Effect of Antenatal Zinc Supplementation on Impetigo in Infants in Bangladesh

Abstract

We sought to determine the effects of maternal zinc supplementation on skin infections among infants in poor urban areas of Dhaka, Bangladesh. A double-blind, placebo-controlled, randomized trial was conducted among 199 and 221 Bangladeshi infants whose mothers were administered 30 mg daily of zinc or placebo, respectively. The mothers received zinc supplementation from 12–16 weeks’ gestation until delivery, and the infants were followed up until 6 months of age. Among the infants of mothers who received zinc supplementation during the antenatal period, 10.6% had at least 1 episode of impetigo compared with 19.6% of the infants of mothers in the placebo group (P = 0.01). Infants in the treatment group had a 54% reduction in incidence rate of episodes of impetigo when compared with infants in the placebo group (P = 0.01). The effect of zinc supplementation was more pronounced among male infants (64% reduction) and intrauterine growth restricted and low birth weight infants (73% reduction) and among infants of mothers with increased parity (60% reduction) or decreased socioeconomic status (71% reduction).

Zinc is critical to many processes involved in fetal and childhood growth, health, and survival,^1,2 and low maternal blood zinc concentrations during pregnancy are linked with increased risk of low birth weight (LBW) (,2500 g).³ Zinc supplementation may reduce risk of preterm birth and improve childhood growth rates and immune responses to infectious diseases.⁴ Maternal zinc supplementation has been shown to reduce risks of acute diarrhea, dysentery, and impetigo, especially in young children with growth retardation.^4–6

Zinc is needed for normal keratinization and healing of skin. It has been suggested that zinc may play an important role in maintaining epidermal barrier function and protecting the skin from infection.⁷ Here, we report effects of maternal zinc supplementation on skin infections in urban poor infants in Dhaka, Bangladesh.

METHODS

In all, 559 pregnant women were enrolled between 12 and 16 weeks’ gestation. They were stratified by parity and randomly assigned to receive 30 mg of elemental zinc or a cellulose placebo tablet daily, as described previously.^5,8

A total of 198 and 220 infants from mothers in the zinc and placebo groups, respectively, were included in postnatal surveillance; 2 infants had unrecorded treatment group status and were excluded from analysis. Following birth, 6 monthly home visits were conducted by trained nonmedical interviewers who asked mothers questions specific for respiratory infections, diarrhea, fever, and skin infections during the past week.^5,8

To diagnose impetigo, mothers were shown 7 clinical color photographs of impetigo of varying degrees of severity and points in its natural history, and asked whether their infant had skin lesions resembling those in the pictures during the past week. Questions were asked regarding location and characteristics of skin lesions (ie, presence of papules, blisters, drainage, crust, erosion, and tenderness around the lesions) and constitutional signs and symptoms (swollen axillary or inguinal lymph nodes, fever, irritability, scarlatiniform rash). The main outcome measure was primary impetigo; secondarily infected lesions were excluded. Impact of maternal zinc supplementation on incidence of infant impetigo was a secondary study outcome,⁵ and thus, analyses were exploratory.

Number of episodes of impetigo was determined, ranging from 0 to 4, and 2 outcome variables were derived; one for whether the infant ever had impetigo overall and another for the number of times the infant had impetigo. Using STATA 9.2, multiple forward and backward stepwise logistic regression was performed, including the following variables: treatment (zinc vs. placebo), number of episodes of acute lower respiratory infection (ALRI), number of episodes of acute diarrhea, infant’s birth weight (g), infant’s sex, maternal age and parity, and socioeconomic status. These variables were regressed on the dependent variable of ever having impetigo. Both forward and backward stepwise regression yielded the same, noncollinear variable combination— treatment, infant’s weight, and number of episodes of ALRI—as did regression analysis using separate birth weight categories (<2500 g, >2500 g), parity categories (nulliparous, primiparous, multiparous), and asset categories (very poor, poor, middle income). Odds ratios for ever having impetigo were then determined for the following variables, confirmed by goodness-of-fit analysis: treatment, number of episodes of ALRI, infant’s birth weight, and maternal assets.⁹ Logistic regression was used to examine the effect of treatment on the outcome of ever having impetigo, stratifying based on birth weight (using first the categories LBW [<2500 g] and normal birth weight [NBW] [≤2500 g]), and then looking at categories of NBW, intrauterine growth restricted (IUGR)/LBW,⁸ and preterm). The same stratification was done for sex of the infant, parity, and maternal assets.

To examine differences between treatment groups in frequencies of impetigo among infants, a Poisson regression model was used with the total number of episodes (0 to 4) as the dependent variable, and with treatment, number of episodes of ALRI, infant’s birth weight, and maternal assets as independent variables. Additional analysis of the data was again done using stratification by birth weight, infant sex, maternal assets, and parity to examine independent variable category’s effects on incidence rate ratios of episodes of impetigo between the zinc and placebo groups, adjusted for assets, birth weight, and episodes of ALRI, except when stratifying by that particular variable. P values <0.05 were considered statistically significant. The study was approved by the Ethical Review Committee of ICDRR,B and the Committee on Human Research of the Johns Hopkins Bloomberg School of Public Health.

RESULTS

Baseline characteristics of the mothers and their infants were comparable across treatment groups (Table, Supplemental Digital Content 1, https://links.lww.com/INF/B30).

Of the infants in the zinc group, 10.6% had at least 1 episode of impetigo compared with 19.6% infants in the placebo group (P = 0.01) (Table 1), a 54% [95% confidence interval 18%–77%] adjusted reduction (P = 0.01) in odds of ever having impetigo (Table, Supplemental Digital Content 2, https://links.lww.com/INF/B31). For each increased episode of ALRI, there was a 1.5 times increased odds of ever having impetigo (P = 0.004) (Table, Supplemental Digital Content 2, https://links.lww.com/INF/B31).

TABLE 1: Episodes of Impetigo in Infants in Maternal Placebo and Zinc Supplementation Groups

Among LBW infants, odds of ever having impetigo were reduced 73% (8%–91%) in the treatment group (P5 0.04) (Table, Supplemental Digital Content 3, https://links.lww.com/INF/B32). Among NBW infants, there also was a 54% (3%–78%) reduction in odds of having impetigo (P = 0.04); there was no impact on odds of having impetigo (P = 0.83) in preterm infants. There was a 64% (26%–89%) reduction in odds ratio of ever having impetigo among boys in the treatment group (P = 0.01); no reduction was found for girls. A 71% (18%–76%) reduction in odds of ever having impetigo was observed in the very poor group (P = 0.01) receiving zinc treatment. Among treated multiparous women, there was a 60% (8%–83%) reduction in odds of their infants ever having impetigo (P = 0.03).

There was a 47% (17%–66%) adjusted reduction in incidence rate of impetigo in treated infants (P = 0.01) (Table, Supplemental Digital Content 4, https://links.lww.com/INF/B33), and a 36% (14%–52%) reduction in incidence rate of episodes of impetigo between individuals with more assets compared with those with less (P = 0.003) was observed. In addition, there was a 1.4 times greater incidence rate of episodes of impetigo for each increased episode of ALRI (P = 0.001).

Comparing treatment and placebo groups, infants with IUGR-LBW (72% reduction [25%–90%], P = 0.01) and LBW (63% reduction [11%–83%], P = 0.01) had statistically significant reductions in incidence rate of impetigo (Table, Supplemental Digital Content 5, https://links.lww.com/INF/B34). Among NBW and premature infants, there was no statistically significant difference in incidence of impetigo. There was a 49% (8%–72%) decrease in incidence rate for boys (P = 0.03), but no effect for girls. In very poor women only, there was a 62% (24%–81%) reduction in incidence between the treatment and placebo groups (P = 0.01). Among treated primiparous women, there was a statistically significant (59% [11%–81%]) reduction of incidence rate (P = 0.03).

DISCUSSION

Maternal zinc supplementation during the last 2 trimesters of pregnancy reduced both the incidence of impetigo and the odds of ever getting impetigo in infants during their first 6 months of life, especially among LBW male offspring of poor mothers. The biologic mechanism for zinc’s efficacy on infant impetigo may include strengthening of the epidermal barrier, especially during a period of rapid cell development and growth, or overall improved immunity to infection.

Increased episodes of ALRI seemed to increase the odds of impetigo, but this requires further research. The infants may have an impaired immune system to begin with, particularly if they are repeatedly contracting ALRI; perhaps in this context, respiratory infection leads to seeding of skin sites with infectious agents that subsequently cause skin infection, or vice versa.

Stratification showed that odds of impetigo were reduced in boys, but not in girls. The mechanism for this is not clear but may suggest that boys are more amenable to an improved zinc status in utero after maternal supplementation, perhaps interfacing with underlying biologic mechanisms that give newborn girls a survival advantage over boys.¹⁰ Zinc treatment seems to have more benefits among offspring of women of lower socioeconomic status, who are likely to be the most micronutrient deficient, compared with those belonging to higher income groups.

Poisson regression also showed that maternal assets, episodes of ALRI, treatment status, and birth weight were all associated with incidence rate of impetigo among infants. However, only infants falling in either the IUGR-LBW or LBW category had statistically significant reductions in incidence rate of impetigo. This differs from the logistic model where NBW infants also showed a significant reduction. This may suggest that LBW infants have a greater overall incidence of impetigo that may be reduced through maternal zinc supplementation, but that the overall effect of zinc on reducing the odds of ever having impetigo applies to both normal and underweight infants.

A potential limitation to this study is that pregnant women sometimes refused to look at pictures of infection because they felt that it might curse their baby. This perceived threat may have created recall bias in postnatal reporting. Data quality may be improved in future studies through culturally sensitive communications that alleviate maternal fears of jinxing the baby.

This study supports the recent conclusion that in view of possible benefits of maternal zinc supplementation on reducing risk of premature delivery and improving infant birth weight among undernourished women, and the lack of reported adverse effects, zinc should be included in maternal antenatal supplements in populations at risk for zinc deficiency.¹⁹ Future studies should include assessment of impact of maternal zinc supplementation on culture-proven impetigo, including in mothers, and maternal-infant transmission of impetigo-causing strains of pathogens.