | 0.01 | ||
| 4 | 2378 | 2 | 0.18 | 0.17 | 0.04–0.80 | |||
| Death by 12 months | 1 | 1388 | 30 | 1.00 | 1.00 | |||
| 2 | 1235 | 11 | 0.42 | 0.51 | 0.25–1.03 | |||
| 3 | 1399 | 10 | 0.33 | 0.40 | 0.19–0.84 | 0.19 | ||
| 4 | 1427 | 15 | 0.49 | 0.71 | 0.36–1.38 | |||
| Death in 2nd year of life | 1 | 1343 | 15 | 1.00 | 1.00 | |||
| 2 | 1265 | 13 | 0.92 | 1.25 | 0.57–2.76 | |||
| 3 | 1442 | 12 | 0.75 | 1.03 | 0.46–2.35 | 0.03 | ||
| 4 | 1483 | 4 | 0.24 | 0.31 | 0.10–0.98 | |||
| HIV infected by 24 monthsc | 1 | 1716 | 18 | 1.00 | 1.00 | |||
| 2 | 1755 | 15 | 0.83 | 0.88 | 0.44–1.78 | |||
| 3 | 1757 | 11 | 0.60 | 0.66 | 0.30–1.42 | 0.41 | ||
| 4 | 1890 | 22 | 1.12 | 1.28 | 0.65–2.49 | |||
| HIV infected or death by 24 monthsc | 1 | 1716 | 24 | 1.00 | 1.00 | |||
| 2 | 1755 | 20 | 0.84 | 0.85 | 0.47–1.56 | |||
| 3 | 1757 | 13 | 0.54 | 0.54 | 0.27–1.08 | 0.81 | ||
| 4 | 1890 | 23 | 0.89 | 0.91 | 0.50–1.67 |
Quartile divisions (µmol/l) were ≤0.36, 0.37–0.45, 0.46–0.57, ≥0.58 (at 6 weeks) and ≤0.37, 0.38–0.49, 0.50–0.62, ≥0.63 (at 6 months).
aFrom separate univariate Cox proportional hazards models with time-varying indicators for quartiles of vitamin A concentrations.
bFrom separate multivariable Cox proportional hazards models with time-varying indicators for quartiles of vitamin A and the following adjusting variables: maternal trial regimen, education and CD4 count during pregnancy, and the following child variables; low birth weight, CD4 counts and plasma vitamin B-12 and albumin. Model for all children included an indicator for time-varying HIV status.
cP-value for trend based on median concentration in each quartile. Among those not known to be HIV infected at 6 weeks. In this group quartile divisions (µmol/l) were ≤0.37, 0.38–0.46, 0.47–0.57, ≥0.58 (at 6 weeks) and ≤0.39, 0.38–0.51, 0.52–0.63, ≥0.64 (at 6 months).
Table 3.
Association of plasma vitamin A in children with morbidity
| Quartiles of plasma vitamin Aa | Morbidity rate (episodes/child-year) | Relative riskb (univariate) | Relative riskc (multivariate) | 95% CIc | P-value for trendc |
|---|---|---|---|---|---|
| All children | |||||
| Respiratory infectiond | |||||
| 1 | 0.57 | 1.00 | 1.00 | ||
| 2 | 0.61 | 0.88 | 0.94 | 0.67–1.31 | |
| 3 | 0.58 | 0.62 | 0.70 | 0.48–1.01 | |
| 4 | 0.55 | 0.71 | 0.78 | 0.52–1.17 | 0.18 |
| Diarrheae | |||||
| 1 | 1.38 | 1.00 | 1.00 | ||
| 2 | 1.53 | 1.06 | 1.10 | 1.00–1.21 | |
| 3 | 1.50 | 1.05 | 1.09 | 0.99–1.21 | |
| 4 | 1.45 | 1.06 | 1.11 | 0.99–1.23 | 0.15 |
| HIV infected | |||||
| Respiratory Infection | |||||
| 1 | 0.95 | 1.00 | 1.00 | ||
| 2 | 1.21 | 0.73 | 0.85 | 0.45–1.61 | |
| 3 | 0.98 | 0.52 | 0.55 | 0.27–1.14 | 0.14 |
| 4 | 0.86 | 0.59 | 0.56 | 0.25–1.26 | |
| Diarrhea | |||||
| 1 | 1.80 | 1.00 | 1.00 | ||
| 2 | 2.06 | 1.08 | 1.13 | 0.95–1.35 | |
| 3 | 1.94 | 1.01 | 1.09 | 0.90–1.33 | 0.60 |
| 4 | 1.87 | 1.01 | 1.07 | 0.89–1.30 | |
| HIV un-infected | |||||
| Respiratory Infection | |||||
| 1 | 0.42 | 1.00 | 1.00 | ||
| 2 | 0.46 | 1.13 | 1.11 | 0.74–1.68 | |
| 3 | 0.49 | 0.82 | 0.81 | 0.50–1.30 | 0.57 |
| 4 | 0.46 | 0.91 | 0.91 | 0.55–1.48 | |
| Diarrhea | |||||
| 1 | 1.22 | 1.00 | 1.00 | ||
| 2 | 1.39 | 1.07 | 1.08 | 0.96–1.21 | |
| 3 | 1.41 | 1.09 | 1.08 | 0.95–1.22 | 0.26 |
| 4 | 1.33 | 1.10 | 1.10 | 0.96–1.26 | |
Quartile divisions for plasma vitamin A (µmol/l) were ≤0.36, 0.37–0.45, 0.46–0.57, ≥0.58 (at 6 weeks) and ≤0.37, 0.38–0.49, 0.50–0.62, ≥0.63 (at 6 months).
aFrom generalized estimating equations using a binomial distribution and a log link function.
bFrom separate multivariate models and adjusted for maternal trial regimen, maternal education, maternal baseline CD4 count and the following child factors: age, HIV status (as time-varying), low birth weight, baseline CD4 count, plasma vitamin B-12 and albumin.
cP-value for trend based on median concentration in each quartile.
dDefined as presence of cough along with rapid respiratory rate.
eDefined as three or more watery stools in a 24 h period.
Table 4.
Association of plasma vitamin B-12 in children with mortality and HIV transmission
| Endpoint | Quartiles of vitamin B-12 | Child months at risk | No. of outcomes | Hazard ratioa (univariate) | Hazard ratiob (adjusted) | 95% CIb (adjusted) | P-valuec | |
|---|---|---|---|---|---|---|---|---|
| Death by 24 months | All children | 1 | 2536 | 37 | 1.00 | 1.00 | ||
| 2 | 2500 | 25 | 0.69 | 0.97 | 0.57–1.63 | |||
| 3 | 2467 | 20 | 0.56 | 1.04 | 0.55–1.96 | 0.86 | ||
| 4 | 2606 | 18 | 0.48 | 0.98 | 0.50–1.92 | |||
| HIV infected | 1 | 792 | 32 | 1.00 | 1.00 | |||
| 2 | 577 | 19 | 0.80 | 0.84 | 0.47–1.51 | |||
| 3 | 438 | 14 | 0.78 | 0.85 | 0.40–1.79 | 0.91 | ||
| 4 | 368 | 14 | 0.95 | 1.01 | 0.48–2.16 | |||
| HIV un-infected | 1 | 1744 | 5 | 1.00 | 1.00 | |||
| 2 | 1922 | 6 | 1.09 | 1.52 | 0.44–5.33 | |||
| 3 | 2029 | 6 | 1.03 | 1.56 | 0.42–5.76 | 0.68 | ||
| 4 | 2238 | 4 | 0.63 | 0.85 | 0.20–3.71 | |||
| Death by 12 months | 1 | 1290 | 25 | 1.00 | 1.00 | |||
| 2 | 1238 | 14 | 0.58 | 0.77 | 0.40–1.49 | |||
| 3 | 1213 | 10 | 0.42 | 0.85 | 0.37–1.98 | 0.71 | ||
| 4 | 1269 | 11 | 0.45 | 1.21 | 0.52–2.81 | |||
| Death in 2nd year of life | 1 | 1246 | 2 | 1.00 | 1.00 | |||
| 2 | 1262 | 11 | 0.91 | 1.29 | 0.53–3.11 | |||
| 3 | 1254 | 10 | 0.83 | 1.40 | 0.50–3.89 | 0.43 | ||
| 4 | 1337 | 7 | 0.54 | 0.78 | 0.26–2.33 | |||
| HIV infected by 24 monthsd | 1 | 1512 | 13 | 1.00 | 1.00 | |||
| 2 | 1547 | 17 | 1.31 | 1.46 | 0.69–3.07 | |||
| 3 | 1466 | 12 | 0.95 | 1.00 | 0.43–2.34 | 0.37 | ||
| 4 | 1624 | 12 | 0.87 | 0.79 | 0.33–1.90 | |||
| HIV infected or death by 24 monthsd | 1 | 1512 | 17 | 1.00 | 1.00 | |||
| 2 | 1547 | 19 | 1.12 | 1.24 | 0.63–2.43 | |||
| 3 | 1466 | 16 | 0.97 | 1.07 | 0.51–2.25 | 0.42 | ||
| 4 | 1624 | 14 | 0.78 | 0.76 | 0.34–1.68 |
Quartile divisions (pmol/l) were ≤213.06, 213.07–299.62, 299.63–431.28, ≥431.29 (at 6 weeks) and ≤223.61, 223.62–323.75, 323.76–455.63, ≥455.64 (at 6 months).
aFrom separate univariate Cox proportional hazards models with time-varying indicators for quartiles of vitamin B-12 concentration.
bFrom multivariable Cox proportional hazards models with time-varying indicators for quartiles of vitamin B-12 and the following adjusting variables: maternal trial regimen, maternal education, maternal CD4 count during pregnancy and the following child variables; low birth weight, CD4 counts and plasma vitamin A and albumin.
cP-value for trend based on median concentration in each quartile Multivariate model for all children included an indicator for time-varying HIV status.
dAmong those not known to be HIV infected at 6 weeks. In this group quartile divisions (pmol/l) were ≤223.46, 223.46–313.30, 313.31–442.23, ≥442.24 (at 6 weeks) and ≤242.21, 242.22–347.11, 347.12–476.22, ≥476.23 (at 6 months).
Table 5.
Association of plasma vitamin B-12 in children with morbidity
| Quartiles of plasma vitamin B-12 | Morbidity rate (episodes/child-year) | Relative riska (univariate) | Relative riskb (multivariate) | 95% CIb | P-value for trendc |
|---|---|---|---|---|---|
| All children | |||||
| Respiratory infectiond | |||||
| 1 | 0.59 | 1.00 | 1.00 | ||
| 2 | 0.62 | 0.84 | 0.86 | 0.57–1.31 | |
| 3 | 0.54 | 1.04 | 1.11 | 0.70–1.78 | |
| 4 | 0.58 | 0.95 | 1.04 | 0.64–1.67 | 0.66 |
| Diarrheae | |||||
| 1 | 1.52 | 1.00 | 1.00 | ||
| 2 | 1.48 | 0.98 | 1.01 | 0.90–1.14 | |
| 3 | 1.50 | 1.02 | 1.07 | 0.92–1.24 | |
| 4 | 1.40 | 0.96 | 1.04 | 0.88–1.23 | 0.72 |
| HIV infected | |||||
| Respiratory infection | |||||
| 1 | 1.07 | 1.00 | 1.00 | ||
| 2 | 1.06 | 0.95 | 0.74 | 0.32–1.67 | |
| 3 | 0.90 | 1.35 | 0.90 | 0.38–2.15 | |
| 4 | 0.75 | 1.50 | 1.11 | 0.49–2.55 | 0.62 |
| Diarrhea | |||||
| 1 | 2.13 | 1.00 | 1.00 | ||
| 2 | 1.99 | 0.94 | 0.89 | 0.75–1.06 | |
| 3 | 1.74 | 0.90 | 0.83 | 0.63–1.09 | |
| 4 | 1.73 | 0.87 | 0.85 | 0.64–1.12 | 0.24 |
| HIV un-infected | |||||
| Respiratory infection | |||||
| 1 | 0.37 | 1.00 | 1.00 | ||
| 2 | 0.48 | 0.86 | 0.87 | 0.51–1.49 | |
| 3 | 0.45 | 1.08 | 1.20 | 0.69–2.11 | |
| 4 | 0.56 | 0.97 | 1.14 | 0.63–2.06 | 0.47 |
| Diarrhea | |||||
| 1 | 1.26 | 1.00 | 1.00 | ||
| 2 | 1.32 | 1.05 | 1.08 | 0.92–1.26 | |
| 3 | 1.44 | 1.16 | 1.20 | 1.01–1.43 | |
| 4 | 1.35 | 1.09 | 1.17 | 0.95–1.43 | 0.18 |
Quartile divisions (pmol/l) were ≤213.06, 213.07–299.62, 299.63–431.28, ≥431.29 (at 6 weeks) and ≤223.61, 223.62–323.75, 323.76–455.63, ≥455.64 (at 6 months).
aFrom generalized estimating equations using a binomial distribution and a log link function.
bFrom separate multivariate models and adjusted for maternal trial regimen, maternal education, maternal baseline CD4 count and the following child factors: age, HIV status (as time-varying), low birth weight, baseline CD4 count, plasma vitamin A and albumin.
cP-value for trend based on median concentration in each quartile.
dDefined as presence of cough along with rapid respiratory rate.
eDefined as three or more watery stools in a 24 h period.
Discussion
We found that plasma vitamin A concentrations in children born to HIV-infected women in Tanzania were inversely associated with mortality up to 24 months of age. Higher plasma vitamin A concentration was protective against mortality among HIV-uninfected children in our study. Our results agree with those from a meta-analysis of eight community-based trials of vitamin A supplements to children aged 6–72 months which showed a 30% reduction in the risk of death in the supplemented group as compared to the control group [24]. Randomized trials that provided vitamin A supplements to neonates showed mixed results with two trials from India and Indonesia showing a protective effect on mortality while a third from Zimbabwe showing no effect or a harmful effect [2–5].
In our study, plasma vitamin A was not significantly associated with mother-to-child transmission (MTCT). In randomized trials from Tanzania [25] and Zimbabwe [5] vitamin A supplementation had an increased risk of transmission and/or mortality, while in two other trials vitamin A supplementation had no significant effect on transmission [26, 27]. Our findings on the relationship between plasma vitamin A and health outcomes are based on observational data analysis; although we have adjusted for a number of confounding variables, the presence of residual confounding cannot be excluded. Also, our study had limited statistical power to examine the association between plasma vitamin A and HIV transmission, especially in the first 6 months. We found no statistically significant association between vitamin A concentrations and symptoms indicative of respiratory infection or diarrhea. In a meta-analysis of five trials, vitamin A supplementation had no effect on the risk of acute respiratory tract infection [28]. However in trials from Ecuador [29] and Indonesia [30], vitamin A was associated with an apparently higher risk of respiratory infections among children who were ‘better nourished’ with a protective association among under-nourished children. A meta-analysis of eight clinical trials showed no overall effect of vitamin A supplementation on diarrhea with an equal number of studies showing positive and negative effects [31]. Our findings of a null relationship between plasma vitamin A and morbidity in the presence of a significant inverse association with mortality may also suggest that vitamin A is more important for reducing severity of infections rather than their incidence. One limitation of our study is that low plasma vitamin A concentration might be a result of an acute phase response and be a marker of advanced underlying disease [32]. However, this does not appear to be a major factor in our study as the protective effect of higher vitamin A concentrations were evident after adjusting for serum albumin levels which is known to be a negative acute phase reactant.
Vitamin A is essential for optimal functioning of the immune system [33]. Animal studies have shown a reduction in the number of circulating natural killer (NK) cells which are important for protection against viral infections during experimental vitamin A deficiency [34].
We did not find any association between plasma vitamin B-12 concentrations and mortality or morbidity. Low serum B-12 levels were associated with faster progression to AIDS in a cohort of HIV-infected bisexual and homosexual men in North America [35]. Multivitamin supplements including vitamin B-12 given to mothers during pregnancy and lactation have been shown to increase plasma levels of these vitamins in their children and was also protective against child morbidity in that cohort [36, 37]. Our study was limited by the use of plasma vitamin B-12 concentrations as the measure of vitamin B-12 deficiency as plasma levels may not accurately reflect bioavailable vitamin B-12 and total homocysteine or methymalonic acid in the blood may be better markers of B-12 status [38]. The prevalence of vitamin B-12 deficiency in this cohort of children (8% at 6 weeks and 7% at 6 months of age) was also lower than that noted in studies from South America.
Results from this study add to the evidence from clinical trials among children 6 months and older about the beneficial effect of vitamin A in children even in settings with high HIV prevalence. Periodic vitamin A supplementation of children starting at 6 months is an inexpensive and simple intervention and has a major role in reducing child mortality in countries where vitamin A deficiency is a public health problem. Results from clinical trials currently underway that are assessing the effect of supplementation among children with B-complex vitamins including vitamin B-12 will be able to shed more light on the role of vitamin B-12 in child health.
Funding
National Institute of Child Health and Human Development (NICHD R01 32257), the Fogarty International Center (NIH D43 TW00004) and the Harvard School of Public Health.
Acknowledgements
We gratefully acknowledge the contributions of the field staff in data collection and the women and children enrolled in the study.
References
- 1.Aguayo VM, Baker SK. Vitamin A deficiency and child survival in sub-Saharan Africa: reappraisal of challenges and opportunities. Food Nutri Bull. 2005;26:348–55. doi: 10.1177/156482650502600404. [DOI] [PubMed] [Google Scholar]
- 2.Rahmatullah L, Tielsch RD, Thulasiraj RD, et al. Impact of supplementing newborn infants with vitamin A on early infant mortality: community based randomized trial in southern India. Br Med J. 2003;327:254. doi: 10.1136/bmj.327.7409.254. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Humphrey JH, Agoestina T, Wu A, et al. Impact of neonatal vitamin A supplementation on infant morbidity and mortality. J Pediatr. 1996;128:489–496. doi: 10.1016/s0022-3476(96)70359-1. [DOI] [PubMed] [Google Scholar]
- 4.Malaba LC, Iliff PJ, Nathoo KJ, et al. Effect of postpartum maternal or neonatal vitamin A supplementation on infant mortality among infants born to HIV-negative mothers in Zimbabwe. Am J Clin Nutr. 2003;81:454–60. doi: 10.1093/ajcn.81.2.454. [DOI] [PubMed] [Google Scholar]
- 5.Humphrey JH, Iliff PJ, Marinda ET, et al. Effects of a single large dose of vitamin A given during the postpartum period to HIV-positive women and their infants on child HIV infection, infection-free survival and mortality. J Infect Dis. 2006;193:860–71. doi: 10.1086/500366. [DOI] [PubMed] [Google Scholar]
- 6.Daulaire NM, Starbuck ES, Houston RM, et al. Childhood mortality after a high dose of vitamin A in a high risk population. Br Med J. 1992;304:207–10. doi: 10.1136/bmj.304.6821.207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Rahamn MM, Mahalanabis D, Wahed MA, et al. Administration of 25,000 IU vitamin A doses at routine immunization in young infants. Eur J Clin Nutr. 1995;49:439–45. [PubMed] [Google Scholar]
- 8.West KP, Jr., Katz J, Shreshtha SR, et al. Mortality of infants <6 months of age supplemented with vitamin A: a randomized double masked trial in Nepal. Am J Clin Nutr. 1995;62:143–8. doi: 10.1093/ajcn/62.1.143. [DOI] [PubMed] [Google Scholar]
- 9.W.H.O./CHD. Randomized trial to assess benefits and safety of vitamin A supplementation linked to immunization in early infancy. Lancet. 1998;352:1257–63. [PubMed] [Google Scholar]
- 10.Gross R, Benade S, Lopez G. The International Research on Infant Supplementation Initiative. J Nutr. 2005;135:628S–30S. doi: 10.1093/jn/135.3.628S. [DOI] [PubMed] [Google Scholar]
- 11.Siekmann JH, Allen LH, Bwibo NO, et al. Kenyan school children have multiple micronutrient deficiencies, but increased plasma B-12 is the only detectable micronutrient response to meat or milk supplementation. J Nutr. 2003;133:3972S–80S. doi: 10.1093/jn/133.11.3972S. [DOI] [PubMed] [Google Scholar]
- 12.Garcia-Casal MN, Osorio C, Landaeta M, et al. High prevalence of folic acid and vitamin B12 deficiencies in infants, children, adolescents and pregnant women in Venezuela. Eur J Clin. Nutr. 2005;59:1064–1070. doi: 10.1038/sj.ejcn.1602212. [DOI] [PubMed] [Google Scholar]
- 13.Beach RS, Mantero-Atienza E, Shor-Posner G, et al. Specific nutrient abnormalities in asymptomatic HIV-1 infection. AIDS. 1992;6:701–8. doi: 10.1097/00002030-199207000-00013. [DOI] [PubMed] [Google Scholar]
- 14.Baum MK, Shor-Posner G, Lu Y, et al. Micronutrients and HIV-1 disease progression. AIDS. 1995;9:1051–6. doi: 10.1097/00002030-199509000-00010. [DOI] [PubMed] [Google Scholar]
- 15.Remacha AF, Riera A, Cadafalch J, et al. Vitamin B12 abnormalities in HIV infected patients. Eur J Hematol. 1991;47:60–4. doi: 10.1111/j.1600-0609.1991.tb00562.x. [DOI] [PubMed] [Google Scholar]
- 16.Fawzi WW, Msamanga GI, Spiegelman D, et al. Randomised trial of effects of vitamin supplements on pregnancy outcomes and T cell counts in HIV-1-infected women in Tanzania. Lancet. 1998;351:1477–82. doi: 10.1016/s0140-6736(98)04197-x. [DOI] [PubMed] [Google Scholar]
- 17.De Leenheer AP, De Bevere VO, De Rutyer MG, et al. Simultaneous determination of retinol and alpha-tocopherol in human serum by high-performance liquid chromatography. J Chromatogr. 1979;162:408–13. doi: 10.1016/s0378-4347(00)81528-5. [DOI] [PubMed] [Google Scholar]
- 18.Cox D. Regression models and life tables. J R Stat Soc. 1972;34:187–220. [Google Scholar]
- 19.Andersen P, Gill R. Cox's regression model counting process: a large sample study. Ann Statist. 1982;10:1110–20. [Google Scholar]
- 20.Diggle P, Liang K, Zeger S. Analysis of Longitudinal Data. London, UK: Oxford University Press; [Google Scholar]
- 21.Miettinen OS. Theoretical Epidemiology. New York: John Wiley and Sons; 1985. [Google Scholar]
- 22.Sommer A, Davidson FR. Assessment and control of vitamin A deficiency: The Annecy Accords. J Nutr. 2002;132:2845S–50S. doi: 10.1093/jn/132.9.2845S. [DOI] [PubMed] [Google Scholar]
- 23.Rogers LM, Boy E, Miller JW, et al. High prevalence of cobalamin deficiency in Guatemalan schoolchildren: associations with low plasma holotranscobablamin II and elevated serum methylmalonic acid and plasma homocysteine concentrations. Am J Clin Nutr. 2003;77:433–40. doi: 10.1093/ajcn/77.2.433. [DOI] [PubMed] [Google Scholar]
- 24.Fawzi W, Chalmers T, Herrera M, et al. Vitamin A supplementation and child mortality: a meta-analysis. JAMA. 1993;269:898–903. [PubMed] [Google Scholar]
- 25.Fawzi WW, Msamanga Gi, Hunter DJ, et al. Randomized trial of vitamin supplements in relation to transmission of HIV-1 through breastfeeding and early child mortality. AIDS. 2002;16:1935–44. doi: 10.1097/00002030-200209270-00011. [DOI] [PubMed] [Google Scholar]
- 26.Coutsoudis A, Pillay K, Spooner K, et al. Randomized trial testing the effect of vitamin A supplementation on pregnancy outcomes and early mother-to-child transmission in Durban, South Africa. AIDS. 1999;13:1517–24. doi: 10.1097/00002030-199908200-00012. [DOI] [PubMed] [Google Scholar]
- 27.Kumwenda K, Miotti PG, taha TE, et al. Antenatal vitamin A supplementation increases birth weight and dcreaes anemia among infants born to human immunodeficiency virus-infected women in Malawi. Clin Infect Dis. 2002;35:618–24. doi: 10.1086/342297. [DOI] [PubMed] [Google Scholar]
- 28.Vitamin A and Pneumonia Working Group. Potential interventions for the prevention of childhood pneumonia in developing countries: a meta-analysis of data from field trials to assess the impact of vitamin A supplementation on pneumonia morbidity and mortality. Bull World Health Organ. 1995;73:609–19. [PMC free article] [PubMed] [Google Scholar]
- 29.Sempertegui F, Estrella B, Camaniro V, et al. The beneficial effects of weekly low-dose vitamin A supplementation on acute lower respiratory infections and diarrhea in Ecuadorian children. Pediatrics. 1999;104:e1. doi: 10.1542/peds.104.1.e1. [DOI] [PubMed] [Google Scholar]
- 30.Dibley MJ, Sadjimin T, Kjolhede CL, et al. Vitamin A supplementation fails to reduce incidence of acute respiratory illness and diarrhea in preschool-age Indonesian children. J Nutr. 1996;126:434–42. doi: 10.1093/jn/126.2.434. [DOI] [PubMed] [Google Scholar]
- 31.Grotto I, Mimouni M, Gdalevich M, et al. Vitamin A supplementation and childhood morbidity from diarrhea and respiratory infections: a meta-analysis. J Pediatr. 2003;142:297–304. doi: 10.1067/mpd.2003.116. [DOI] [PubMed] [Google Scholar]
- 32.Baeten JM, McClelland RS, Richardson BA, et al. Vitamin A deficiency and the acute phase response among HIV-1-infected and uninfected women in Kenya. J Acquir Immune Defic Syndr. 2002;31:243–9. doi: 10.1097/00126334-200210010-00016. [DOI] [PubMed] [Google Scholar]
- 33.Reifen R. Vitamin A as an anti-inflammatory agent. Proc Nutr Soc. 2002;61:397–400. doi: 10.1079/PNS2002172. [DOI] [PubMed] [Google Scholar]
- 34.Green HN, Mellanby E. Vitamin A as an anti-infective agent. Br Med J. 1928;2:691–6. doi: 10.1136/bmj.2.3537.691. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Tang AM, Graham NMH, Chandra RK, et al. Low serum vitamin B-12 concentrations are associated with faster human immunodeficiency virus type 1(HIV-1) disease progression. J Nutr. 1997;127:345–51. doi: 10.1093/jn/127.2.345. [DOI] [PubMed] [Google Scholar]
- 36.Baylin A, Villamor E, Rifai N, et al. Effect of vitamin supplementation to HIV-infected pregnant women on the micronutrient status of their children. Eur J Clin Nutr. 2005;59:960–968. doi: 10.1038/sj.ejcn.1602201. [DOI] [PubMed] [Google Scholar]
- 37.Fawzi WW, Msamanga GI, Wei R, et al. Effects of providing vitamin supplements to human immunodeficiency virus-infected, lactating mothers on child's morbidity and CD4 cell counts. Clin Infect Dis. 2003;36:1053–62. doi: 10.1086/374223. [DOI] [PubMed] [Google Scholar]
- 38.Wickramasinghe SN, Fida S. Correlations between holo-transcobalamin II, holo-haptocorrin, and total B-12 in serum samples from healthy subjects and patients. J Clin Pathol. 1993;46:537–9. doi: 10.1136/jcp.46.6.537. [DOI] [PMC free article] [PubMed] [Google Scholar]