Zinc status in HIV infected Ugandan children aged 1-5 years: a cross sectional baseline survey

Zinc deficiency is wide spread in low-income countries and is responsible for 4% of childhood deaths and 1% of the burden of disease in Africa, Latin America and Asia [1]. Populations in sub-Saharan Africa and South East Asia have the greatest risk of zinc deficiency because of inadequate zinc intake in about one third of the population [2].

Zinc is a component of various metallo-enzymes, proteins and cell membranes; and plays an important role in immune regulation. Zinc deficiency increases susceptibility to oxidative stress and impaired cell membrane function [3, 4]. In HIV infected adults, low serum zinc has been associated with more advanced HIV disease and increased mortality [5–7].

Whereas some studies of HIV infected children in high-income countries have indicated that micronutrient deficiencies are uncommon [8], the reverse is true in low-income countries. The zinc status of HIV infected children in Uganda has not been reported. In this paper we report the magnitude of zinc deficiency and associated factors in a group of HIV infected children in Uganda.

Of the 247 children analysed for serum zinc, 134 (54.3%) had low zinc < 10μmol/L. Their mean (SD) age was 33.4 (13.8) with a range of 12.0 to 65.5 months. The male to female ratio was approximately 1:1.

Serum zinc and factors associated with low zinc concentrations

The mean (SD) serum zinc concentration was 10.0 (2.9) μmol/L with a range of 5.6 - 29.5 μmol/L. There was no linear relationship between age, WHZ and haemoglobin versus zinc status as shown in Figure 2, 3 and 4. In addition there was no significant difference in mean age, anthropometry, haemoglobin and CD4+ cell count among children with low serum zinc (< 10μmol/L) compared to those with normal zinc levels (table 2).

Variable (mean, SD)	Serum zinc < 10 μmol/L	95% CI	Serum zinc = 10 μmol/L	95% CI
Age in months	32.2 (14.0)	29.8 to 34.6	34.7 (13.6)	32.2 to 37.2
Weight for height z score	-0.2 (1.4)	-0.5 to 0.0	0.1 (1.5)	-0.2 to 0.4
Weight for age z score	-1.4 (1.3)	-1.6 to -1.2	-1.0 (-1.3)	-1.2 to -0.7
Height for age z score	-2.3 (1.4)	-2.5 to -2.0	-1.9 (1.8)	-2.3 to -1.6
Haemoglobin (g/dl)	9.7 (1.7)	9.5 to 10.0	10.2 (1.5)	9.7 to 10.5)
Absolute CD4+ cell count (cells/μ L)	1056 (613)	950 to 1162	1214 (610)	1099 to 1330
*On HAART (n,%)	13 (29.5)		31 (70.5)

*Children on HAART are expressed as a percentage (%).

Factors associated with low serum zinc concentrations at bivariate analysis were: being HAART naïve, reported fever, underweight, WHO stage 3 or 4 disease and elevated CRP (table 3). The mean (SD) serum zinc among CRP negative children was 10.2 (2.5) compared to 9.8 (3.6) in CRP positive children. This difference was not statistically significant.

Variable	No. of children	Low zinc < 10μmol/L (n; %)	Unadjusted OR (95%CI)	p-value	Adjusted OR (95%CI)
Age < 36 months Age ≥ 36 months	146 101	86 (58.2) 49 (48.5)	1.5 (0.9 - 2.5)	0.153	1.0 (0.6 - 1.8)
Male Female	128 119	75 (58.6) 59 (49.6)	1.4 (0.9 - 2.4)	0.162	0.7 (0.4 - 1.1)
On HAART No HAART	44 203	13 (29.5) 121(59.6)	3.5 (1.7 - 7.1)	0.000	3.7 (1.8 - 7.7)
Reported fever No fever	44 203	30 (68.2) 104 (51.2)	2.0 (1.0 - 4.1)	0.046	2.2 (1.1 - 4.6)
Current diarrhoea No Diarrhoea	20 227	15 (75.0) 119 (52.4)	2.7 (1.0 - 7.7)	0.062	2.2 (0.8 - 6.7)
Cough present No cough	141 106	81 (57.4) 53 (50.0)	1.4 (0.8 - 2.2)	0.249
WHO stage 3 or 4 WHO stage 1 or 2	62 185	42 (67.7) 92 (49.7)	2.1 (1.2 - 3.9)	0.018	1.5 (0.8 - 2.9)
WHZ score < -2 WHZ score ≥ -2	24 223	15 (62.5) 119 (53.4)	1.5 (0.6 - 3.5)	0.519
WAZ score < -2 WAZ score ≥ -2	56 191	38 (67.9) 96 (50.3)	2.1 (1.1 - 3.9)	0.022	1.2 (0.6 - 2.6)
HAZ score < -2 HAZ score ≥ -2	136 111	80 (58.8) 54 (48.6)	1.5 (0.9 - 2.5)	0.124	1.3 (0.8 - 2.3)
CRP < 6 mg/L CRP ≥ 6 mg/L	147 100	72 (49.0) 62 (62.0)	1.7 (1.0 - 2.9)	0.051	1.6 (0.9 - 2.7)
Hb < 11 g/dl Hb ≥ 11 g/dl	170 77	98 (57.6) 36 (46.8)	1.6 (0.9 - 2.7)	0.130	1.1 (0.6 - 2.0)
CD4 < 20% CD4 ≥ 20%	85 162	52 (61.2) 82 (50.6)	1.5 (0.9 - 2.5)	0.139	1.1 (0.6 - 1.9)

In the multivariate model the only significant independent predictors of low zinc were: being HAART naïve and reported fever. Of the 44 children on HAART, 13 (29.5%) had low serum zinc compared to 121 of 203 (59.6%) in the non-HAART group. This difference was statistically significant [Adjusted OR 3.7 (95% CI 1.8- 7.7)]. When fever was excluded from the model elevated CRP was a significant predictor of low zinc, illustrating that fever and CRP are closely related.

The prevalence of zinc deficiency in this group of HIV infected children aged 1-5 years attending paediatric HIV clinics in Uganda was high (54.3%). This was higher than what was reported in a previous study of children aged 6-36 months with persistent diarrhoea but of undetermined HIV status at Mulago hospital, Uganda [15]. While we used a cut off of 10 μmol/L to indicate zinc deficiency, the previous study had used 4.73 μmol/L assuming a normal serum zinc level of 8.99 (SD 2.13) μmol/L measured from a control group of healthy children. What was regarded as normal is below the mean (SD) serum zinc of 10.0 (2.9) reported in our study. An earlier community study in Kampala had found a serum zinc range of 8.4 - 20.9μmol/L [mean(SD) 10.1(3.2)] among children aged 4 -14 years[16]. Based on these two previous studies, it is possible that the zinc status in our study is similar to zinc levels in the general population of children in a similar age group in Uganda. The prevalence was twice as high in children who had not yet started HAART compared to those receiving HAART. This was not influenced by the duration on HAART. This implies that HAART may protect against zinc deficiency but will not completely eliminate it.

Compared to other studies of HIV infected children in Africa, the prevalence of zinc deficiency was higher than what was reported in South Africa and Rwanda [17, 18]. Our study had a larger sample size and older children compared to the two studies which enrolled children from 2 months of age.

We did not find any association between age and zinc status in our study. All the children enrolled in this study were above one year of age, very few of whom were still breastfeeding. Other studies have shown that micronutrient deficiencies are less likely to occur below 24 months of age [17] and that breastfeeding is protective [19] against low serum zinc. Further more, there was no significant difference in mean zinc concentrations between girls and boys. Similar findings were reported from children whose HIV status was not known in low-income families in California [20].

Surprisingly, there was no significant association between zinc status and diarrhoea in the current study, possibly because there were very few children with diarrhoea. Studies in other developing countries have shown an association between low zinc and diarrhoea [21]. As expected, the presence of fever was significantly associated with low zinc levels. Fever is an indicator of infection and this may be associated with increased acute phase proteins and consequently reduced serum zinc [22]. Other common illnesses such as cough and skin rash were not associated with low zinc either.

There was a weak association between underweight and zinc status although a previous Uganda study of children with persistent diarrhoea [15] and a South Africa study of HIV infected children [17] did not find any association between zinc and nutritional status. Some authors have shown that severely malnourished children have a higher anti-oxidant activity and this is associated with low serum zinc [19]. Although anaemic children were more likely to have low zinc, the association was not significant. Other studies in zinc deficient low-income countries such as India and Cambodia have reported a close association between zinc deficiency and anaemia [19, 23]. The lack of association in our study could not be explained.

Advanced HIV disease (WHO stage 3 and 4) was associated with low serum zinc but probably confounded by other factors. Advanced HIV disease is more likely to be associated with recurrent acute infections and an elevated acute phase response interpreted as low zinc. Our findings are similar to a South African study where the prevalence of zinc deficiency increased with HIV disease staging [6]. CD4+ T cell count had no association with zinc status in the current study. This finding is similar to what was reported in studies of HIV infected adults at Tufts university and Medical Centre in the nutrition for health living cohort [24] and in South Africa [6].

Children who were on HAART were less likely to be zinc deficient. This is further supported by a study in adults that showed that patients receiving HAART have better micronutrient indices including zinc than those not yet on HAART [25]. However, two other studies indicated that zinc deficiency remains highly prevalent in HIV infected adults on HAART [7, 24].

We acknowledge that infant, child feeding practices and maternal HIV status influence the diet and mode of feeding [26], and may subsequently affect zinc intake. Adult studies have shown that the time of blood collection and feeding influence zinc levels[27], factors not controlled for in this study. The exclusions though numerous, were not systematic and therefore less likely to influence the results of our study. Low zinc levels were associated with fever or CRP implying that infection and acute phase proteins may have influenced the zinc status of the study children. Other researchers have previously confirmed that serum zinc is affected by the serum protein level and any acute-phase reactions[28]. Although the concentration of serum zinc gives limited information on the total zinc content in the body, there is no better way of determining zinc status that has been established[29]. The generalisability of this study is limited to the HIV infected children since we did not have a control group of HIV un-infected children.

Coupled with the already existing poor nutritional and immunological status, low zinc status in Ugandan HIV infected children is likely to remain a significant contributor to increased morbidity especially among those not yet receiving HAART.

While almost two thirds of untreated HIV infected children were zinc deficient, zinc deficiency occurred in only a third of those on HAART. Increased access to HAART among HIV infected children living in Uganda might reduce the prevalence of zinc deficiency in this population.