Effects of Duration and Age at Onset of Type 1 Diabetes on Preclinical Manifestations of Nephropathy

The International Diabetic Nephropathy Study is a prospective cohort study of 243 children and young adults with type 1 (insulin-dependent) diabetes. Subjects between the ages of 9 and 30 years at three study sites (Montreal, Minneapolis, and Paris) with a duration of type 1 diabetes of 2–20 years, normal blood pressure, and albumin excretion rate <100 μg/min (all but eight had rates <20 μg/min) were recruited and underwent a renal biopsy. A detailed description of the study methods is given elsewhere (24). The primary goal was to relate changes in glomerular morphometry (as they reflect the early, prealbuminuric phase of diabetic nephropathy) to a variety of potential etiologic determinants, including age at onset, disease duration, degree of glycemic control, blood pressure, and lipid abnormalities. A study of the relationships among these factors and of biopsy at study entry is the subject of a recent publication (25). In this article, we investigate the relationships among age at onset, pre- and postpubertal duration of disease, and morphometric glomerular abnormalities at baseline. Tanner staging was assessed at the time of the first study visit, usually within 3 months of the baseline biopsy, in 176 (72%) of the 243 study subjects.

The following three morphometric measures obtained from renal biopsy tissue were used as the principal outcomes in this investigation because of their demonstrated importance as precursor lesions for more overt clinical manifestations of diabetic nephropathy (26,27): the volume fraction of the glomerulus occupied by the mesangium [Vv(Mes/glom)], glomerular basement membrane (GBM) width, and the surface density of peripheral glomerular capillary basement membrane per glomerulus [Sv(PGBM/glom)]. Increases in relative mesangial volume reduce the glomerular surface available for filtration (18,19). The methods for obtaining the morphometric measurements are detailed in an earlier report (25). The sample size of the study was based on estimates of the expected change in Vv(Mes/glom) over the 5-year period between the baseline and 5-year biopsies (24).

To estimate the effect of duration on all morphometric measures, we stratified the study population into three subgroups based on biologic age at onset of diabetes: 1) the prepuberty group consisted of 184 subjects who were prepubertal at the time of enrollment and also, therefore, at the time of diabetes onset (n = 18) [males: both pubic hair and genital Tanner stages = 1, or (in the absence of Tanner staging) age <12 years; females: both pubic hair and breast Tanner stages = 1 or age <11 years] or reported diagnosis of diabetes before age 12 years in males or age 11 years in females (n = 166); 2) the puberty group consisted of 35 subjects who reported a diagnosis of diabetes at age 12–15 years in males or 11–14 years in females; and 3) the postpuberty group consisted of 24 subjects who were enrolled postpuberty (males: pubic hair and/or genital Tanner stage 5 or age >15 years; females: pubic hair and/or breast Tanner stage 5 or age >14 years) and whose diabetes was also diagnosed after age 15 years in males and after age 14 years in females.

For all analyses, estimations were performed using multiple regression, with the duration of diabetes as the principal independent variable. Because age and disease duration are highly collinear (and nearly completely so, when examined within subgroups defined by age at disease onset), regression models did not include age terms. The resulting measure of effect is expressed as the mean difference in each morphometric measure for every 1-year increase in disease duration. Several regression models were assessed for goodness of fit, including a single straight line, a quadratic, two straight lines with different slopes, and combinations of a straight line and a quadratic. In all instances, residual analyses were performed to assess the goodness of fit of the models. Multivariate models were used to adjust the effect estimates for sex, center, and baseline HbA_1c, diastolic blood pressure, height, and BMI. Because the estimates before and after adjustment were extremely similar, however, only the unadjusted estimates will be presented in the results.

Of the 243 subjects (118 males and 124 females) in the cohort, 184 were in the prepuberty group, 35 in the puberty group, and 24 in the postpuberty group. The baseline characteristics and morphometric outcomes are shown in Table 1 for the overall cohort and for each of the age-at-onset subgroups. The ages at onset and at biopsy increased across the three subgroups, as expected from their definition. The duration of diabetes at the time of biopsy was lower in the puberty group. Baseline HbA_1c values were similar in the three groups, while systolic and diastolic blood pressures were higher in the postpuberty group, reflecting their older age.

Vv(Mes/glom) increased with increasing duration of diabetes; a combination of a straight line and quadratic relationship provided a significantly (P < 0.0001) better overall fit of the data than a single straight line (Fig. 1). The inflection joint for the two model segments was at 14.2 years. This pattern was similar in the three age-at-onset subgroups, although few subjects in the puberty group had durations beyond the inflection point. These data suggest a nonlinear relationship between disease duration and Vv(Mes/glom) that is independent of age of disease onset but that begins to increase substantially only after 14 or 15 years duration. The values of the coefficients (and their 95% CIs) for both curve segments in the overall cohort and in each subgroup are shown in Table 2.

The results for GBM width are shown graphically in Fig. 2. GBM width increased with increasing disease duration; a single quadratic function with a constantly decreasing slope provided the best fit for these data (P = 0.033 vs. a straight line). Both the linear (duration) and quadratic (duration²) terms for disease duration were significant predictors of GBM width in the overall sample (Table 3). Results were similar in the three age-at-onset subgroups, although the CIs were wide in the puberty and postpuberty groups.

The results for Sv(PGBM/glom) are shown graphically in Fig. 3. As disease duration increased, Sv(PGBM/glom) decreased; a downward-curving quadratic provided the best fit (P = 0.002 vs. a straight line). The quantitative results are given in Table 3. Consistent with the results for Vv(Mes/glom) and GBM width, no difference was observed in the duration-Sv(PGBM/glom) relation according to age at onset of diabetes.

Previous studies have yielded inconsistent findings as to whether nephropathy in type 1 diabetes progresses more slowly before puberty than thereafter, but these studies have generally been limited to later stages of progression (i.e., after development of albuminuria) and have suffered from several important methodologic limitations. Our results suggest that it is total duration of disease, rather than duration before or after puberty, that affects the rate of progression of the GBM and mesangial lesions characteristic of diabetic nephropathy.

Two key features of our study have enabled us to clarify this issue and help to explain the conflicting evidence. First, our cohort consisted of young diabetic patients with albumin excretion rates <100 μg/min, and all but 8 of the 243 were normoalbuminuric (24), thus permitting us to study the effects of disease duration in young diabetic subjects before the clinical onset of nephropathy. Second, our morphometric analysis of renal biopsy specimens provides a quantitative assessment of glomerular histologic lesions that have previously been shown to be associated with clinical nephropathy (26,27). Third, we have controlled for such potentially confounding variables as sex, BMI, blood pressure, and glycemic control at the time of biopsy in order to isolate the effect of disease duration and its timing on the development of critical glomerular lesions. As a result, we have been able to show that the subclinical development of the principal glomerular lesion of diabetic nephropathy, i.e., the increase in Vv(Mes/glom), is nonlinear with respect to disease duration. Vv(Mes/glom) appears to progress slowly in the first 14 or 15 years of the disease, regardless of age (or pubertal stage) at onset. These results are consistent with those previously reported by Donaghue et al. (12) for microalbuminuria and Andersen et al. (2) for frank proteinuria. Similarly, Ellis et al. (22) did not find differences in Vv(Mes/glom) in diabetic patients who were prepubertal versus those who were pubertal or postpubertal at the time of biopsy, although none of the subjects in that study had a disease duration >12 years.

Our results help explain the conflicting findings of previous studies regarding the effects of pre-versus postpuberty disease duration. Because it is impossible to accumulate 14 or 15 years of disease duration before puberty, all of the disease duration beyond 14–15 years will be included in the postpubertal component. The postpubertal effect will thus appear larger, on average, than the prepubertal effect, i.e., the “protective effect” of prepubertal duration is an artifact of the shape of the duration-nephropathy relation.

The apparent increased effect of disease duration after 14–15 years, even if real, might represent a period effect. In other words, disease duration earlier than 14–15 years before study enrollment may reflect an earlier calendar period during which type 1 diabetes was less aggressively treated; therefore, poorer glycemic control typical of disease management in the more distant past (which is not reflected in our baseline HbA_1c measurements) would be confounded with longer disease duration. Subjects with shorter disease duration have been treated for most or all of their duration in the recent past, thus creating a period effect by falsely suggesting that the early years of disease had a lesser influence on progression of nephropathy.

The apparent nonlinearity in the rate of development of mesangial expansion, a critical lesion in diabetic nephropathy (26), may have important pathophysiologic implications. For example, Fioretto et al. (28) found that successful pancreas transplantation did not result in the reversal of established diabetic nephropathy lesions during the first 5 years of normoglycemia, whereas substantial reversal was seen after 10 years (29). The reason for this delay in reversal of lesions is unknown but could represent the same pathophysiologic processes involved in the delayed expression of early diabetic nephropathy lesions. Thus, lesions may develop only after renal extracellular matrix molecules become heavily glycosylated, after which they may induce a reduction in their local degradation (30,31). Conversely, lesions may become reversible only after long-lived glycosylated extracellular matrix molecules are replaced by nonglycolated molecules following re-establishment of normoglycemia.

Our study has several limitations. First, despite the overall sample size of 243 subjects with renal biopsies, the number of subjects is rather small in our puberty (n = 35) and postpuberty (n = 24) groups. The sample size of the International Diabetic Nephropathy Study was based on estimated changes in Vv(Mes/glom) over the 5-year period between the baseline and follow-up biopsies (24), not a comparison of baseline morphometric measures in the three age-at-onset groups. As shown by the wide CIs, the small sizes of the puberty and postpuberty groups provided low statistical power for detecting effects of disease duration. A second limitation is that our puberty group has few subjects with a duration beyond 14–15 years, again providing low power for detecting differential effects of prolonged disease duration in that group. Finally, Tanner staging was missing in 28% of the subjects, which may have led to misclassification of pubertal status at disease onset in a few subjects. For these reasons, and particularly in regard to sample size, the current findings should be interpreted with caution.

Members of our group have previously demonstrated that increase in mesangial fractional volume [Vv(Mes/glom)] is associated with the development of proteinuria, hypertension, and renal insufficiency (26,27). Moreover, the International Diabetic Nephropathy Study Group has shown elsewhere that previous duration of type 1 diabetes is strongly associated with Vv(Mes/glom) and other diabetic glomerular lesions at the time of the baseline biopsy (25). Microalbuminura has been reported in cohorts of children and adolescents with type 1 diabetes (15,16,19,32,33) and precedes the occurrence of frank proteinuria by ∼5 years (2). Nonetheless, we have no way of ensuring that the relationships we have found for histologic lesions of diabetic nephropathy will persist with respect to development of microalbuminuria, proteinuria, and renal insufficiency. A second renal biopsy obtained after 5 years of follow-up of our cohort will enable us to examine the progression of glomerular morphometric measures between the first and second biopsy. In this way, we will be able to assess the extent of further increases in Vv(Mes/glom) and other glomerular lesions with additional postpubertal years of type 1 diabetes.

Other members of the International Diabetic Nephropathy Study Group:

Christine Aebi, Mimi Belmonte, Robert Gardiner, Diane Laforte, Constantin Polychronakos, Alicia Schiffrin, and Atul Sharma, McGill University, Montreal, Canada; Khalil Khoury, Université de Sherbrooke, Sherbrooke, Canada; Jan Braaten and Kenneth Faught, University of Ottawa, Ottawa, Canada; Paul Czernichow, Université de Paris VII, Paris, France; Marie-Claire Gubler, Hôpital Necker-Enfants Malades, Paris, France; Philippe Passa, Hôpital Saint-Louis, Paris, France; Rebecca Carpenter, Blanche Chavers, Youngki Kim, Krishna Saxena, Joseph Sockalosky, Marty Spencer, Michael Steffes, and Robert Vernier, University of Minnesota, Minneapolis, MN.

This work was primarily supported by grants from the Juvenile Diabetes Foundation International. Clinical studies in Minneapolis were supported by a General Clinical Research Center (GCRC) grant (M01-RR00400) from the National Center for Research Resources and the National Institutes of Health. Support was also provided by the Emma Howe Foundation and the Robert Palmer Fund. M.S.K. and S.S. are recipients of Senior Investigator Awards from the Canadian Institutes of Health Research (CIHR).

John Basgen performed the EM morphometric measurements and was assisted in EM preparation by Tom Groppoli. Jean Bucksa headed the clinical research laboratory. Susan Kupcho performed the urinary albumin measurements. Joyce Stein helped with the study administration. Trudy Strand, Marlys Nolander, Patricia Moynihan, Vicky Siefert, and Michele Watrin performed the coordinator duties in Minnesota, while Brigitte Maruca assisted in this capacity in Montreal and Christine Delcroix, MD, and Dominique Simon, MD, performed the coordinator duties in Paris. Moira Mills coordinated the Montreal laboratory efforts. Hélène Beaufils and Veronique Beaudoin performed many of the clinical studies in Paris. We also thank the nurses at the General Clinical Research Center at Fairview-University Medical Center and at the International Diabetes Center in Minneapolis and Michelle Proulx at St. Paul Children’s Hospital for their excellent patient care. We are especially grateful to Raquel Meerovici who performed the data cleaning and entry.