Clustering of Risk Factors in Parents of Patients With Type 1 Diabetes and Nephropathy

All patients are part of the Finnish Diabetic Nephropathy (FinnDiane) study, which is a nationwide multicenter study that seeks to determine genetic and clinical risk factors for micro- and macrovascular complications in type 1 diabetes. For this study, 2,355 patients with information for either available parent and classifiable diabetic renal status by May 2005 were selected from the FinnDiane database. The study design is cross-sectional (all data and samples collected at a baseline visit) and includes patients from 70 centers in Finland. Of the patients, 51% were male, mean (±SD) age was 41 ± 11 years, and duration of diabetes was 28 ± 9 years. Information was available from 2,353 mothers and 2,323 fathers (4,676 parents total). The local ethics committees approved the study, and it was performed in accordance with the Declaration of Helsinki. Written informed consent was obtained from each patient.

Data on medication, cardiovascular status, and diabetes complications of the patients with type 1 diabetes were registered from a standardized questionnaire that was completed by the patient's attending physician. Type 1 diabetes was defined as an onset of diabetes at ≤35 years of age and permanent insulin treatment initiated within 1 year of diagnosis. Renal status was defined based on UAER from at least two of three overnight or 24-h urine collections (normal UAER <20 μg/min or <30 mg/24 h [n = 1,140], microalbuminuria 20–200 μg/min or 30–300 mg/24 h [n = 435], macroalbuminuria >200 μg/min or >300 mg/24 h [n = 527], or end-stage renal disease [n = 253]), as on dialysis (n = 64), or as having a transplanted kidney (n = 189). Patients with normal UAER were required to have a duration of diabetes >15 years in order to ensure normal renal status. Diabetic nephropathy was defined as macroalbuminuria or end-stage renal disease. Since only one-third of patients with microalbuminuria eventually will develop overt diabetic nephropathy (22), we aimed to keep the diabetic nephropathy group as homogenous as possible and chose to include patients with microalbuminuria in the group without diabetic nephropathy. Renal function was estimated by the Cockcroft-Gault formula, adjusted for body surface area (23). As a measure of insulin sensitivity, we used an equation for the estimated glucose disposal rate (24) modified for use with A1C instead of HbA₁ (25).

Anthropometric data (weight, height, and waist and hip circumferences) and blood pressure were collected by a trained nurse. Blood pressure was measured twice in the sitting position and a mean value calculated.

Fasting blood samples were drawn and analyzed for A1C and serum creatinine. A1C was determined by standardized assays at each center (normal range 4.0–6.0%) and serum creatinine assessed by enzymatic methods at a central laboratory. UAER was determined centrally from 24-h urine collections in all patients, except those with end-stage renal disease, using radioimmunoassay and, since November 2002, with immunoturbidimetry.

Parental information was obtained from diabetic patients by a standardized questionnaire. Parental age, history of diabetes, hypertension, myocardial infarction, and stroke were asked separately for each parent. Parental CVD was defined as history of myocardial infarction or stroke. If the parent had diabetes, the age at onset and mode of treatment was registered, and based on this information parental diabetes was classified as type 1 diabetes (age at onset <35 years and insulin treatment), type 2 diabetes (age at onset >50 years or treatment with oral hypoglycemic agents or diet), or nonclassifiable (no data available on age at onset or treatment). If the parent had died, the cause and time of death was registered. Parental cardiovascular mortality was defined as death from myocardial infarction, heart failure, ruptured aortic aneurysm, or cerebrovascular event.

Data on mortality were available from all of the 4,676 parents and, of them, 1,635 (35%) had died. Age at death was 69 ± 13 years for mothers and 64 ± 13 years for fathers.

To assess the association of different combinations of a parental history of hypertension, CVD, and diabetes with diabetic nephropathy in the offspring, we used a parental risk score. Each parent was given 1 point for each positive history of hypertension, CVD, and diabetes. If the parent had no history of those traits, the score was 0, but if the parent was positive for all three, the score was 3. Thus, the maximum parental score was 6 and minimum 0. Combinations of hypertension and CVD, as well as hypertension and diabetes, were also calculated, with a maximum score of 4. There were only a few patients with high scores, and these scores were therefore pooled.

A total of 1,370 (29%) of the 4,676 parents have participated in the FinnDiane study. They answered a standardized questionnaire regarding their medical history, meaning the information given by diabetic patients could be directly validated. Data were thus validated for 625 fathers and 745 mothers of 789 patients with type 1 diabetes. The time difference between the point when data were given by the diabetic patient and when the parent attended the study was 3.1 ± 2.3 years for mothers and 2.9 ± 2.5 years for fathers. In some cases, the parents had been diagnosed with hypertension, CVD, or diabetes during this period, and the data were corrected accordingly. The overall sensitivity of the parental data were 83% and specificity 98%. For maternal and paternal data, the respective sensitivities for hypertension were 84 and 82%, for myocardial infarction 75 and 91%, for stroke 63 and 42%, and for diabetes 91 and 87%.

Continuous variables were analyzed with ANOVA if normally distributed (mean ± SD) or Kruskal-Wallis if not normally distributed (presented as median with interquartile range). The statistical significance of difference in categorical variables between groups was tested with χ² test. To assess the trend in proportions over categories, χ² test for trend analysis was used (26). In 2 × 2 tables, a Mantel-Haenszel analysis was performed to adjust for age by dividing age into four groups. The odds ratios (ORs) for individual parental factors were calculated using logistic regression analyses adjusted for age. In multiple logistic regression analyses, diabetic nephropathy was used as the dependent variable and different familial risk factors as independent variables, adjusted for sex, duration of diabetes, and A1C. Results are presented as OR with 95% CI. Notably, in the multivariate analyses in Table 1, parental diabetes, CVD, and mortality were not included due to collinearity. Mortality was analyzed with Kaplan-Meier survival analysis, and statistical significance of difference was tested with log-rank test. In the analyses for cardiovascular mortality, only cases with known cause of death were included. All analyses were performed using SPSS version 12.0.1 (SPSS, Chicago, IL). A P value <0.05 was considered statistically significant.

We investigated parental factors associated with diabetic nephropathy in offspring in 4,676 parents of 2,355 type 1 diabetic patients, 780 with diabetic nephropathy (59% men) and 1,575 without diabetic nephropathy (48% men, P < 0.001). Patients with diabetic nephropathy, compared with patients without diabetic nephropathy, were older (42 ± 9 vs. 41 ± 12 years, P = 0.002), had an earlier onset of diabetes (12 ± 7 vs. 14 ± 8 years, P < 0.001), had a longer duration of diabetes (30 ± 8 vs. 27 ± 10 years, P < 0.001), had a similar BMI (25.2 ± 3.9 vs. 25.1 ± 3.3 kg/m², P = 0.594), had a higher systolic (147 ± 21 vs. 133 ± 16 mmHg, P < 0.001) diastolic (84 ± 11 vs. 79 ± 9 mmHg, P < 0.001) blood pressure, had a higher UAER (499 [range 171–1,274] vs. 11 [6–29], P < 0.001), had a higher A1C (8.9 ± 1.5 vs. 8.3 ± 1.3%, P < 0.001), had a lower estimated glucose disposal rate (4.0 ± 1.6 vs. 6.5 ± 2.4 mg · kg⁻¹ · min⁻¹, P < 0.001), and had a lower estimated creatinine clearance (60 ± 34 vs. 101 ± 28 ml/min per 1.73 m², P < 0.001). Patients with diabetic nephropathy also had more coronary heart disease (15 vs. 4%, P < 0.001), strokes (7 vs. 1%, P < 0.001), and amputations (11 vs. 1%, P < 0.001). In patients with diabetic nephropathy, both mothers (67 ± 10 vs. 65 ± 12 years, P = 0.001) and fathers (66 ± 1 vs. 64 ± 1 years, P = 0.005) were older than those without diabetic nephropathy.

Data on hypertension, CVD, and diabetes were available in 89% (92% alive/81% deceased), 87% (90/80), and 92% (93/90) of mothers and in 82% (89/74), 83% (88/78), and 94% (95/94) of fathers. The type of diabetes could be determined in 91% of mothers and in 85% of fathers.

The prevalence of parental hypertension, CVD, diabetes, and mortality is presented in Table 1, which also shows that parental, maternal, and paternal hypertension, as well as parental type 1 diabetes and parental mortality, were associated with diabetic nephropathy in a univariate logistic regression analysis adjusted for age. Data were further analyzed using two multivariate models, one for parental data and one for maternal and paternal data, adjusted for duration of diabetes, A1C, and sex. In these models, parental hypertension and type 1 diabetes, especially in the mothers, were associated with diabetic nephropathy.

If both parents had hypertension, the prevalence of diabetic nephropathy in the diabetic offspring was 41%, while it was 33% if only one of the parents had hypertension and 29% if none of the parents had hypertension (P = 0.001, χ² for trend). The prevalence of diabetic nephropathy in offspring was 42% if both, 34% if one, and 30% if none of the parents had CVD (P = 0.006, χ² for trend). The corresponding values regarding parental diabetes were 45, 35, and 32% (P = 0.075, χ² for trend). If both, compared with none, of the parents had hypertension, the adjusted OR for diabetic nephropathy was 1.56 (95% 1.13–2.15) (Table 2).

Based on parental hypertension–CVD score, the prevalence of diabetic nephropathy in the offspring was 28% in patients with a parental score of 0 points, 30% with a parental score of 1 point, 36% with a parental score of 2 points, and 42% with a parental score of 3 or 4 points (P < 0.001 for trend). According to the parental hypertension–diabetes score, the prevalence of diabetic nephropathy in the offspring was 29% (0 points), 32% (1 point), 35% (2 points), and 50% (3 or 4 points) (P < 0.001 for trend). According to the parental hypertension–CVD–diabetes score, the prevalence of diabetic nephropathy in offspring was 27% (0 points), 31% (1 point), 33% (2–3 points), and 50% (4–6 points) (P < 0.001 for trend). The ORs for the association of the different parental scores with diabetic nephropathy are presented in Table 2. After adjustment for duration of diabetes, sex, and A1C, the parental hypertension–CVD score was no longer associated with diabetic nephropathy.

Parental mortality and paternal mortality were more prevalent in patients with than without diabetic nephropathy. No difference was observed in maternal mortality or parental, maternal, and paternal CVD mortality (Table 1). For fathers to patients with diabetic nephropathy, Kaplan-Meier curves showed reduced overall survival (log-rank P = 0.04) and reduced cardiovascular survival (log-rank P = 0.03). No difference was observed in parental (P = 0.11) or maternal (P = 0.81) overall survival or in parental (P = 0.08) or maternal (P = 0.66) survival from cardiovascular death.

This study confirms previous observations that familial factors play a role in the development of diabetic nephropathy. The association with individual traits studied (familial hypertension, CVD, and diabetes) appeared weaker than previously reported. A marked familial clustering of hypertension, CVD, and diabetes was, however, associated with diabetic nephropathy in offspring with type 1 diabetes.

The question why only a proportion of patients with type 1 diabetes develop diabetic nephropathy is still unresolved. Diabetic nephropathy undoubtedly clusters in families, but also a number of closely related traits seem to occur more frequently in families of offspring with diabetic nephropathy (6,9,13,14). This suggests that hereditary factors play a role. However, despite several attempts to define the traits that have an impact on diabetic nephropathy, there is still no consensus. The reason may be that most studies have been either underpowered or have included an ethnic admixture. Therefore, the present study was performed in a large homogeneous population that enabled us to study maternal and paternal risk traits separately. This is important because disease transmission from mothers to offspring and from fathers to offspring may differ (27). Notably, because genes of diabetes and its complications may show variability across populations, findings in one global region may not necessarily apply to other regions.

Hypertension in the parents, especially in the mothers, was associated with diabetic nephropathy when the traits were assessed individually. Notably, although paternal hypertension was only weakly associated with diabetic nephropathy, the risk increased if both parents were hypertensive. This suggests that although maternal hypertension may be more strongly linked to diabetic nephropathy, paternal hypertension also plays a role. Very few studies have addressed this issue, but Hadjadj et al. (20) observed higher blood pressure in mothers of patients with diabetic nephropathy, although there was no difference in the prevalence of hypertension.

Maternal stroke showed a weak relationship with diabetic nephropathy, but it is of note that the prevalence of stroke was rather low, as was the sensitivity of the questionnaire to detect a history of stroke. The finding is, however, in agreement with data on hypertension, a strong risk factor for stroke (28).

We have previously reported a link between parental type 2 diabetes and development of diabetic nephropathy in offspring with type 1 diabetes (21). This is in line with findings from the EURODIAB study, in which the finding was isolated to female patients (8). In the present study, we found no difference in the prevalence of parental type 2 diabetes despite substantial power. However, there was a tendency toward an increasing risk of diabetic nephropathy as the load of parental diabetes increased (diabetes in neither, one, or both parents). Furthermore, the combination of both parental diabetes and parental hypertension, compared with each alone, was clearly associated with an increased risk of diabetic nephropathy in the offspring.

It is possible that the lack of a difference in the prevalence of type 2 diabetes in the present study can be due to a selection bias caused by the increased cardiovascular mortality observed in fathers of diabetic nephropathy patients. Furthermore, the use of patient-reported data, instead of parent-reported data or direct assessment of parental glucose metabolism, may dilute our results. Nevertheless, if the true association would be as strong as we previously reported in our smaller dataset (21), it would probably have been observed with the present sample size and methodology. Therefore, the link between parental diabetes and diabetic nephropathy is probably not as strong as we previously reported, which may explain some of the negative findings in smaller studies (13,14,17,20). However, the link between the cluster of parental hypertension, diabetes, and CVD on one hand and diabetic nephropathy on the other may indicate that a familial predisposition to all of these conditions may be more important than the diabetic state in isolation. In other words, the common denominator may be, for instance, insulin resistance instead of diabetes, the latter being only one of several consequences of the former.

A novel finding was the higher prevalence of parental type 1 diabetes among patients with diabetic nephropathy compared with those without (4.3 vs. 2.9%). This is in contrast to a small Swedish study with a rather high prevalence of type 1 diabetes in the parents but with no association to diabetic nephropathy (29). Notably, the association between type 1 diabetes in the parents and diabetic nephropathy in the offspring in the present study was confined to the mothers.

Parental CVD has been associated with diabetic nephropathy in a few small studies in which CVD was defined as a combined end point including morbidity and mortality (13,14), although most studies have failed to replicate this finding (9,16,17). When data for CVD were analyzed individually, the present study did not show an association between parental CVD and diabetic nephropathy. On the other hand, the sensitivity of the questionnaire for parental CVD was lower than that for hypertension and diabetes. Therefore, we cannot rule out that use of more sensitive measures could detect an association between parental CVD and diabetic nephropathy. Such a view is supported by the fact that the prevalence of diabetic nephropathy increased from 30, to 34, to 42% if none, one, or both of the parents had CVD, respectively.

The present study is the first to have sufficient power to assess maternal and paternal survival separately. Using Kaplan-Meier survival analysis, we observed an association between paternal but not maternal mortality and diabetic nephropathy. This is partly in accordance with studies that showed an increase in parental overall and CVD mortality among patients with diabetic nephropathy (15,18,21). Other studies that were not based on survival analysis did not show a difference in the prevalence of parental mortality (8,16,17). Importantly, the new observation is that reduced survival among parents to patients with diabetic nephropathy seems to result from an increase in overall and CVD mortality in the fathers.

Nevertheless, the present study has some limitations. The parental information was received from the diabetic patients and not from the parents directly or from their medical records. We were, however, able to validate the parental data in as many as 45% of the parents still alive, and the overall sensitivity of the data was as high as 83%, with a specificity of 98%. Another important aspect was that the parental data were not complete but did exceed 80% in all cases. Finally, the use of sensitive measures, such as 24-h blood pressure monitoring, oral glucose tolerance tests, and angiographies could enable detection of occult disease, but this should be counteracted by the large number of subjects assessed in the present study.

Altogether, the associations we find in this study are not as pronounced as those in some earlier, smaller studies that showed associations between diabetic nephropathy and different familial risk factors. The discrepancy might reflect an improvement in the treatment of patients with type 1 diabetes with regard to both glycemic control and blood pressure, notably the increased use of ACE inhibitors and angiotensin II receptor blockers. This could possibly postpone the development of diabetic nephropathy and lead to a misclassification of nephropathy status. Interestingly, in the Diabetes Control and Complications Trial, a more prominent familial clustering of diabetes complications was observed in the conventionally treated patients (A1C 9.2%) than in the intensively treated patients (A1C 7.3%), which would support a role for improved treatment (30). Notably, the mean A1C value of 8.5% in our population was rather high; therefore, that glycemic control alone would explain the observed discrepancy is unlikely.

In conclusion, diabetic nephropathy in type 1 diabetes is associated with a cluster of parental hypertension, CVD, and diabetes, as well as with paternal CVD mortality. Although maternal or paternal transmission of these traits may have different impact on the risk of diabetic nephropathy in the offspring with type 1 diabetes, it seems that the more the traits cluster in the families, the higher the risk for diabetic nephropathy.

This study was supported by grants from the Folkhälsan Research Foundation, Samfundet Folkhälsan, Wilhelm and Else Stockmann Foundation, Liv och Halsa Foundation, Finnish Medical Society (Finska Läkaresällskapet), Perklén Foundation, and European Commission (QLG2-CT-2001-01669).

The skilled technical assistance of laboratory technicians Tarja Vesisenaho, Sinikka Lindh, Anna Sandelin, Hannele Hilden, Helina Perttunen-Mio, and Virve Naatti is gratefully acknolwedged. We also acknowledge all the physicians and nurses at each center participating in the collection of patients (31 or online appendix [available at http://care.diabetesjournals.org]).