Canadian Pediatric Endocrine Group extension to WHO growth charts: Why bother?

Abstract

The Canadian Pediatric Endocrinology Group (CPEG) has produced complementary growth curves based on the 2010 ‘WHO Growth Charts for Canada’. In response to concerns from CPEG members and the general paediatric community regarding the presentation of the WHO data, complementary curves were generated, which the authors believe will enhance clarity, reduce potential errors in classification and enable users to better track short-term changes, particularly for weight in older children. Specifically, these curves extend weight-for-age beyond 10 years of age, restore additional percentiles within the normal range, remove extreme percentiles and harmonize the choice of body mass index percentiles with adult definitions of overweight and obesity. All modifications followed strict WHO methodology and used core data from the United States National Center for Health Statistics. The curves retain the clean appearance of the 2010 Canadian curves and are available from the CPEG website (http://cpeg-gcep.net).

The Canadian Pediatric Endocrinology Group (CPEG) has produced complementary growth curves based on the 2010 ‘WHO Growth Charts for Canada’ (1). Our members believe that our modifications enhance these 2010 curves, particularly for paediatric clinicians. CPEG recognizes and appreciates the methodology for data collection and analysis used by the WHO to produce growth curves that better reflect contemporary trends and recommendations in breastfeeding, and an ethnically diverse population of children (2–4). We also recognize the efforts of the Dietitians of Canada, the Canadian Paediatric Society, the College of Family Physicians of Canada, Community Health Nurses of Canada and others to integrate and promote these new growth curves in Canada (1). Given a longstanding subspecialty interest and expertise in growth, both normal and disturbed, our members have been intimately involved in introducing the WHO curves across the country. Based on experience in clinical practice, concerns have been raised by our membership and by the general paediatric community with respect to how the data are displayed on the new WHO curves (5). In response, the CPEG has generated complementary curves that we believe, in some instances, enhance clarity, minimize potential for errors in classification and enable clinicians to better track short-term changes, particularly regarding weight in older children (Figures 1 and 2). Adapted from their WHO counterparts, these curves differ in the following ways:

Figure 1).

CPEG growth charts for boys aged 2 to 19 years. A) Height-for-age and weight-for-age; B) BMI-for-age

Figure 2).

CPEG growth charts for girls aged 2 to 19 years. A) Height-for-age and weight-for-age; B) BMI-for-age

Continuation of the weight-for-age curve beyond 10 years of age

The continuation of the weight-for-age curves beyond 10 years of age will enable a simultaneous assessment of changes in both weight and height on the same page. While we recognize the clinical importance of promoting the use of body mass index (BMI) rather than weight in this age group, we believe that BMI and weight-forage are complementary measures that should be followed in parallel with height-for-age. This is particularly important for dietitians, gastroenterologists and other clinicians caring for children with conditions in which short-term changes in weight may impact linear growth and reflect disease activity. The ability to view both height and weight on one page is key to this pattern recognition. While we wholeheartedly embrace the enhanced use of BMI curves to track and diagnose obesity, we still need to track weight, height and BMI concurrently in older children. Extension of the weight curves does not diminish the utility of the BMI norms. These extended curves are based on the same North American ‘core data’ used to generate the WHO reference curves for school-aged children and adolescents, through strict application of WHO exclusion criteria and curve-fitting methods. For details, consult the CPEG Statistical Methods and Models manual (http://cpeg-gcep.net).

Inclusion of additional percentile lines within the normal ranges

The WHO Growth Curves Adapted for Canada depict one, two and three SDs above and below the mean, corresponding to percentiles 0.1, 3, 15, 50, 85, 97 and 99.9. This has resulted in fewer percentile lines between the 3rd and 97th percentile lines (2). We aimed to return to the more familiar Centers for Disease Control and Prevention (CDC) approach by including a fuller range of percentile lines (3, 10, 25, 50, 75, 90 and 97), enabling a more precise description within the normal range (6). The addition of the 25th and 75th percentile lines enables health care providers to more easily detect aberrations in growth at an earlier stage. This is particularly critical for weight-for-age, where failure-to-thrive is often defined by crossing two percentile lines (7). In returning to these traditional percentile lines, we hope to avoid inadvertent delays in the detection of abnormal growth (8).

Removal of extreme percentile lines (0.1 and 99.9)

Similarly, we have opted to remove the 0.1 and 99.9 percentile lines to avoid compression of the curves and to promote clarity in plotting and interpretation. There is the risk that growth between percentiles 0.1 and 3 or between percentiles 97 and 99.9 may be interpreted as normal simply because it is ‘on the curve’. Such misinterpretations persist in the community even though the Dietitians of Canada clearly define (for individuals two to 19 years of age) weight-for-age <3rd percentile as ‘underweight’ and height-for-age <3rd percentile as ‘stunted’, in both cases warranting further assessment (1). More extreme percentiles may actually be misleading: Although data from large numbers of children were used in the creation of the charts, percentiles 0.1 and 99.9 each refer to one in 1000 children. Such extreme percentiles are simply not well estimated with only a mean (± SD) of 673±204 boys per yearly interval and 646±185 girls per yearly interval between five and 19 years of age. As a result, these more extreme percentiles are based on very limited data.

Modification of the BMI percentile lines

As noted above, we continue to strongly endorse the use of the WHO BMI curves as an important tool. These differ from the CDC curves in their definition of obesity (6). Importantly, the WHO curves align more closely at 19 years of age with standard adult definitions for overweight (25 kg/m²) and obesity (30 kg/m²) (9). As highlighted above, the elimination of the extreme percentile lines (0.1 and 99.9) emphasizes more healthy BMIs while retaining the WHO definitions.

CONCLUSION

After discussion with other Canadian stakeholders (eg, Canadian Paediatric Society, Association of Quebec Pediatricians, Dietitians of Canada), CPEG has released these complementary curves. With the exception of extending the weight-for-age curves beyond 10 years of age, we have simply reformatted existing WHO curves to achieve the goals outlined above. The extension of the weight-for-age reference beyond 10 years of age follows from strict application of the WHO methodology to the core North American data used to create the original WHO reference curves (2). These revised curves retain the clean, well-designed appearance of the curves published by the Dietitians of Canada and are available from the CPEG website in both French and English (http://cpeg-gcep.net). A useful anthropometric calculator for calculating percentiles and z-scores is also available from this website.