Adriëtte Oostvogels

132 Chapter 5 our study as the population of Moroccan and Turkish descent more often has diabetes compared to the Dutch population. 29 This study had some limitations as well: we did not have detailed information on how many and which family members (aunts/uncles or grandparents) were affected with what kind of diabetes (type 1 or type 2). In the Netherlands, of all patients diagnosed with diabetes, about 84% has type 2 diabetes mellitus (T2DM), although in the younger groups (30-40 years), this percentage is about 50%. 29 Little information is available on the distribution of type 1 and type 2 in the ethnic groups. Therefore it is unclear if the higher C-peptide levels found in the more ethnic diverse group of children with both maternal and paternal second-degree FHD are a result of FHD type 1 or type 2. Moreover, studies showed that the risk of adverse outcomes is increased with the number of family members that have diabetes. 4,7,30 In the group of children with both maternal and paternal second-degree FHD, it could be that these children simply had more affected family members than children in the other groups, thereby overestimating the effect of both maternal and paternal FHD. Third, 63.6% of the C-peptide values in the total sample were below the detection limit of 0.34 nmol/L. Therefore we predicted C-peptide based on characteristics of the whole group. 31 When repeating the analyses with the non-imputed values of C-peptide, we only found an association between paternal only and both maternal and paternal second-degree FHD and higher C-peptide levels.(Supplementary Table 4) However, as the number of children in the group with both maternal and paternal second-degree FHD was smaller than forty cases, these analyses could have been underpowered. Comparison to other studies To our knowledge, only eight studies have considered the FHD of grandparents, aunts and/or uncles on outcomes of body composition and/or glucose metabolism in childhood or adolescence. 14,16-21,32 In contrast to our adjusted models, Jouret et al. found an increased risk of overweight in preschool children of whom one of the grandparents had diabetes. 16 We cannot compare the results directly as, in this study, no distinction was made between maternal, paternal or both maternal and paternal FHD. 16 Similar to our study, five studies found negative effects of second-degree FHD on measures of glucose metabolism. 14,17, 19-21 Only two studies did not find associations between FHD and glucose metabolism. 18,32 Of all these aforementioned studies, only two separated effects of maternal, paternal or both maternal and paternal FHD and found, in contradiction to our results, stronger effects for maternal FHD. 19,20 In one of these studies, the children with maternal FHD (mean age = 11 years) were 2.5 years older than children without FHD (mean age = 8.5 years), therefore these differences can also have resulted from differences in maturation. 20 Moreover, both these studies did not investigate the combined effect of both maternal and paternal FHD. Studies