Acceptability of three novel lipid-based nutrient supplements among Malawian infants and their caregivers

We tested the acceptability of three new lipid-based nutrient supplements (LNSs) in two independent phases among 18 8-12-month-old healthy rural Malawians and their caregivers. In phase 1, acceptability was assessed by offering three new LNSs in random order, and an LNS already determined to be acceptable, Nutributter ®, each added to 30g of warm maize porridge over three consecutive days. In phase 2, infants from each village were provided one of the new supplements for a 2-week home-use trial. Outcome measures included the amount consumed, time completion of the dose and the maternal rating of likeability on a 5-point scale. The supplements were rated acceptable if consumption was over 50% of the offered dose in phase 1. The mean (95% confidence interval) proportion of the LNS test meals consumed under direct observation was 88% (82-94%) for LNS-10gM, 90% (84-95%) for LNS-20gM, 87% (79-95%) for LNS-20gNoM, and 86% (83-90%) for Nutributter. The median (25th and 75th centile) time (minutes) for completing the offered test meal was 4 (2, 7) for LNS-10gM, 5 (3, 6) for LNS-20gM, 4 (3, 8) for LNS-20gNoM and 4 (2, 6) for Nutributter. During both phases, almost all caregivers rated all study foods very likeable for themselves and their children, with mean scores slightly lower among the caregivers than among the infants. In the home-use phase, the test foods were almost exclusively used by the study participants with minimal sharing with siblings and other household members. Some infants were reported to prefer the new investigational products over traditional complementary food. Considering that the novel LNS was largely acceptable. Efficacy trials are now needed to assess their impact on child growth and development. © 2011 Blackwell Publishing Ltd.

The study consisted of two phases. In phase 1, eligible participants consumed three directly observed test meals of investigational LNS in a random order on three consecutive days, followed on a fourth day by a similar meal containing Nutributter® (Nutriset S.A.S, Malaunay, France), a commercially available LNS known to be acceptable to infants (Adu‐Afarwuah et al. 2007). The primary outcome was the proportion of the test dose consumed in a 15‐min period; secondary outcomes comprised median time to completion of the test meal, maternal estimate of the supplement’s palatability to the child (on a hedonic 5‐point scale) and the caregiver’s own score of the supplements (with the same hedonic scale). In phase 2, a different group of participants were allocated to receive one of the three investigational LNS products for a 2‐week supplementation period, followed by individual structured interviews and focus group discussions. This phase was used to produce descriptive information on LNS consumption patterns among the study participants. The study was conducted in Mangochi district of Malawi, south‐eastern Africa. In this area, exclusive breastfeeding for a full 6 months is almost non‐existent, and infant diet is typically complemented with thin maize porridge (phala, 10% dry weight of maize flour) beginning at 2–6 months of age. Underweight and stunting are very common; in a recent prospective cohort study of 813 newborns, underweight (weight‐for‐age z‐score ≤2) and stunting (height‐for‐age z‐score ≤2) prevalence was 10% and 50% by 6 months of age and 40% and almost 80% by 18 months of age, respectively (Maleta et al. 2003). There is one rainy season between December and March during which the staple food maize is grown. Data collection for the trial was done during the harvest season, i.e. in April–May 2009. Inclusion criteria for both trial phases included age at least 8 but less than 12 months, current breastfeeding and consumption of complementary porridges, residence in the study area throughout the follow‐up period and signed informed consent from at least one caregiver. Exclusion criteria were moderate or severe wasting, presence of oedema, congenital malformations, history of peanut allergy or any severe allergic reaction, severe illness warranting hospitalization, concurrent participation in another clinical trial or any symptoms of food intolerance within 30 min after the ingestion of a 6‐g test dose of Nutributter (given to all potential participants to exclude the possibility of peanut allergy). For enrolment, trained health surveillance assistants from a local health centre contacted families who were known to live in one purposively selected village in the study area and to have a baby of the correct age. Caregivers, whose infant’s age was appropriate and verified from an under‐five clinic card, were invited for further screening to the health centre. At the enrolment session, caregivers were given detailed information on the trial contents, and the infants and children were fully assessed for eligibility. Those meeting eligibility criteria and whose caregivers signed a consent form were given a trial identification number and randomized into phase 1 of the study. For phase 2, six villages were randomly selected in the catchment area, and two villages were assigned to each of the three investigational products. Caregivers of age‐eligible infants and children were invited to the nearest health centre where, upon confirmation of all eligibility criteria, the infants and children were included in the study. The caregivers were the primary respondents. For phase 1, a statistician not involved in the implementation of the study pre‐made a randomization list that indicated the order of the daily test foods for each participant, and ensured that each investigational product was preceded and followed in the consumption order equally often by all other investigational products. The identification numbers and the corresponding sequence of various test meals were sealed in individual opaque envelopes. When a participant was enrolled, the caregiver chose one of the envelopes, and its content determined the choice of test meals for each day. In phase 2, group allocation was systematic and based on the village of residence of the participant. The supplements included three newly developed investigational LNSs and one earlier developed product (Nutributter) as a positive control. Of the investigational products, LNS‐20gM differed from Nutributter mainly in its EFA content, whereas LNS‐20gNoM contained less protein because it contained no milk powder, and LNS‐10gM was more concentrated in micronutrients (Table 1). For each supplement, the intended daily ration contained approximately 100% of the recommended daily allowance of several micronutrients. All supplements were produced by Nutriset S.A.S. LNS‐10gM, LNS‐20gM and LNS‐20gNoM were made of vegetable oil, dried skimmed milk powder (except LNS20gNoM), peanut paste, sugar, maltodextrines, and mineral and vitamin mix. Nutributter was formulated using the same ingredients plus whey. Nutrient content of the three investigational LNS‐products and Nutributter LNS, lipid‐based nutrient supplement; IU, international unit; RE, retinol equivalents. The products were packed in identical 140‐g opaque containers, each marked with one of eight different letter codes (A, B, C, D, E, F, G, H) corresponding to the different LNS composition (two different letter codes per product), in order to keep the research assistants masked to product formulation. For consumption, all LNS products were to be mixed in complementary porridge made from maize flour produced locally from milling whole grain maize. In phase 2, the flour was packed in opaque plastic bags, each containing 500 g of the maize flour. All test meals were consumed at the study office, where the caregivers reported on five consecutive mornings at approximately 8 am. The first day involved a run‐in procedure, during which the children ate a test meal (consisting of a randomly selected investigational product), and data were collected but were not included in the final analysis. The actual data collection happened on the following 4 days, with the three investigational products being offered in random order and Nutributter (positive control) on the last day. Because it was known that Nutributter was liked by children, and it is considerably sweeter than the test LNS products, it was always offered on the last day to avoid any possible interference with consumption of or hedonic responses to the test products. Research assistants prepared test meals from either 15 g of LNS‐20gM, 15 g of LNS‐20gNoM, 7.5 g of LNS‐10gM or 15 g of Nutributter, added to 30 g of freshly prepared, warm maize porridge. At each test meal, the mother was first asked to breastfeed the participating infant; thereafter, there was a 60‐min waiting period, during which no foods or liquids were offered to the infant. The porridge–LNS mixture was then given to the mother in a pre‐weighed bowl, and she was asked to feed the child with a standard spoon. The test feeding continued under direct observation of a research assistant until the participant completed the given portion or a maximum of 15 min had elapsed from the beginning of the meal. Food consumption was calculated by subtracting the final weight of the bowl, leftovers and utensils from their combined weight at the beginning of the meal. At the end of each test meal, the caregiver was asked to indicate how much she thought her child liked the food. The scoring was done on a 5‐point hedonic scale, graphically illustrated in a series of human face symbols with various degrees of smile or discontent. On the fifth test day, after the children had consumed their test meal of Nutributter, 20 mothers were given five 5‐g test doses of LNS products in a random order and were asked to rate their taste with the previously mentioned 5‐point hedonic scale. The five LNS products, given without any added porridge, consisted of the same four products given to the infants and an extra one that was being developed specifically for pregnant and lactating women (data not shown). In phase 2, caregivers were provided with a 2‐week supply of one of the three investigational LNS products. They were instructed to feed their infants twice daily with a portion that consisted of approximately two tablespoons of porridge mixed with either one teaspoon (5 g) of LNS‐10gM or two teaspoons (10 g) of LNS‐20gM or LNS‐20gNoM. After the 2‐week supplementation period, the caregivers were interviewed with structured questionnaires about the LNS consumption patterns in their household. During these individual interviews, the caregivers were also asked to rate the acceptability of the provided supplement by their infants and by themselves, using the same 5‐point hedonic scale that was used during the directly observed test meals. Because the amount consumed was to be established from phase 1, during which a scale and trained staff were available, in phase 2 of the study, the focus was on collecting self‐reported experiences at the end of the 2‐week supplementation period. To obtain more thorough insights into the consumption patterns, we organized six focus group discussions. At each session, approximately eight mothers from one village got together with a trained facilitator to discuss maternal perceptions on the taste, likeability, acceptability and feasibility of LNS products for child nutrition. Qualitative analysis was used to highlight themes that were related to caregivers’ opinions on the longer‐term use of the various investigational LNS products. For phase 1, the purpose was to determine whether each of the three investigational products had adequate acceptability defined as a minimum of 50% mean intake from an offered dose. The test dose, on the other hand, was set to represent 150% of the intended single portion in an eventual intervention where LNS would be given in two equal daily portions. The test dose was higher than the recommended portion for home use because we anticipated that measurement precision in the homes might not be perfect, and hence the children would occasionally be given larger portions. Consumption of a minimum of 50% of two LNS portions comparable with the test meal would mean that the participant obtained from the supplement at least 75% of the recommended daily allowance for the micronutrients included in it, which, together with normal feeding, was assumed to ensure appropriate micronutrient intake. The sample size was based on the desire to test the hypothesis that mean consumption of each type of LNS would be at least 50% of the amount offered. We assumed that the standard deviation of consumption would be 30% of the amount offered. The sample size of 18 would therefore allow us to reject the null hypothesis with 90% power if the true mean were at least 75%. For maternal rating of the products, a row balanced Latin square design of 20 mothers was included to even out order and carryover effects because they consumed the food on the same day, although only 18 mothers were required per sample size calculation. The order and carryover effect was not a concern for the children as they tested the products on different days, hence the sample size remained 18. For phase 2, the selected sample size was based on the plan to organize a focus group discussion in all the six villages where participants were enrolled. In an earlier research, it has been shown that about eight participants per focus group interview is a socially feasible number of participants (Barbour & Kitzinger 1999 ), which would dictate a minimum of 48 participants. Raw data were initially recorded on paper forms and their coherence was checked daily by a senior research assistant. The data were then double‐entered into a tailor‐made microsoft access 2003 (Microsoft Corporation, Redmond, WA, USA) database. The two entries were electronically compared; all extreme and otherwise susceptible values were confirmed or corrected. Statistical analysis was completed with stata for Windows (version 11.0) (StataCorp LP, College Station, TX, USA). For the main end point (amount in grams of test food consumed during the 15‐min test feeding session), which is a continuous variable, we calculated the mean [95% confidence interval (CI)] intake for each of the test foods separately, whereas for ordinal data, we calculated the median for each group. For qualitative data, individual interviews and focus group discussion sessions were transcribed and translated and back translated from Chiyao to English, and were analysed using Atlas.ti (version 6, Scientific Software Development GmbH, Berlin, Germany) for emerging themes. All children were analysed in the group to which they were initially randomized, i.e. on an intention to treat basis. The trial was performed according to International Conference on Harmonisation‐Good Clinical Practice guidelines, and it adhered to the principles of Helsinki declaration and regulatory guidelines in Malawi. Before the onset of enrolment, the trial protocol was reviewed and approved by the College of Medicine Research and Ethics Committee (University of Malawi) and the Ethical Committee of Pirkanmaa Hospital District (Finland). Key details of the protocol were published at the clinical trial registry of the National Library of Medicine, Bethesda, MD, USA (http://www.clinicaltrials.gov, trial identification is {“type”:”clinical-trial”,”attrs”:{“text”:”NCT00885144″,”term_id”:”NCT00885144″}}NCT00885144). All suspected serious adverse events (SAEs) were monitored by the research team without the use of an external Data Safety Monitoring Board. SAE was defined as any untoward medical occurrence that either resulted in death or was life‐threatening or required inpatient hospitalization or prolongation of existing hospitalization, or resulted in persistent or significant disability/incapacity or other serious medical condition.

N/A

• Dima Health