Infant feeding and health-related quality of life in healthy Chinese infants: results from a prospective, observational cohort study

Breast milk is considered the best source of nutrition for infants [1], but for a variety of reasons many parents choose to supplement breastfeeding with some formula-feeding or to provide feedings exclusively with infant formula. Optimizing health outcomes of formula-fed infants is therefore a public health priority [2]. Infant health-related quality of life (HRQOL), a broad concept that encompasses aspects of physical, psychological, and social function [3], is an important outcome that may be directly affected by an infant’s feeding regimen in at least two ways. First, the infant’s feeding regimen may influence feeding tolerance and gastrointestinal (GI) symptoms. For example, compared with breastfed infants, infants who are fed standard formula have harder stools [4, 5], which can lead to constipation and discomfort. Second, feeding regimen may influence immune function and susceptibility to illnesses or infections [6]. For example, breast milk has been shown to support immunity through transmission of maternal immune agents and by promoting a unique balance of microflora in the gut [7] that may offer immune benefits [8]. However, despite a need to better understand infant HRQOL in relation to feeding regimen, there is little research on this association to date [9].

Infant formula containing structured lipid (i.e., fat with an increased proportion of palmitic acid in the sn-2 position) and oligofructose (a non-digestible soluble dietary fiber) has been developed to promote the formation of softer stools that are more like those of breastfed infants [10, 11]. In addition, oligofructose is a prebiotic [12] that has been shown to influence the growth of bifidobacteria to levels that are within the range of levels found in breastfed infants [13]. The aim of this study was to assess HRQOL in infants consuming an infant formula with increased sn-2 palmitate and oligofructose (study formula) either alone or as a supplement to breastfeeding, and to compare this to HRQOL in infants consuming breast milk exclusively. Although there are methodologic difficulties of measuring HRQOL in infants [14], several valid and reliable instruments exist to examine infant HRQOL through proxy report by the parent or caregiver [9, 15‐17]. We used the Infant and Toddler Quality of Life Questionnaire (ITQOL) [17] to assess HRQOL as a function of feeding regimen in a large prospective study of Chinese infants. We hypothesized that infants fed the study formula either alone or as a supplement to breast milk would have the same HRQOL as infants fed breast milk alone.

The study included 136 exclusively breastfed, 140 exclusively study formula-fed, and 151 mixed-fed infants (Fig. 1). Feeding regimen groups were similar with regard to age at study enrollment, sex, and ethnicity (Table 1). Exclusively study formula-fed infants were born at a slightly earlier gestational age, on average, than exclusively breastfed or mixed-fed infants. All study infants were Asian and none attended daycare during the study. Infants in the mixed-fed group had a generally stable feeding profile over the study interval (percentage of daily feedings as breastfeeds = 58.3 % ± 20.1 % at study day 1 and 58.1 % ± 19.8 % at day 48). In regard to parental characteristics, mothers and fathers of study formula-fed infants had slightly but significantly fewer years of education compared to the other groups, with no differences in occupational status (Table 1). There was no difference (p = 0.48) in monthly household income among the three groups. The average number of adults living in the household was high and did not differ among groups (3.1 ± 1, p = 0.96). Similarly, the proportion of infants delivered by Cesarean section was high (~55 %) and not different across groups (p = 0.21).

Parent SF-12v2 summary scores, which were assessed as a covariate in this study, increased from day 1 to day 48 and differed only slightly across feeding regimen groups (Table 2). The MCS score was shown to be moderately positively correlated with all ITQOL parent-focused concepts at day 1 (all r ≥ 0.35, p < 0.05, Appendix) and at day 48 (all r ≥ 0.37, p < 0.05, Appendix).

Comparisons of ITQOL concept scores across feeding regimen groups revealed no significant differences at day 1 (Table 3). On day 48, the study formula-fed group was found to be significantly lower than both the breastfed group and the mixed-fed groups on the Temperament and Moods score, the General Health Perceptions score, and the Parent Impact—Time score (Table 3). Study formula-fed infants were also found to be significantly lower than breastfed infants (but not mixed-fed infants) on the Parent Impact—Emotional score at day 48 (Table 3). These differences ranged from 3 to 5 points on scores standardized to range from 0 to 100.

In this study, we found high HRQOL scores at both study visits, regardless of feeding regimen. A few small differences were observed at day 48 between study formula-fed and exclusively breastfed infants on two infant-focused HRQOL concepts (Temperament and Moods, General Health Perceptions) and two parent-focused HRQOL concepts (Parent Impact—Time and Parent Impact—Emotional). Similar small differences were observed at day 48 between the study formula-fed and mixed-fed groups for Temperament and Moods, General Health Perceptions, and Parent Impact—Time. All of these differences, while statistically significant, were quite modest – ranging from 3 to 5 points on a 100-point scale. Prior studies consistently suggest that changes in HRQOL of a half a standard deviation or greater represent meaningful changes in people’s experiences; this appears to hold across a variety of measures [26]. We observed ITQOL concept score differences of between a third and a half a standard deviation between feeding regimen groups, differences that are below this clinically meaningful threshold. Although these differences may not be clinically relevant, it is possible that other unmeasured factors such as maternal stress, postpartum depression, sleep quality, or other social determinants of health in the study formula-fed group may have contributed to both the parents’ initial choice of feeding regimens and the slightly lower Parent Impact—Emotional and Parent Impact—Time scores in the study formula-fed group. Likewise, Temperament and Moods may have been lower in the formula-fed group due to the parent’s rationale for initiating formula-feedings or other infant health measures, which this study was not designed to assess. The observed differences in HRQOL may also be due to differences in infant feeding mode (bottle vs. breastfeeding) rather than nutrition (formula or breast milk) or other unmeasured factors.

In general, HRQOL scores for our study population were high, indicating good overall health and well-being in this sample. For comparison, a study in infants 13 months of age with severe asthma-like symptoms reported mean ITQOL concept scores of 61 for Overall Health, 70 for Temperament and Moods, and 66 for Bodily Pain [27]. In addition, the mean SF-12v2 scores of the parents of infants in the present study were close to the expected values for a general adult population (scores are norm-based on the US general population with a mean score of 50 and a standard deviation of ten; scores in Chinese adults have been shown to be very similar [23, 28]). A variety of health-promoting factors contribute to infant well-being, and although decisions about infant feeding are important, their impact may be attenuated in the presence of these other factors. For example, social factors may influence infant well-being (real and/or perceived), and the mean number of adults in the household of infants in this study (~3) suggests a high caregiver-to-infant ratio which may have contributed to high HRQOL scores. Less healthy infants or those in less favorable social conditions may be more sensitive than our study population to the potential effects of feeding regimens on HRQOL. Hence, future research should evaluate the relationship between feeding regimens and HRQOL in other infant populations, especially those with lower HRQOL.

Self-assessed HRQOL is impossible for infants due to their young age. Therefore infant HRQOL can be assessed only by proxy, typically by parental or caregiver report. However the quality of life of the proxy may influence their perception of the infant’s quality of life. In order to address this potential bias, we included maternal quality of life (assessed using SF-12v2 MCS score) as a covariate in the analysis of ITQOL scores across feeding groups. This approach may be broadly applicable to other studies of infant HRQOL, and the use of a single global summary score such as the MCS may be preferred over assessing multiple individual factors (e.g. family income, occupation, and educational level) that may influence parental rating of infant HRQOL.

Our study adds to a limited body of literature assessing the impact of feeding regimen on quality of life in infants. Most published studies have assessed the impact of breastfeeding on parental quality of life and have not evaluated effects on infant quality of life [29‐31]. Manificat and Dazord [32] examined quality of life among 105 infants (mean age 5.7 months) as a function of feeding, using the Quality of Life of the Infant Scale [9] and reported that maternal overall perception of infant quality of life was independent of the mode of feeding. However, duration of breastfeeding greater than 3 months was associated with higher maternal-reported scores on four items, including infant nervousness, infant anxiousness, family cohesiveness, and infant total quality of life (visual analogue scale) [32]. This finding of differences on some, but not all, items is similar to our results.

Strengths of our study include use of the ITQOL, which was chosen based on extensive review of alternative measures. We chose the ITQOL over other measures [9, 15] because it has been most extensively used, is appropriate for measurement of HRQOL in healthy infants [20], and provides clearly defined concept scores. The inclusion of a mixed-fed group is an additional strength; most studies exclude this group despite widespread use of mixed feeding regimens and practitioner encouragement of partial breastfeeding when exclusive breastfeeding is not possible [33]. Finally, assessment of HRQOL in both parents and infants is important for appropriate adjustment for the potential confounding influence of parental HRQOL; the correlations between the SF-12v2 and the ITQOL parent-focused concepts highlight the importance of assessing both of these factors.

Limitations of our study include potential lack of generalizability to the overall population of Chinese infants given the high educational level of study parents, inclusion of infants consuming only one type of infant formula, and high HRQOL of study infants. There is also potential for residual confounding in our observational study due to unmeasured factors that influence both choice of feeding regimen and HRQOL. In addition, HRQOL was reported by parents not blinded to feeding regimen; it is possible that reporting of HRQOL was influenced by parents’ beliefs about the effects of feeding regimens on infant HRQOL, rather than based purely on observations of the infant’s well-being. Despite these limitations, our observational approach had the advantage of assessing feeding regimens and their outcomes in real-world settings.

Our study represents an important contribution to the limited literature on infant HRQOL, and highlights the feasibility and utility of measuring HRQOL in this population. We found high HRQOL concept scores in this population of healthy infants, regardless of which feeding regimen parents had chosen. Only a few small differences in HRQOL concept scores were observed between the breastfed, study formula-fed and mixed-fed infants and the effect size of these differences suggests they may be of limited clinical relevance. Furthermore, no significant differences were observed between groups in Overall Health, Physical Abilities, Growth and Development, or Bodily Pain/Discomfort scores, areas where nutrition might be expected to play a role. Based on these results, we conclude that, in this sample of healthy Chinese infants, there was little to no effect of the specific feeding regimens assessed in this study on infant HRQOL. Future research should examine relationships between feeding regimens and HRQOL in other infant populations including those with lower overall health and well-being.

ANCOVA, analysis of covariance; GI, gastrointestinal; HRQOL, health-related quality of life; ITQOL, infant and toddler quality of life questionnaire; MCS, mental component summary; PCS, physical component summary; SF-12v2, short form 12 health survey, version 2; WHO, World Health Organization