This paper investigated childhood obesity risk using the UK-WHO charts in a sample of normal UK infants not selected on the basis of any defining characteristics. Our primary finding was that UK infants, on average, demonstrated a striking pattern of accelerated BMI growth against the UK-WHO charts. That pattern was characterised by an average gain between one and 12
months of age of 1.0 BMI Z-score, equivalent to upward crossing through approximately 1.5 centile bands (e.g., 50th
centile to the 85th
centile). Subsequently, we observed a greater risk of high infant BMI values and upward centile crossing using the UK-WHO charts compared to the previously used UK90 growth reference charts. The study design, which compared one sample of infants born at one point in time (2008–2009) to two different growth charts, means that the presented differences in obesity risk were entirely due to difference between the charts (e.g., source sample, statistical design) not because of any change in risk within the population. Our findings highlight the fact that the source sample of a growth chart has a fundamentally important role to play in childhood obesity risk classification; the switch from the UK90 reference to the UK-WHO charts will result in more infants being diagnosed as being on a trajectory toward childhood obesity. Indeed, combined data from the 2008/2009 and 2009/2010 National Diet and Nutrition Survey (NDNS) in the UK showed an obesity prevalence in two to three year old infants almost double that found in children aged four to 10
], assumedly because the infants were assessed using the UK-WHO charts whilst the children were assessed using the UK90 reference, not because of any real change in risk within the population. This is a perfect example of how the introduction of the UK-WHO charts may lead to misinterpretation of obesity risk in the UK.
Breastfed infants demonstrate slower growth than their bottle-fed counterparts [14
], with evidence suggesting that this may be because they are better at self-regulating their total energy intake [26
]. According to national statistics (13), the average infant in the UK does not follow the WHO feeding regime of exclusive breastfeeding to at least four months of age [10
]. When their growth is assessed against a chart based on data from infants who did follow the WHO feeding regime, they therefore demonstrate a pattern of accelerated growth indicative of increased risk for childhood obesity. This rationale is supported by the available literature, which shows that the choice to bottle-feed rather than breastfeed contributes most to accelerated infant growth [28
], which in turn contributes most to the development of childhood obesity [5
The Z-scores of White infants in the present study roughly approximate to those observed by Wright et al. [17
], who used data from the Avon Longitudinal Study of Parents and Children (ALSPAC) and the Gateshead Millennium baby Study (GMS) to test the WHO standard in the UK, at the cross-sectional ages where those researchers had data. The main strengths of the present study are that we modelled longitudinal data to supply continuous comparative information and assessed the risk of childhood obesity risk traits not included in the Wright et al. paper. The fact that we compared to the UK-WHO charts and not the WHO standard is actually of little importance because the only difference would have occurred at birth, where the UK-WHO charts are based on recalculated UK90 reference data [12
]. If anything, using the WHO standard data at birth instead of recalculated UK90 reference data would have resulted in a lower degree of subsequent accelerated weight gain because, in the present study, birth weight Z-scores according to the WHO standard were approximately 0.2 units higher than those using the recalculated UK90 reference data (mean values for White infants (sexes combined) +0.22 and −0.04, respectively). Wright et al. [17
] found that ALSPAC and GMS infants appeared large at birth using the WHO standard and focused on that finding, but not on the subsequent apparent accelerated growth which has major implications on the interpretation of childhood obesity risk in the UK.
The present paper helps inform the practitioner, who ultimately faces the interpretive dilemma of wondering whether a specific growth pattern should give cause for concern, about average growth patterns and risks of childhood obesity according to the UK-WHO charts. The correct interpretation of apparent accelerated growth, however, presupposes that the practitioner knows that the UK-WHO chart is essentially a growth standard, knows the differences between a standard and a reference, and knows the difference in how to interpret growth when using a standard compared to when using a reference. Because the number of UK infants who are diagnosed as being on a trajectory toward childhood obesity must have increased since the introduction of the UK-WHO charts, an increased effort to integrate the dissemination of useful information to parents with growth monitoring practice is necessary.
Not being able to test the relative contribution of factors responsible for accelerated growth of UK infants against the UK-WHO charts is perhaps the greatest limitation of the present study, because it would have allowed us to definitively answer the question “why do UK infants demonstrate obesogenic growth trajectories (against the UK-WHO charts)”?. Other limitations include a sample composed exclusively of infants from one city in the UK, which may limit generalisability of the results, and weight and BMI gain variables that did not account for regression to the mean. Gain variables would normally be calculated as the residuals from the general linear regression of size at age “T” on size at age “T-1” [24
]. However, the relative risk analysis in the present study needed to include data from the same individuals against the UK-WHO charts and against the UK90 reference charts, so any initial regression would need to account for this non-independence; no formula has yet been proposed.