Impact of Daily Zinc Supplementation to Infants Born With Low Birth Weight on Death and Severe Disease

Objectives:

i.To determine the impact of daily administration of 1 RDA (5 mg) of elemental zinc to LBW infants from 4 weeks to 6 months of life on all-cause mortality.

ii.To determine the impact of zinc supplementation on severe morbidity requiring hospitalization.

Rationale:

i.The prevalence of LBW is high in Asia, particularly in South Asian countries like India, Pakistan and Bangladesh. LBW infants have a high neonatal and infant mortality and morbidity rates.

ii.A recent preliminary study showed that zinc supplementation of LBW infants reduced mortality (Sazawal et al., 2001).

iii.A targeted intervention in LBW infants starting at about 4 weeks of age is programmatically highly attractive. It can be combined with immunization program and with recent recommendation to give iron and folate to infants.

iv.The fact that mortality rate is very high in LBW infants means a relatively smaller study will give us results suitable for a policy decision.

Study population and location:

The infants were recruited at birth at the maternity ward of M.R. Bangur General Hospital where more than 6000 deliveries take place in a year. Located in Kolkata this hospital serves a district of West Bengal. Preliminary data collected by us from four hospitals in West Bengal showed that little over 37% of the babies born there have a birth weight of less than 2500 g. Taking conservative estimate of 35% of babies born to be of LBW (< 2500 g), more than 2000 LBW babies are expected to be born at this hospital in a year. LBW infants born at this hospital was recruited in this study and followed up at home. Mainly the urban and peri-urban poor use these Government hospitals. The hospital services are rendered free of charge by the Government.

Sample size:

One preliminary zinc supplementation trial in small-for-date infants reported as high as 67% reduction in the mortality rate with zinc supplementation between 2 and 9 months of age. We consider this effect as highly optimistic. Therefore, we calculated the required sample size based on more conservative estimates at various levels of expected reduction in mortality rate associated with zinc supplementation of LBW infants.

Taking a figure of infant mortality rate of 40/100 in LBW infants and assuming that at least 50% of these deaths occurred between 4 weeks and 6 months of age. We expect 20 deaths per 100 infants during the observation period for this study. Assuming a reduction in the mortality rate by 35% with 95% confidence and 80% power, we need a total sample size of 970 infants, which includes 10% excess to account for deviated course (e.g. Loss to follow-up, withdrawal of consent). Assuming a mortality rate reduction of 30% with similar assumption the total (plus 10% for deviated courses) sample size should be 1360 in LBW infants. If we expect a reduction in mortality rate by 25%, the total sample size would be 1900 infants which include 10% for deviated course.

Inclusion criteria:

1. Weight at birth less than 2500 g
2. Infant is likely to stay in the city for the next 6 months
3. Absence of gross congenital malformation

Exclusion criteria:

1. Perinatal illness requiring hospital stay after birth for more than 7 days
2. Written informed consent was not provided by parents
3. Presence of gross congenital anomaly
4. Resides at a distance that makes it difficult to reach for home visit (estimated time to reach in more than 3 hours)

Randomization and blinding:

The inclusion criteria were ascertained and those eligible were considered for inclusion into the study. The child visited the clinic or a home visit was made at 3-4 weeks by a team of health workers and if still alive the infant was randomly assigned to the zinc or to the placebo group. A master randomization code was prepared by using permuted blocks of random numbers using block lengths of 4, 6 and 8. This code was held by a person not involved with the study who numbered the bottles containing the formulation or placebo for each child according to the randomization chart and the bottles were labeled only with the serial numbers of the child in the study. The final code was kept with a person not involved with the study. The zinc syrup and the placebo were similar in appearance, consistency and taste.

Intervention:

Each infant in the study received placebo or 5 mg of elemental zinc (1 RDA) as zinc acetate syrup daily from 4 weeks of age. For routine immunization and other care, parents were advised to attend the baby clinic of the same hospital or another hospital in the locality. In case of any reported or detected illness by the field workers, free drugs were provided and hospitalization facilitated when required.

Enrollment of subjects:

We estimate that 80% of the LBW babies can be realistically included in the study. We have earlier estimated that approximately a little over 2000 LBW babies were born in a year in each hospital. We recruited a total of 2012 infants, 1005 in the zinc group and 1007 in the placebo group.

After obtaining an informed verbal consent, the infant eligible for enrollment at the maternity ward was escorted home by a health worker to note the exact location of their stay. The mother was advised to bring the infant to the hospital at 4 weeks where baseline information on the infant were recorded and 8 weeks' supply of the zinc/placebo formulation was provided to the parents with instruction for use. The mother, under supervision, gave the first dose at the clinic. If the mother did not return at 4 weeks after birth, field workers were visited home and the medicine was provided as above. The field workers visited the home 8 weeks later recorded the events about the infant and provided another 8 weeks to medicine. After 16 weeks of medication, the infant visited the clinic or a field visit was made to ascertain the state of the infant and was provided with 8 weeks of medication and evaluation was recorded, after 8 weeks.

Measurement of study outcome:

Death if any was recorded at each contact. The cause of death was ascertained by a verbal autopsy. Available hospital records were examined and a team of pediatricians independently reviewed verbal autopsy questionnaires. The diagnosis was made by consensus or through discussion. Similarly, illness events requiring hospitalization were documented at each contact. A questionnaire was used and hospital records examined, including all prescriptions. At each contact either at the clinic or at home, the details of illness events and treatment received were recorded in a standardized data forms and documentations such as prescriptions, medicine bottles or cartons were physically checked and recorded. Anthropometric measurements were recorded at 3 and 6 months of receiving the medicines. All documentations on morbidity were evaluated at fortnightly meetings of the health workers with the two pediatricians involved the study (DM \& AS) and the cause's of hospital admissions, deaths and morbidities were reviewed and a diagnosis was made \& recorded

Ethical considerations:

Written informed consent was obtained from the parents. Ethics Review Committee of the Society for Applied Studies, Kolkata approved the protocol.

Data management and statistical analysis:

The data were entered into a desktop computer and edited using a software EPI Info (CDC, Atlanta, USA and WHO, Geneva) and analyzed by EPI Info and another software named Stata (Stata Corporation, Texas, USA). The main comparison was between those who received zinc and those who received placebo. Baseline information was compared to evaluate the comparability of the two groups. These include socio-economic variables, type of delivery, gestation, birth weight, breast-feeding status etc. The two groups were compared for all-cause mortality. The proportion of infants having illness requiring hospitalization in the two groups was compared. For this comparison, only the first episode requiring hospitalization in a particular infant was considered. Although the follow-up period would be very similar, some degree of variable follow-up was expected which was addressed by using person-time calculation between 4 weeks and end of the study. The analysis was based on intention-to-treat.

Results:

About 20,248 neonates were evaluated for eligibility at the maternity ward of MR Bangur Hospital of whom 5,942 had <2,500g weight. A total of 2,012 infants were admitted into the study, 1,005 being in the zinc group and 1,007 in the placebo group. A total of 909 infants in the zinc group and 913 in the placebo group completed 6 months of follow-up. The birth weight, type of delivery, birth order and mothers education were comparable in the two groups. Weight, length, head and midarm circumference at 3 and 6 months were similar in the two groups. In a subgroup of infants' blood hemoglobin levels were determined on completion of the study and the mean values in the two groups were similar. In a subgroup of infants zinc supplemented group should a substantially higher mean zinc concentration compared to placebo. The numbers of deaths in the two groups were not significantly different. Hospitalization rates with diarrhea, ARI and all causes are similar in the two groups. The number of infants with one or more episodes of diarrhea was significantly less among zinc supplemented group (by 17%, 95% CI:1 to 35 %). The number of infants with one or more episodes of acute respiratory infection was similar in two groups (p=0.76).