Maternal and paternal transmission of type 2 diabetes: influence of diet, lifestyle and adiposity
Ali Abbasi, MD, Department of Epidemiology, University Medical Center Groningen, Hanzeplein 1, PO Box 30.001, 9700 RB Groningen, the Netherlands.
(fax: 0031 50 361 4493; e-mail: email@example.com).
Abstract. Abbasi A, Corpeleijn E, van der Schouw YT, Stolk RP, Spijkerman AMW, van der A DL, Navis G, Bakker SJL, Beulens JWJ (University Medical Center Groningen, Groningen; University Medical Centre Utrecht, Utrecht; and National Institute for Public Health and the Environment (RIVM), Bilthoven, the Netherlands) Maternal and paternal transmission of type 2 diabetes: influence of diet, lifestyle and adiposity. J Intern Med 2011; 270: 388–396.
Objective. Transmission of family history of type 2 diabetes to the next generation is stronger for maternal than paternal diabetes in some populations. The aim of the present study was to investigate whether this difference is explained by diet, lifestyle factors and/or adiposity.
Methods. We analysed 35 174 participants from the Dutch contribution to the European Prospective Investigation into Cancer and Nutrition, a prospective population-based cohort (aged 20–70 years) with a median follow-up of 10.2 years. Parental history of diabetes was self-reported. Occurrence of diabetes was mainly identified by self-report and verified by medical records.
Results. Amongst 35 174 participants, 799 incident cases of diabetes were observed. In age- and sex-adjusted analyses, hazard ratio (HR) and 95% confidence intervals (CIs) for diabetes by maternal and paternal diabetes were 2.66 (2.26–3.14) and 2.40 (1.91–3.02), respectively. Maternal transmission of risk of diabetes was explained by diet (9.4%), lifestyle factors including smoking, alcohol consumption, physical activity and educational level (7.8%) and by adiposity, i.e. body mass index and waist and hip circumference (23.5%). For paternal transmission, the corresponding values were 2.9%, 0.0% and 9.6%. After adjustment for diet, lifestyle factors and adiposity, the HRs for maternal (2.20; 95% CI, 1.87–2.60) and paternal (2.23; 95% CI, 1.77–2.80) transmission of diabetes were comparable.
Conclusions. Both maternal and paternal diabetes are associated with increased risk of type 2 diabetes, independently of diet, lifestyle and adiposity. The slightly higher risk conferred by maternal compared to paternal diabetes was explained by a larger contribution of diet, lifestyle factors and adiposity.
The prevalence of type 2 diabetes, a leading cause of cardiovascular morbidity and mortality, is increasing worldwide . Parental – paternal and/or maternal – history of diabetes is a major determinant of increased risk of diabetes [2–5]. Family history may reflect complex relationships between genetic factors and environmental conditions that are important for developing diabetes . Thus, parental history of diabetes includes environmental risks (e.g. nongenetic familial behaviours, lifestyle and obesity) beyond the genetic risk factors for diabetes .
A greater risk from maternal type 2 diabetes compared to paternal diabetes has been reported in some [7–9] but not all studies [2, 4, 5]. A variety of explanations for this greater importance of maternal diabetes have included the following: genomic imprinting (i.e. the differential expression of inherited susceptibility genes in paternal or maternal generation ); mutations in mitochondrial DNA, which are maternally inherited ; and metabolic programming during intrauterine exposure . It is still not clear to what extent modifiable factors such as diet, lifestyle and obesity can explain the association between maternal or paternal diabetes and risk of diabetes. To our knowledge, only one prospective study amongst female nurses has investigated the contribution of excess adiposity and certain dietary habits . No such longitudinal data are available in men.
The aim of this study was to prospectively investigate the association between parental history of diabetes – maternal and/or paternal – and risk of incident type 2 diabetes in a population-based cohort of male and female adults, i.e. the Dutch contribution to the European Prospective Investigation into Cancer and Nutrition (EPIC-NL). The EPIC-NL study was suitable for this purpose because detailed data on diabetes risk factors such as diet and lifestyle factors were collected in this cohort .
Study population and design
The EPIC-NL cohort (n = 40 011) includes the Monitoring Project on Risk Factors for Chronic Diseases (MORGEN) and Prospect cohorts, initiated between 1993 and 1997. Details of the EPIC-NL study design, recruitment and procedures have been described in more detail previously . Briefly, Prospect is a prospective population-based cohort study of 17 357 women aged 49–70 years who participated in a breast cancer screening programme. In the MORGEN study, 22 654 individuals aged 20–59 years were recruited from Amsterdam, Doetinchem and Maastricht. A new random sample of about 5000 participants was examined each year. These rounds of enrolment add up to this number of individuals. The participation rates were 34.5% for Prospect and 45.0% for MORGEN.
At baseline, a general questionnaire and a food frequency questionnaire (FFQ) were sent by post to all participants and these were returned after completion at the medical examination. We excluded 1150 participants with missing data on baseline characteristics or extreme values for energy intake (<450 or >6000 kcal day−1) and 2360 participants with unknown history of parental diabetes. Further subjects were excluded because of prevalent type 2 diabetes (n = 507) or missing recordings of censoring time (n = 820). Follow-up time was calculated from the date of enrolment to the date of diabetes diagnosis or death. All other participants were censored at the end of follow-up (January 2006). Finally, 35 174 participants were included in the cross-sectional and prospective analyses.
All participants gave written informed consent prior to study inclusion. Both cohort studies complied with the Declaration of Helsinki. Prospect was approved by the Institutional Review Board of the University Medical Center Utrecht and MORGEN was approved by the Medical Ethics Committee of the Netherlands Organization for Applied Scientific Research.
The general questionnaire contained questions on demographic characteristics and risk factors for the presence of chronic diseases. For both cohorts, coding of this information was standardized and merged into one uniform database. Body weight, height and waist and hip circumference were measured according to standard procedures. Smoking status was categorized into current, past and never smoker. Physical activity was assessed using a questionnaire validated in an elderly population and categorized as inactive, moderately inactive, moderately active and active, according to the Cambridge Physical Activity Score . Low education level was defined as primary education, lower vocational education or advanced elementary education. Blood pressure was measured twice on the left arm. The mean of the two blood pressure measurements was used in the analysis. In the Prospect study, systolic and diastolic blood pressures were measured with the participants in the supine position using a Boso Oscillomat (Bosch & Sohn, Jungingen, Germany), whereas a random-zero sphygmomanometer (Hawksley & Sons, Lancing, UK) with the participant in the sitting position was used in the MORGEN cohort. The comparability of these different measurement procedures has been described in more detail previously . The assessment of the Prospect cohort slightly overestimated blood pressure compared with the MORGEN cohort. Hypertension was defined based on self-report of diagnosis by a physician, measured hypertension (≥140 mmHg systolic blood pressure or ≥90 mmHg diastolic blood pressure) or the use of blood pressure-lowering medication. Hyperlipidaemia was defined based on self-report of diagnosis by a physician or the use of lipid-lowering therapy.
In both cohorts, daily food intake was determined using the same validated FFQ [16, 17], which contains questions on the usual frequency of consumption of 79 main food groups during the year preceding enrolment. Overall, the questionnaire enables estimation of the average daily consumption of 178 foods. Intakes of different nutrients were adjusted for total energy intake using the regression residual method .
Assessment of parental history of diabetes
Parental history of diabetes was obtained by self-report. Participants were asked whether their biological mother and/or father had (whether alive or deceased) previously been diagnosed with diabetes. Parental history of diabetes was categorized as none, any parent(s) (mother and/or father), maternal only, paternal only or both.
Assessment of type 2 diabetes
Occurrence of diabetes during follow-up was self-reported via two follow-up questionnaires at 3- to 5-year intervals in the MORGEN and Prospect studies. In the Prospect study, incident cases of diabetes were also detected as glucosuria via a urinary glucose strip test, which was sent out with the first follow-up questionnaire. Diagnoses of diabetes were also obtained from the Dutch Center for Health Care Information, which holds a standardized computerized register of hospital discharge diagnoses. Follow-up was complete until 1 January 2006. Potential cases identified by these methods were verified against general practitioner or pharmacist records (Prospect only) via postal questionnaires . Diabetes was defined as being present when the diagnosis was confirmed by either of these methods. For 89% of participants with potential diabetes, verification information was available, and 72% were verified as having type 2 diabetes and were thus included in the analysis .
Baseline descriptive statistics of the continuous variables were reported as mean ± standard deviation (SD) and groups were compared using two-tailed Student’s t-test or anova. Categorical variables were presented as numbers and percentages and a chi-squared test was used to test the differences between participants without parental history of diabetes and those with each category of parental history of diabetes for these variables.
Generalized linear models were used to assess the cross-sectional associations between parental diabetes and baseline parameters of obesity, including body mass index (BMI) and waist and hip circumference in participants. Multivariable models were adjusted for cohort (Prospect or MORGEN), age, sex, lifestyle factors (smoking, alcohol consumption, physical activity level and educational level), total energy intake and energy-adjusted dietary factors. The dietary factors included the amount of intake of fat, protein, carbohydrate, cholesterol, fibre, vitamin C and vitamin E. The estimated marginal means and 95% confidence intervals (CIs) were reported and linear regression β coefficients and 95% CIs were calculated for each category of parental history of diabetes compared to the reference group of participants without any parental history of diabetes.
The association between parental history of diabetes and incident diabetes in participants was assessed by Cox proportional hazard regression. In the crude model (controlled for cohort), hazard ratios (HRs) and 95% CIs for diabetes were calculated for each category of parental diabetes against a reference group of participants without parental history of diabetes. In Model 1, basic adjustments were made for age and sex. We assessed the effect of sex by including the interaction of sex with parental diabetes in this model. Moreover, the stratified analyses for sex were fitted in adjusted models for age and other covariates. Lifestyle factors were added in Model 2. Total energy intake and energy-adjusted dietary factors were added in Model 3. Parameters of obesity were subsequently included in the final model (Model 4). We then separately added each factor to Model 1 to determine its contribution to the association between parental history of diabetes and risk of diabetes. Inclusion of these factors in the model would be expected to attenuate the HR related to parental diabetes. We calculated the percentage attenuation of the HR for each category of parental diabetes. Percentage attenuation of HR was calculated as: (HR before addition − HR after addition)/(HR before addition − 1) × 100. A P value of 0.05 or less from two-sided tests was considered statistically significant. All statistical analyses were carried out using Statistical Package for Social Sciences version 16 (SPSS Inc, Chicago, IL, USA) and Stata software version 10.0 (Stata-Corp LP, College Station, TX, USA).
Baseline characteristics of the study population are summarized in Table 1 by parental diabetes status. When compared with participants without parental history of diabetes, those who reported paternal and/or maternal diabetes were older and more likely to be woman, had a higher BMI, waist and hip circumference and blood pressure, a lower alcohol consumption and education level, were less physically active and were more likely to experience cardiovascular morbidity. Parental diabetes was associated with a lower intake of total energy and carbohydrates, whereas the intake of protein, fat, fibre, vitamin C and vitamin E was higher in participants with parental diabetes.
Table 1. Baseline characteristics by parental history of diabetes in the EPIC-NL study (n = 35 174) a
|No. (%) of participants||28 696 (81.6)||6478 (18.4)||2187 (6.2)||3941 (11.2)||350 (1.0)|
|Age (years)||48.4 (12.3)1,*||51.6 (10.0)||49.7 (10.8)2,*||52.6 (9.4)||52.0 (9.7)|
|Woman||21 026 (73.3)1,*||5148 (79.5)||1682 (76.9)||3178 (80.6)||288 (82.3)|
|Body mass index (kg m−2)||25.4 (3.9)1,*||26.4 (4.1)||26.0 (3.9)2,*||26.6 (4.2)||27.2 (4.6)|
|Waist circumference (cm)||84.5 (11.2)1,*||86.8 (11.5)||85.8 (11.3)2,*||87.3 (11.6)||88.4 (12.0)|
|Hip circumference (cm)||103.0 (7.9)1,*||104.7 (8.5)||103.9 (7.9)2,*||105.1 (8.7)||105.9 (9.7)|
|Systolic blood pressure (mmHg)||125 (19)1,*||128 (19)||126 (19)2,*||130 (20)||128 (19)|
|Diastolic blood pressure (mmHg)||77.4 (10.6)1,*||78.8 (10.4)||77.9 (10.6)2,*||79.4 (10.4)||78.2 (10.4)|
|Alcohol consumption (g week−1)||11.4 (15.6)1,*||9.9 (14.4)||10.6 (14.4)||9.7 (14.5)||7.9 (12.0)|
|Current smoker||8560 (29.8)1,*||1873 (28.9)||614 (28.1)||1171 (29.7)||88 (25.1)|
|Low educational levelb||15 606 (54.4)1,*||4229 (65.3)||1250 (57.2)2,*||2734 (69.4)||245 (70.0)|
| Inactive||2450 (8.5)1,*||626 (9.7)||185 (8.5)||402 (10.2)||39 (11.1)|
| Moderately inactive||8298 (28.9)||1837 (28.3)||649 (29.7)||1094 (27.8)||94 (26.9)|
| Moderately active||8045 (28.0)||1763 (27.2)||604 (27.6)||1049 (26.6)||110 (31.4)|
| Active||9903 (34.5)||2252 (34.8)||749 (34.2)||1396 (35.4)||107 (30.6)|
| Total energy intake (kcal day−1)||2076.6 (636.7)1,*||1992.3 (601.0)||2004.4 (604.8)||1991.4 (602.4)||1926.6 (556.9)|
|Nutrient intake (g day−1)d|
| Protein||75.5 (10.9)1,*||76.7 (10.9)||76.0 (10.7)2,**||76.9 (11.0)||78.5 (11.7)|
| Fat||77.3 (11.3)1,*||78.4 (11.5)||78.0 (10.8)||78.5 (11.9)||79.1 (11.2)|
| Saturated fat||32.4 (5.8)1,*||33.0 (5.9)||32.6 (5.6)2,**||33.1 (6.0)||33.1 (5.6)|
| Monounsaturated fat||29.4 (5.1)1,**||29.6 (5.2)||29.6 (5.0)||29.6 (5.4)||29.9 (5.2)|
| Polyunsaturated fat||14.9 (3.8)1,*||15.1 (4.0)||15.1 (3.9)||15.1 (4.0)||15.5 (4.1)|
| Cholesterol (mg day−1)||215.9 (58.3)1,*||221.7 (62.6)||218.1 (59.3)2,***||221.9 (64.1)||230.7 (63.9)|
| Carbohydrates||222.3 (30.7)||220.8 (30.7)||221.2 (30.0)||220.6 (31.1)||220.7 (31.3)|
| Mono- and disaccharide||112.5 (29.3)1,***||111.7 (29.5)||111.2 (28.5)||112.1 (29.9)||110.5 (30.9)|
| Fibre||23.3 (4.8)1,*||23.7 (4.8)||23.4 (4.7)2,***||23.8 (4.8)||23.7 (4.6)|
| Vitamin C (mg day−1)||109.2 (45.1)1,*||111.6 (46.1)||110.1 (44.1)||112.4 (47.0)||112.7 (47.8)|
| Vitamin E (mg day−1)||12.2 (3.2)1,*||12.4 (3.3)||12.3 (3.3)||12.4 (3.3)||12.7 (3.5)|
|Hyperlipidaemiae||2268 (7.9)1,*||617 (9.5)||165 (7.5)||414 (10.5)||38 (10.9)|
|Hypertensionf||5889 (20.5)1,*||1621 (25.0)||489 (22.4)||1023 (26.0)||109 (31.1)|
|Myocardial infarction||434 (1.5)1,**||121 (1.9)||28 (1.3)||82 (2.1)||11 (3.1)|
|Stroke||309 (1.1)1,*||101 (1.6)||26 (1.2)||65 (1.6)||10 (2.9)|
|Cancer||1187 (4.1)||264 (4.1)||80 (3.7)||171 (4.3)||13 (3.7)|
Table 2 shows the association between parental history of diabetes and parameters of obesity in participants. We calculated adjusted means of parameters of obesity in each category of parental diabetes accounting for cohort, age, sex, diet and lifestyle factors. Subjects with maternal and/or paternal diabetes had higher BMI (β coefficient, 0.65; 95% CI, 0.55–0.75), waist circumference (β coefficient, 1.88; 95% CI, 1.62–2.15) and hip circumference (β coefficient, 0.93; 95% CI, 0.72–1.13) compared with participants without parental diabetes. This association was stronger for those with both maternal and paternal history of diabetes.
Table 2. Cross-sectional association between parental history of diabetes and body mass index (BMI) and waist and hip circumference (n = 35 174)
|BMI (kg m−2)|
| Estimated means (95% CI)a,b||25.8 (25.7–25.8)||26.4 (26.3–26.5)||26.2 (26.0–26.3)||26.5 (26.4–26.7)||27.0 (26.6–27.3)|
| β coefficient (95% CI)||0||0.65 (0.55–0.75)||0.39 (0.23–0.55)||0.75 (0.63–0.88)||1.20 (0.81–1.56)|
|Waist circumference (cm)|
| Estimated means (95% CI)a,b||87.8 (87.7–87.9)||89.7 (89.4–89.9)||89.1 (89.7–89.5)||89.9 (89.6–90.2)||90.9 (89.9–91.9)|
| β coefficient (95% CI)||0||1.88 (1.62–2.15)||1.28 (0.86–1.70)||2.12 (1.79–2.44)||3.12 (2.10–4.13)|
|Hip circumference (cm)|
| Estimated means (95% CI)a,b||102.9 (102.8–103.1)||103.9 (103.7–104.1)||103.4 (103.1–103.8)||104.1 (103.8–104.3)||104.5 (103.7–105.3)|
| β coefficient (95% CI)||0||0.93 (0.72–1.13)||0.50 (0.17–0.83)||1.11 (0.86–1.37)||1.55 (0.75–2.35)|
During a median follow-up of 10.2 years, we observed 799 incident cases of type 2 diabetes (rate of 2.2 per 1000 person-years). In the unadjusted analysis, participants with parental history of diabetes had an approximately 3-fold higher incidence rate of diabetes compared with those who had no parents with diabetes (Table 3; 1.7 vs. 5.0 per 1000 person-years, P < 0.001). Despite the sex differences in each category of parental diabetes, there was no significant interaction of sex with parental diabetes (HR of interaction term, 0.86; 95% CI, 0.60–1.22). In sex-stratified analyses, multivariable-adjusted HRs of diabetes for maternal and paternal history of diabetes were 2.50 (95% CI, 1.72–3.62) and 2.14 (95% CI, 1.72–3.62), respectively, in male participants. In women, these values were 2.15 (95% CI, 1.79–2.6) and 2.32 (95% CI, 1.8–3.0), respectively.
Table 3. Parental history of diabetes, contributing factors and risk of incident type 2 diabetes over 10 years (n = 35 174)
|Incidence rate, per 1000 person-year||1.7||5.0||3.9||5.1||10.9|
|Crude modela||1 (ref.)||2.90 (2.51–3.34)||2.36 (1.88–2.97)||2.91 (2.46–3.43)||6.21 (4.42–8.71)|
|Model 1||1 (ref.)||2.75 (2.39–3.17)||2.40 (1.91–3.02)||2.66 (2.26–3.14)||5.89 (4.20–8.27)|
|Model 2||1 (ref.)||2.64 (2.29–3.05)||2.40 (1.90–3.01)||2.53 (2.14–2.98)||5.49 (3.91–7.71)|
|Model 3||1 (ref.)||2.58 (2.24–2.97)||2.37 (1.88–2.98)||2.46 (2.09–2.91)||5.02 (3.57–7.06)|
|Model 4||1 (ref.)||2.32 (2.01–2.68)||2.23 (1.77–2.80)||2.20 (1.87–2.60)||3.92 (2.78–5.50)|
In total, crude HRs of incident diabetes for maternal and paternal history of diabetes were 2.91 (95% CI, 2.46–3.43) and 2.36 (95% CI, 1.88–2.97), respectively, when compared with those who reported no parental diabetes. Model 1 in Table 3 shows that adjustment for age and sex modestly attenuated (13.1% reduction) the risk of diabetes by maternal diabetes (HR, 2.66; 95% CI, 2.26–3.14]), whereas this did not contribute to the risk conferred by paternal diabetes (HR, 2.40; 95% CI, 1.91–3.02). After multivariable adjustment (Model 4) for age, sex, diet, lifestyle factors and parameters of obesity, risk of diabetes was comparable for maternal (HR, 2.20; 95% CI, 1.87–2.60) and paternal history of diabetes (HR, 2.23; 95% CI, 1.77–2.80).
It is interesting that age, sex, diet, lifestyle factors and parameters of obesity contributed more to the association between maternal diabetes and risk of diabetes (overall attenuation of 37.1%) than paternal diabetes (overall attenuation of 9.6%). Therefore, we separately added each factor to Model 1 to assess its contribution to the association between category of parental diabetes and risk of diabetes. Parameters of obesity explained 23.5% and 9.6%, respectively, of the risk estimation of diabetes by maternal and paternal diabetes. Risk estimation of maternal diabetes was partly explained (9.4%) by energy intake and dietary determinants, whereas this accounted for only 2.9% of the association between paternal diabetes and risk of diabetes. After adjustment for lifestyle factors, an attenuation of 7.8% was observed in the association between maternal diabetes and risk of diabetes, whereas risk estimation of paternal diabetes was not affected by lifestyle factors.
In this prospective cohort with over 10 years of follow-up, we found that both maternal and paternal history of diabetes were associated with baseline diabetes risk factors and with an increased risk of incident type 2 diabetes in participants, independent of diet, lifestyle and adiposity. However, the association between maternal diabetes history and risk of diabetes was slightly stronger in the age- and sex-adjusted model compared with paternal history. More than one-third of the maternal transmission of diabetes was explained by age, sex, diet, lifestyle factors (smoking status, alcohol consumption, physical activity and educational level) and parameters of obesity. The association between paternal diabetes and incident diabetes, however, was explained only modestly (approximately 10%) by diet and parameters of obesity.
The main strengths of our study are its large sample size, prospective design, verification of incident diabetes and extensive information about participants’ diet and lifestyle factors. Nevertheless, our study has some limitations. The EPIC-NL cohort almost exclusively comprised Caucasians from the Netherlands, and it is unclear whether our findings could be extended to other ethnic groups. Another limitation is that parental history of diabetes was obtained by self-report which is the usual method in single-generation cohorts. Furthermore, we excluded individuals with missing data or unknown parental history of diabetes. Having unknown family history has been shown to be more common for paternal than for maternal diabetes . However, the baseline characteristics of excluded individuals were similar to those who were included in our analysis. Therefore, it is unlikely that this would have led to recall bias or misclassification by category of parental diabetes of participants who did or did not develop diabetes. We relied on self-reported information about lifestyle and diet, which may be subject to misclassification. However, both the physical activity questionnaire and the FFQ have been validated previously [14, 16, 17]. These studies showed that both questionnaires could be used to rank individuals according to their physical activity or diet. We therefore believe that this did not greatly influence our results. Finally, individuals with type 2 diabetes may remain undiagnosed for several months to years and diagnosis of diabetes is always challenging in observational studies. Some cases of type 2 diabetes may have been undetected, resulting in underestimation of the association between parental diabetes and risk of diabetes in participants.
We first investigated the association between parental history of diabetes and baseline parameters of obesity in cross-sectional analyses. In multivariable-adjusted models, maternal or paternal diabetes was associated with higher BMI and waist and hip circumference. Having both maternal and paternal diabetes was associated with higher parameters of obesity. These findings are in agreement with those of other studies demonstrating that the presence of a maternal or paternal history of diabetes is associated with greater adiposity [21, 22] and weight gain , thus suggesting that diabetes and obesity share some common heritable determinants [5, 24].
A slightly more important role of maternal, compared with paternal, transmission of diabetes was shown in the present study in the age- and sex-adjusted model. This difference (approximately 25%) was explained by diet and adiposity as well as age, sex and lifestyle factors for maternal diabetes. Of note, multivariable models have not been used to assess these factors in previous studies investigating the increased importance of maternal transmission [8, 9, 21]. Amongst these studies, adiposity substantially explained the risk of diabetes transmitted by maternal diabetes, whereas other factors contributed to a lesser extent.
In the present study, both maternal and paternal transmission of diabetes were explained to some extent by obesity parameters, with a 2-fold higher contribution for maternal than for paternal diabetes. The contribution of dietary determinants was also larger in the association between maternal diabetes and risk of diabetes when compared with paternal diabetes. In addition, the influence of age, sex and lifestyle factors was confined to the association between maternal diabetes and risk of diabetes. This finding was not observed in a recent analysis from the Nurses’ Health Study (NHS). In the NHS, BMI rather than waist and hip circumference largely explained the association between both maternal and paternal history of diabetes and risk of type 2 diabetes. In addition to BMI, it was suggested that a higher intake of red meat and sugar-sweetened beverages and lack of alcohol consumption may explain part of the association between family history of diabetes and risk of diabetes . Of interest, in our study, parental history of diabetes was related to a lower intake of total energy and carbohydrates but a higher energy-adjusted intake of fat, protein, fibre, vitamin C and vitamin E. These differences modestly explained the association between maternal diabetes and risk of diabetes, whereas the effect was minimal for paternal transmission. The NHS included a sample of female nurses with limited variation in socio-economic status and with relatively healthy lifestyle behaviours. This selected sample may have led to an underestimation of the extent to which lifestyle factors explain parental transmission of diabetes .
The strong risk conferred by maternal or paternal diabetes was comparable after accounting for diet, lifestyle factors and adiposity. Of note, these factors differently explained maternal and paternal transmission of risk of diabetes. Our findings are consistent with previous evidence indicating a stronger magnitude of maternal transmission of obesity and its association with many lifestyle factors, compared with paternal transmission . A possible explanation for this is that the mother might have more influence on eating habits and other lifestyle behaviours whilst raising her children. Indeed, there may be more contact hours between mothers and children during childhood and in later life, and therefore, the mother’s lifestyle may be more of an example for her children than the father’s. It has been shown that if the mother has a history of diabetes during pregnancy, her child is less likely to follow certain healthy dietary recommendations . Similarly, those with a maternal history of diabetes may be more prone to have an unhealthy diet and lifestyle throughout their lifetime. Finally, there is evidence to suggest an effect of maternal nutrition and weight maintenance during pregnancy on infant birth weight [27, 28]. Birth weight could in turn influence future risk of chronic diseases such as type 2 diabetes [29, 30].
Diabetes is a polygenic disease in which multiple genetic and environmental components play roles throughout all stages of the disease [6, 31]. Beyond the genetic heritability, parental history of diabetes also carries environmental risk factors. In other words, it seems that parents and children share common lifestyle behaviours which will be continued throughout the children’s lifetime. These components, transmitted by maternal or paternal exposures, may explain the heterogeneity of diabetes transmission in different populations.
We conclude that both maternal and paternal history of diabetes are associated with an increased risk of developing type 2 diabetes, independent of diet, lifestyle and adiposity. The slight excess risk conferred by maternal compared to paternal diabetes is explained by a larger contribution to this association of age, sex, diet, lifestyle factors and adiposity.
Conflict of interest statement
No potential conflicts of interest relevant to this article were reported.
This work was supported by the Netherlands Heart Foundation, the Dutch Diabetes Research Foundation and the Dutch Kidney Foundation. This research was performed within the framework of the Center for Translational Molecular Medicine (CTMM; http://www.ctmm.nl); project PREDICCt (grant 01C-104-07). The EPIC-NL study was funded by ‘Europe against Cancer’ Programme of the European Commission (SANCO), the Dutch Ministry of Health, the Dutch Cancer Society, the Netherlands Organization for Health Research and Development (ZonMW) and World Cancer Research Fund (WCRF). We thank Statistics Netherlands and the PHARMO Institute for follow-up data on cancer, cardiovascular disease and vital status. None of the study sponsors had a role in the study design, data collection, analysis or interpretation, report writing or the decision to submit the report for publication.