Iodine supplementation of lactating women and assessment of infant visual information processing and maternal and infant thyroid function: A randomized trial

Iodine deficiency is one of the major causes of brain damage in childhood. However, iodine supplementation during early pregnancy and lactation can prevent the ill effects of iodine deficiency. This study evaluated maternal and infant thyroid function and infant visual information processing (VIP) in the context of maternal iodine supplementation. A communitybased, randomized, supplementation trial was conducted. Mother infant dyads (n = 106) were enrolled within the first 10 days after delivery to participate in this study. Mothers were randomly assigned either to receive a potassium iodide capsule (225 μg iodine) daily for 26 weeks or iodized salt weekly for 26 weeks. Maternal thyroxine (T4), triiodothyronine (T3), thyroid stimulating hormone (TSH), thyroglobulin (Tg), urinary iodine concentration (UIC), breast milk iodine concentration (BMIC) and infant T4, TSH, UIC and VIP were measured as outcome variables. At baseline, neither mothers nor infants in the two groups were significantly different in any of the biomarkers or anthropometric measurements. Maternal TSH and goiter prevalence significantly decreased following iodine supplementation. The percentage of infants who preferentially remembered the familiar face was 26% in the capsule and 51% in the I-salt groups. Infant sex, length for age Z score, BMIC, maternal education and household food security were strong predictors of novelty quotient. In conclusion supplementation daily for six months with an iodine capsule or the use of appropriately iodized salt for an equivalent time was sufficient to reduce goiter and TSH in lactating women. Higher BMIC and LAZ as well as better household food security, maternal education, and male sex predicted higher novelty quotient scores in the VIP paradigm. The study was a community-based, randomized, supplementation trial conducted in 2013 in Sidama zone, southern Ethiopia. The study population depended on subsistence farming for their livelihood. Their major staple food was enset (Enset ventricosum) followed by unrefined maize [14]. The area was known for severe iodine deficiency prior to national implementation of salt iodization in 2012 [15–17]. Breast feeding behavior, household salt use, and compliance of subjects was reported in our previously published manuscript [12]. This study is registered with ClinicalTrials.gov, number {“type”:”clinical-trial”,”attrs”:{“text”:”NCT03889431″,”term_id”:”NCT03889431″}}NCT03889431. This study enrolled all lactating women in the participating communities who delivered between January and February 2013. A total of 106 healthy mother-infant dyads were included in the study and data collection was completed by August 30, 2013. Eligibility criteria were: the women must be lactating and must have delivered a full-term single infant within the prior 10 days. Mothers and infants must have no history of illness, thyroid dysfunction or iodine supplementation. All mother-infant dyads who met the criteria were invited to volunteer for the study, and enrollment was continued until the required sample size was attained. Ethical approval was obtained from Oklahoma State University (OSU), Hawassa University, the Ethiopian Ministry of Science and Technology, and the Food, Medicine, and Health Care Administration and Control Authority of Ethiopia. Following receipt of ethical clearance, study participants were given a detailed explanation of the objectives of the research and consent was obtained from the women for themselves and their infants before data collection was started. Women were enrolled with the help of health extension workers working in the study area. The 106 mothers with their infants were randomly assigned either to the iodine capsule group or to the iodized salt (I-salt) group using random numbers (Fig 1). (CONSORT flow diagram). The capsule group (n = 53) received 225 μg of iodine daily as a capsule of potassium iodide (Pure Encapsulations, Inc., Boston, MA) for 26 weeks. Iodine capsules were ingested daily by mothers under supervision of the health extension worker. The I-salt group was provided weekly with a 450 g packet of iodized salt (30 to 40 mg I as KIO3 /kg) for their household for 26 weeks (Guts Agro Industry PLC, Addis Ababa, Ethiopia) by the health extension workers. The 450 g of salt provided weekly for each family was planned to be enough that they would not purchase any additional (perhaps inappropriately iodized) salt from the market. There was no requirement to use all of the provided salt. A national survey conducted in 2005 with a multistage cluster sampling design estimated salt consumption per person per day using 24-hour dietary recalls. The national mean (±SD) for salt consumption was 8.4 (5.9) g/day. However, based on the 1,677 households sampled in the SNNPR region (our study region), salt consumption was 17.2 (13.8) g per person per day, or more than twice the national average. One reason for high salt consumption in the region is the custom for use of salt rather than sugar in coffee [18]. Prior to beginning supplementation, baseline data were collected from both mothers and infants. Collection and analysis of breast milk and urine has been described in our previous paper [12]. Maternal blood samples were collected using a disposable 10 cc syringe coated with lithium heparin with a 21 gauge needle (Sarstedt, Inc., Newton, N.C.). The blood was centrifuged and plasma was separated immediately. Plasma was frozen at– 20 0C and analyzed for concentrations of triiodothyronine (T3), thyroxine (T4), thyroid stimulating hormone (TSH) and thyroglobulin (Tg). Each woman was examined for goiter at baseline and end point by a single health officer using palpation based on the following grades: grade 0, no palpable or visible goiter; grade 1, palpable goiter but not visible when neck is in the normal position; grade 2, visible goiter when neck is in the normal position [4]. For mothers weight and height to calculate body mass index (BMI = Wtkg/Htm2), and mid upper arm circumference (MUAC) were measured. Infant anthropometry including head circumference, weight and length were measured. Infants’ weight-for-age (WAZ), length-for-age (LAZ), weight-for-length (WLZ), and head circumference-for-age (HCAZ) were calculated using WHO Anthro software (Version 2.0.4). Anthropometric measurements were done in duplicate. Mothers were interviewed at baseline using an individual questionnaire adapted from the Ethiopian Demographic and Health Survey. The questionnaire included socioeconomic status and demographic characteristics of the women [19]. Additionally, data were collected using the Household Food Insecurity Access Scale (HFIAS) developed by the Food and Nutrition Technical Assistance (FANTA) project of the United States Agency for International Development (USAID) [20]. After 26 weeks, collection of biological samples and anthropometric measurements were repeated. Additionally, at 26 weeks infant blood samples were collected on specimen paper by finger prick for TSH and T4 analysis, and the Visual Information Processing (VIP) test was administered to infants. All the participating mothers lactated throughout the study; thus no mother was excluded for having stopped lactation after enrollment. Testing of infant VIP was modified from an earlier report with Ethiopian infants [21] and from the work of Rose and colleagues [22]. For the VIP, eight pictures of young adult Ethiopian faces without emotional expression were used. One laptop computer controlled the presentation of stimuli. A second computer was used to follow and record the infant’s looking behavior. The infant’s looking behavior was coded live, but a camera in the second computer recorded the VIP test for later reliability testing. VIP measurements were always conducted by the same person. For the VIP test, the infant sat on the mother’s lap half a meter from a 15” laptop screen in a darkened room at the health center. The tester (who was blind to the study and out of sight of the infant) focused the camera in the second laptop on the infant’s eyes. The habituation phase began with the presentation of one face randomly selected from the eight faces. When the infant looked at a picture, its reflection was visible in the infant’s pupil and the experimenter pressed a key on his computer. Response latency (time between stimulus onset and the look) and the duration of the look were recorded by the computer software. When the infant looked away, the experimenter released the key; if the look away lasted 1s or longer, the face disappeared from the screen for 2s and then reappeared. This procedure continued until the criterion for habituation was achieved, namely two consecutive looks that were less than one half the mean of the two longest looks. After habituation, the comparison phase began without interruption. In the comparison phase, the face that appeared during habituation (the “familiar” face) appeared again on one side of the screen and a novel face of the opposite sex appeared on the other side of the screen. The faces remained on the screen regardless of whether or not the infant was looking at one of them. Once the infant accrued a total looking time at the faces of 5s, the faces disappeared for 1s, and then reappeared but on opposite sides of the screen. They remained until an additional 5s of looking had accrued. The infant’s look duration, shifts between faces, and looks away from the pictures were compiled simultaneously by the software. Visual recognition memory is composed of various constructs [22]. During habituation, look duration measures processing speed and attention [23] and average look duration and number of looks reflect the speed at which recognition was achieved [24]. For the comparison phase, the novelty quotient (NQ; preference for the new face) measures recognition memory (21) and total shifts identify how active the infant was in comparing the two faces. A university student, blind to the study, coded recorded videos for 30 randomly selected infants to test coding reliability. For longest look, the correlation was 0.78; correlation for novelty preference was low probably because distinguishing between infant’s looks at the right and left side of the screen was more difficult in the taped version than in person. The T3, T4, TSH and Tg were quantitatively determined by ELISA assay (ALPCO Diagnostics, Salem, NH). Quality control for T3 and T4 were obtained from Bio-Rad Laboratories (Irvine, Ca). Infant’s T4 and TSH also were determined by ELISA assay (Diagnostic Automation, Calabasas, CA). Reference ranges for thyroid hormone biomarkers of mothers and infants are included as supplied by the manufacturers. Hence, for mothers TSH: 0.4–4.2 μIU/mL, T3: 0.75–2.2 ng/mL, T4: 5–13 μg/dL and Tg: 2–50 ng/mL. Reference ranges for infants were: TSH: 0.7–34 μIU/mL and T4: 8–23 μg/dL. Percentages, frequency distributions, means (SD) and medians (IQR) were used to describe the socio-economic status, demographic characteristics, and iodine status of lactating mothers and their infants. All data were checked for normal distribution and skewed data were log-transformed before analysis. To compare the means between the two treatment groups independent-sample t-tests were used for normally distributed data and the Mann-Whitney U test for non-normally distributed data. For comparison of samples at baseline and 26 weeks, paired-sample t-tests were used for normally distributed data and Wilcoxon signed-Rank tests were used for non-normally distributed data. Pearson’s correlation coefficient (r) or Spearman’s rho were used to examine relations between variables. Variables predicting VIP were analyzed by multiple regression. All variables were checked for collinearity using variance inflation factors (VIF). All of the biomarkers analyzed and infants’ anthropometric indices with VIF < 4 were entered into the models. Multiple classification analysis (MCA) was used for predicting novelty quotient (NQ). MCA examines the contribution of each category of the predetermined predictor variables before and after adjustment for the control variables. MCA is effective to control multicollinearity [25]. Post hoc power calculation was made using STATA software for four outcomes (Tg, T3, T4 and TSH) and the power was adequate (>80%) for the Tg and TSH outcomes. However, it was lower than 80% for T3 and T4.