Height of individuals has long been considered as a significant index of nutrition and health of a population [1, 2]. In the last century, a consistent increase in mean height and weight has been found in children and adults, mirroring improvements in nutritional [3] and socio-economic status [4–9]. Although in Europe height has been increasing in most populations [9–12], some recent studies have reported that the increase in height has reached a plateau in Germany [13] or Poland [14]. This plateau was attributed to the fact that the corresponding populations had achieved their full genetic potential and/or that their socio-economic conditions had ceased to improve [13–15].

There is little information regarding the trends of height and weight among developing or transitional countries. Many African countries show no increase or even a decrease in height [1], albeit an increase was noted in some of them, e.g. Kenya and Senegal [1]. An increase in children height and weight was also reported for Brazil [16], Iran [17], Mexico [18], the Cook islands [7] and India [15]. Hence, studies of secular changes in height and weight in populations are useful for providing information on nutritional status in early life, updating growth reference standards, and providing insight with regards to epidemiological trends of cardiovascular disease [19, 20]. Further, to our knowledge, little if no information is available regarding height and weight trends in Eastern Africa.

Thus, in this paper, we report the secular trends in height and weight among representative samples of children and adolescents of the Seychelles, a rapidly developing middle-income island state in the Indian Ocean (Eastern African region).

To our knowledge, this is the first report on secular trends among children and adolescents of the African region. In this study, a secular upward trend for height and weight among children and adolescents of the Seychelles could be demonstrated. For instance, at age 15.5 years, boys/girls were on average 10/13 cm taller and 15/9 kg heavier in 2005–6 than in 1956–7.

In most countries for which trends in children and adolescents' height and weight have been published, a secular increase has been noted [16, 27–33], although some studies have reported a stabilization [13, 14] or even a downward trend [34], likely due to unfavorable conditions such as war [17] or economical crisis [34]. In the Seychelles, the upward trends in height and weight were comparable in magnitude to those found in other countries, i.e. the USA [35], Australia [30], Brazil [7] and Turkey [33] and paralleled the increase in the per capita national gross domestic product, which rose in real terms from US$ 2927 in 1980 to US$5239 in 2004 [21].

The secular trends in height and weight can be affected by nutritional [3] and socio-economic factors [5, 7, 8], although this last statement has been challenged [36]. Other factors such as earlier maturation have also been suggested [27, 30, 37, 38]. Indeed, data from food balance sheets indicate that the per capita calorie availability has increased substantially in Seychelles, from 1800 kcal in 1965 to 2300 kcal in the late 1980s, and above 2400 kcal in the early 2000s [39]. The proportion of carbohydrates has decreased over time (74% of total calories in 1965 and 55% in 2000) while the proportion of fats has increased (16% in 1965 and 32% in 2000) [39]. Also, according to households expenditure surveys, between 1983 and 1993 the consumption of meat products increased by 238% in the Seychelles, while the consumption of fish, fruits and vegetables decreased by 33% [40]. Thus, it is likely that the improvements in socio-economic factors and the changes in food intake and in physical activity levels which occurred in the last decades in the Seychelles might partly explain the increases in height and weight observed in this study, although other factors cannot be ruled out. For instance, in the study conducted in 1956–7, a significant percentage of children presented with stunting [25], a very rare condition in the Seychelles nowadays. As the height of children is influenced by their parents' height, the fact that in the 1950s a substantial proportion of now-become parents were stunted might have reduced the increase in their children's height and weight, irrespective of a current good nutrition environment. Finally, some authors have suggested that secular trends in growth are more responsive (or plastic) to changes in the environment in boys than in girls [4, 41]. This might partly explain the higher increase in height in boys in the first period (1956–7 to 1998–9) of this study.

It has also been described that secular trends in height observed in young children might be carried over into adulthood [42]. Although data on adult height in the Seychelles is not available for the 50-year period under study, two types of information can help address indirectly this issue. First, in girls, the secular increase in height was much smaller at the age of 16 years than for the younger age groups (figure 1); for boys (figure 2) the results are inconclusive as the rapid growth period during puberty typically ends after the age of 16, for which no data is available for comparison over time in this study. Second, based on two population surveys in adults [22, 23], height at age 25–34 years was virtually identical in 1989 and 2004. Overall, these findings support the hypothesis that at least part of the increase in height in children during the past decades reflects accelerated growth during childhood over successive cohorts without a commensurate increase in height in adulthood, i.e. children grew faster during childhood but reached similar height after completion of adolescence, a trend also reported for Greenlandic children [27] and for India, where after a significant positive secular trend in height attained over the first 20 years, the adult height has now plateaued [15]. Finally, it is of interest to note that, in girls, the increase in height per decade was higher for period 1998–9 to 2005–6 than for period 1956–7 to 1998–9, suggesting an acceleration, a feature also reported in Mexico [18], where no secular change in height was noted among primary school children between 1968 and 1978, with a significant increase for period 1978–2000.

Weight increased over time at a much larger pace than height. These findings are in agreement with a marked secular increase in BMI reported in the Seychelles [24] and in other countries [43–46], with some exceptions [47]. This larger increase over time in weight than height might reflect a larger increase over time in adiposity relative to muscle mass [29, 48]. Indeed, the stronger increase in weight relative to the increase in height led to an increase in BMI and the prevalence of obesity among Seychellois children and adolescents. The upward trends in weight and BMI are partly associated with a concomitant decrease in levels of physical activity over time, as shown in Seychelles [24] although data on the relationship between lack of physical activity and overweight are still inconclusive [49]. Still, as no data for physical activity is available for the 1956–7 period, it was not possible to assess the possible impact of changes in physical activity levels on the increase in weight. Also, the effects on weight of an increased energy intake [39, 40] or of improvements in health services such as a decrease in intestinal parasites among children from 95% in 1956–7 [25] to less than 15% in 1997 [50] cannot be ruled out. Overall, these findings are consistent with a shift from substantial under nutrition in children in 1956–7 to an ongoing epidemic of overweight/obesity in the youth population of the Seychelles.

The observed secular increases in height and weight stress the need to adequately update reference growth curves. Indeed, many reference growth curves in children have been developed several decades ago and might no longer reflect the current distribution of body height and weight [29, 32, 33, 51]. However, if updates in height norms can be validly and simply carried out based on current data, the update in weight norms, using current data, is more problematic as many countries currently face an increase in pediatric obesity [52], inclusive Seychelles [24]. Deriving weight standards from the current anthropometric data would artificially decrease the prevalence of obese children and adolescents. Thus, the sex-and age- specific cut off values for overweight and obesity should be established carefully for the pediatric population and it has been advised that universal internationally agreed-upon thresholds be used [53]. Further, increased height in current vs. previous generations of children may have important implications with regards to the measurement and interpretation of indicators that are related to height. For example, cut off values for high blood pressure, which relate strongly with a child's age and height [54], might need to be adjusted if children of same age are taller nowadays than before.

This study has some limitations. First, it was not possible to assess individual data for the period 1956–7, and only average age-specific values were available for comparison. Still, this method has been used by others in order to assess trends for long time periods [14, 32]. Second, no data on pubertal status was collected, preventing the possibility of assessing whether pubertal age has decreased over time. Third, data available for comparison between time periods were limited to children aged <16 years, which prevented to directly examine if the observed height gain in children over successive cohorts tracked into adulthood. Fourth, estimations are only based on 3 cross-sectional surveys over the whole study period, and the period 1998–2006 might be too short to properly trends; further, the increase in anthropometric measurements might be non-linear, but as only three time points were available, a non-linear model could not be applied. Still, as the yearly assessment of height and weight is currently under way, a better trend assessment will be achievable within some years. Also, it will be possible to follow the cohorts of children, thus leading better estimates of growth curves. Finally, the main strengths of this study are the long time interval examined (50 years), the large size of the samples (particularly for 1998–2006), the population-based nature of the data in all periods, and the use of standardized measurements throughout the study period (particularly for 1998–2006).

In summary, a secular upward trend in body height was documented in children and adolescents of both genders in the Seychelles and the rate of increase was not different during the last decade vs. the previous 4 decades. Indirect evidence suggests that this upward height gain in youth over time may not have translated in a commensurate height gain in adults and further studies should be conducted to fully assess the significance of such upward height trends among youth. Conversely, the markedly steeper upward secular trend in weight than height suggests that the young population of the Seychelles is experiencing an epidemic of overweight and obesity.