In this study a large population of Punjabi school going children and adolescent girls (n = 10,050) of ages 8–16 years have been analysed for anthropometric measurements. We found very high prevalence of stunting with notable discrepancies in prevalence assessed by using two international height references. Prevalence of stunting was highest with CDC as compared to WHO reference and study population reference. Overall, the two international references for height-for-age showed a moderate degree of agreement among themselves when they were applied to estimate stunting (Fig. 1). Markedly, we found a poor agreement when the two references i.e. CDC and WHO were compared with the height-for-age cut-offs of study population. The results of present study highlight the importance of choice of references used to estimate the height status of children and adolescents to determine stunting, as it can significantly impact the strategies implemented by policy makers to address this abnormality. An appropriate assessment of stunting can also affect the decision of paediatricians to treat such a condition as now there is growing evidence that catch-up growth is possible till 15 years of age (14), thereby children can be stunted during early adolescence and can recover also by the use of timely strategies.
Even though recently there is growing evidence regarding the catchup growth possibilities in early adolescent period, data highlighting the prevalence of stunting in Pakistani adolescent population is very limited. Mostly stunting had been considered and studied in early childhood i.e. less than five years of age. Few studies conducted to estimate stunting in the school age children have applied WHO or CDC reference of height-for-age and found inconsistent results which reduces their usefulness in categorising height status in local population (20, 27–31, 38). We have tackled this problem in this study by applying two international and one local height cut-off reference to the overall Punjab and its regional population to allow a thoughtful comparison between studies regarding prevalence of stunting.
Moreover, when WHO standards were applied to our Pakistani school age children there was an overestimation of prevalence of stunting (20, 28). We have found the similar trend in our data when CDC and WHO references were applied to determine height status in our study population. There was an increasing trend of stunting with increasing ages in the schoolgirls when CDC and WHO references were used, however prevalence of stunting was comparable across all the age groups when study population cut-offs were applied. Additionally, we also found significant differences in stunting prevalence when we applied these references to the school age girls belonging to different regions of Punjab.
In the present study we found approximately 5% stunting overall in the schoolgirls of Punjab, which is partly in agreement with a previous study where they found 8% of children were stunted (28), but is in contrast to other studies which show stunting prevalence of 16% (6–12 years) (20), 16.7% & 14.3% [5–14 years; according to National Health Survey of Pakistan (NHSP) and Karachi surveys respectively] (38), 15.2% (5–14 years) (27), 45.2% (4–12 years), 36% (6–12 years) (30) and 35% (5–10 years) (31) among school children from various regions of the country. In these previous studies the study population (range = 5–14 yrs) was younger than the present study (8–16 years) and with smaller sample sizes (range n = 200–2072) than ours (n = 10,050). In these previous studies, only WHO reference was applied while ignoring CDC for estimating stunting prevalence. Moreover, we have also observed that when we applied WHO reference for height the overall prevalence of stunting was 17.49%, which is largely in agreement with the stunting prevalence reported in previous studies as 16% (20), 15.2% (27), 16.7% & 14.3% (38) conducted among school age children of rural Sindh, southern Punjab, NHSP and Karachi surveys respectively. Therefore, lower prevalence of stunting in our study cohort as compared to previous studies could be due to the WHO reference applied to estimate stunting. Unfortunately, there is no study available in literature from local population where indigenous/local height cut-off reference to define height status have been employed, rendering it impossible to compare estimates of stunting prevalence obtained by our local cut-offs with previous data. The prevalence of stunting as reported in few previous studies conducted on Pakistani school age children showed higher values and inconsistent results and differed from our study estimates (20, 27, 30, 31, 38).
We also noticed significant differences in the estimates of stunting when school going girls of northern, central and southern Punjab were compared (Figs. 2, 3, 4 & 5). Highest prevalence of stunting was found to be present among the northern Punjab schoolgirls (6.46%, p˂0.05; Fig. 5) and lowest prevalence was estimated among the girls of southern Punjab (3.18%, p˂0.05; Fig. 5). When we further explored the age wise differences in the stunting prevalence among northern Punjab schoolgirls, a strikingly notable finding was the very high stunting estimates of 12.20% and 15.61% among the age groups 12 and 13 years respectively (Fig. 2). The most likely explanation of this region-specific inconsistency could possibly be the differences in socio-economic status, maternal education status, provision of health facilities, heterogenous nature of population, diverse cultural practices, variable food habits, level of child’s growth and development associated with this chronic malnutrition condition manifested in the form of stunting (39). The identification of these regional differences in the prevalence of stunting especially in adolescent girls is extremely important to target the high prevalence areas with the implementation of programs to prevent complications such as poor obstetrical outcomes, obesity and neurological disorders.
This study emphasizes the importance of comparing internationally used references when applied to paediatric population and their degree of agreement with the study population own cut-offs. The need to generate population specific height references has been highlighted because of variations in height pattern among pre-adolescent and adolescent children especially 7–18 years of age (40). This comprehensive meta-analysis of mean height and pattern of linear growth of populations across the world shows that it is not justifiable to use a single reference throughout the globe to estimate height status, especially for the Asians (40). There are huge differences between the height status of adolescents belonging to different populations, thereby suggesting variations in growth pattern depending upon the genetic and environmental factors which exist in between these populations (41). This suggests the dire need to establish population specific references, which are difficult to validate as there is no gold standard available for comparison for this purpose. Moreover, for the authenticity of the population specific cut-offs, longitudinal studies with comprehensive application of references are required to be considered suitable for application especially in terms of predicting adverse health consequences in adulthood. There is very limited data available to assess the adult health status after applying local references. Although there exists a unanimous consensus for the application of standards to estimate stunting in the younger age groups i.e. till 5 years of age, height-for-age reference values have not been specified for the young adolescent group, thereby rendering it impossible to obtain consistent results regarding the prevalence of stunting among school age children.
This study has certain limitations. First, it is a cross-sectional study and stunting prevalence was estimated at one time point. Longitudinal studies are recommended to determine the trends in height status in our population, to generate population specific references and to determine the health outcomes of stunting in adults. Second this study includes only school age girls, we need to generate height reference data also for boys. Third, the study participants are the schoolgirls belonging to the province of Punjab. This study needs to be expanded to encompass data from all the four provinces of Pakistan.