To evaluate HbA1c followed from diagnosis, as a predictor of severe microvascular complications (i.e., proliferative diabetic retinopathy [PDR] and nephropathy [macroalbuminuria]).
In a population-based observational study, 447 patients diagnosed with type 1 diabetes before 35 years of age from 1983 to 1987 in southeast Sweden were followed from diagnosis until 2019. Long-term weighted mean HbA1c (wHbA1c) was calculated by integrating the area under all HbA1c values. Complications were analyzed in relation to wHbA1c categorized into five levels.
After 32 years, 9% had no retinopathy, 64% non-PDR, and 27% PDR, and 83% had no microalbuminuria, 9% microalbuminuria, and 8% macroalbuminuria. Patients with near-normal wHbA1c did not develop PDR or macroalbuminuria. The lowest wHbA1c values associated with development of PDR and nephropathy (macroalbuminuria) were 7.3% (56 mmol/mol) and 8.1% (65 mmol/mol), respectively. The prevalence of PDR and macroalbuminuria increased with increasing wHbA1c, being 74% and 44% in the highest category, wHbA1c >9.5% (>80 mmol/mol). In comparison with the follow-up done after 20–24 years’ duration, the prevalence of PDR had increased from 14 to 27% and macroalbuminuria from 4 to 8%, and both appeared at lower wHbA1c values.
wHbA1c followed from diagnosis is a very strong biomarker for PDR and nephropathy, the prevalence of both still increasing 32 years after diagnosis. To avoid PDR and macroalbuminuria in patients with type 1 diabetes, an HbA1c <7.0% (53 mmol/mol) and as normal as possible should be recommended when achievable without severe hypoglycemia and with good quality of life.
Introduction
The development of diabetic retinopathy and nephropathy is related to diabetes duration and hyperglycemia. The introduction of HbA1c as a biomarker for glycemic control in the early 1980s made it possible to assess the role of hyperglycemia for development of diabetic complications (1). The landmark study Diabetes Control and Complications Trial (DCCT), published in 1993, overwhelmingly showed that diabetic microangiopathy was related to glycemic control assessed as HbA1c (2). However, it is still unclear how tight glycemic control to strive for by means of HbA1c, and guidelines vary among diabetes care organizations (3–5). Too tight control may increase the risk of serious hypoglycemia (6), while less tight control will enhance diabetic microangiopathy. It should be noted that with the advances in diabetes care during the last two decades, the risk of severe hypoglycemia with improved HbA1c is weaker (7). A limitation with previous studies is that glycemic control was not followed from onset of diabetes (8,9). This initial period can account for metabolic memory or legacy effects possibly by glycosylation or epigenetic mechanisms (10,11). Beside diabetes duration and glycemic control, other factors such as age at onset and puberty may be related to development of severe retinopathy and nephropathy (12,13)
Since the early 1980s, we have followed a population-based type 1 diabetes cohort from diagnosis with HbA1c and have reported follow-up results after diabetes durations of 10 and 20 years (14–16). Up to a diabetes duration of 20 years, severe eye complications defined as panretinal laser-treated proliferative retinopathy or nephropathy defined as macroalbuminuria were not detected in patients having a long-term weighted mean HbA1c (wHbA1c) <7.6% (60 mmol/mol). Since the hazard ratio for diabetes complications in relation to HbA1c increases with duration (17), it is of interest to repeat this follow-up after another 10 years.
The aim of this study is to report the association of HbA1c followed 32–36 years from diabetes onset with development of severe retinopathy, defined as proliferative diabetic retinopathy (PDR) and nephropathy (macroalbuminuria) in patients with type 1 diabetes aged 0–34 years at diagnosis, and to formulate HbA1c target goals for avoiding these complications.
Research Design and Methods
Patients
An unselected population comprising all 447 patients diagnosed with type 1 diabetes before 35 years of age, 1 January 1983 to 31 December 1987, in the southeast hospital region in Sweden was studied. That all patients were included was validated with the help of the Swedish Childhood Diabetes Registry (18) and the Diabetes Incidence Study in Sweden (DISS) (19). In January 2019, a follow-up of the original patient cohort was made (i.e., at a diabetes duration of 32 to 36 years from diagnosis). Data were collected from medical records by a questionnaire to their treating physicians with questions about level of retinopathy at last eye examination, albuminuria at last visit, and treatment with ACE inhibitors, angiotensin II receptor blockers, and other antihypertensive drugs. In cases with PDR and/or macular edema, the retinopathy diagnosis was reevaluated by a specialist in ophthalmology. Data about 432 patients could be retrieved from the Swedish National Diabetes Registry (NDR) (20). Fourteen patients had their last clinical visit before the start of NDR 1996, and one patient had not been registered. Information was also retrieved from the Swedish Renal Registry (21), the Swedish National Patient Register, and the Swedish Cause of Death Register. When reviewed in this follow-up, 4 of 451 patients earlier reported (16) had characteristics of type 2 diabetes, high C-peptide or not insulin-dependent, and were therefore excluded. Patients who had moved abroad (n = 7) or were deceased (n = 55) were followed to their last visit.
For the 432 patients reported to NDR, data about blood pressure, lipids, and BMI at follow-up were available in the NDR database. These data are provided in Supplementary Table 1.
Insulin Treatment
In the 1980s, the basal bolus concept and insulin pumps were introduced in the treatment of type 1 diabetes in Sweden, and almost all newly diagnosed patients used intensive insulin regimens from the beginning. During the last decades, rapid and long-acting insulin analogs have been introduced. In the last 5 years, the number of patients using sensor-based glucose testing has increased to ∼80%. The method of insulin administration is registered in NDR, and 25% of the patients in the Vascular Diabetic Complications in Southeast Sweden (VISS) study used insulin pumps at the end of this follow-up.
The Research Ethics Committee of the Faculty of Health Sciences at Linköping University approved the study (Dnr 2017/394–31).
Retinopathy
Retinal screening using fundus photography was planned every other year from onset of diabetes or from 10 years of age for children. The commonly used technique was fundus photography with red-free digital images of one central and one nasal field per eye. The level of retinopathy and macular edema was classified according to the international clinical diabetic retinopathy and diabetic macular edema disease severity scale (22). In this severity scale, diabetic retinopathy is divided into mild, moderate, and severe levels of non-PDR (NPDR) and PDR. The indication for treatment with laser or intraocular injections was either PDR or diabetic macular edema. The date of the first treatment was collected from clinical records. Photographs or reliable data concerning previous therapy with laser or injections for PDR and/or maculopathy were available for 440 (98%) of the patients.
Nephropathy
The patients were screened for proteinuria at their regular clinical visits, at least once every year. The urine sample was analyzed with quantitative immunoturbidometric methods. Microalbuminuria was defined as two positive test results from three samples within 1 year with an albumin excretion rate 20–200 µg/min or albumin/creatinine ratio of 3–30 mg/mmol. Macroalbuminuria was defined as an albumin excretion rate >200 µg/min or albumin/creatinine ratio >30 mg/mmol. For all patients with macroalbuminuria, the medical records were scrutinized to confirm that there was no other kidney disease explaining the condition. Data were available for 441 (99%) of the patients. One patient had IgA nephropathy and one patient with myelomeningocele had hydronephrosis, and they were excluded from the renal evaluation. At the follow-up, 45% of the patients were treated with antihypertensive drugs, mainly blockers of the renin-angiotensin-aldosterone system.
The prevalence of nephropathy was examined at the last visit and grouped as normoalbuminuria, microalbuminuria, or macroalbuminuria. Eight patients who were normoalbuminuric but treated with ACE inhibitors because of previous microalbuminuria were classified as microalbuminuria. When calculating the incidence of macroalbuminuria, the first year when macroalbuminuria became persistent was indicated as onset.
For 425 patients, estimated glomerular filtration rate (eGFR; in mL/min/1.73 m2) data were available in NDR. eGFR in NDR was calculated according to the MDRD formula for adults: eGFR = 175 × (creatinine [µmol/L]/88.4)−1.154 × age−0.203 × 0.742 (if a woman). eGFR <60 mL/min/1.73 m2 was used as a cut point to indicate impaired eGFR.
HbA1c Measurement
HbA1c was measured regularly, two to four times per year. At the start of the study in January 1983, HbA1c was analyzed by Isolab minicolumns (Fast Hb Test System; Isolab Inc., Akron, OH) at the four central laboratories. This was replaced from 1984 to 1987 by high-performance liquid chromatography methods measuring HbA1c with high precision. The analyzing laboratories calculated intermethod calibrations and conversion factors when the methods were changed. Beginning in June 1994, hospital laboratories were participating in an interlaboratory quality program (Equalis, Uppsala, Sweden), in which all laboratories analyzed two samples per month. In 1997, a nationwide standardization was introduced in Sweden with a standardization scheme based on the Mono S method (23). Repeated comparisons were made with National Glycohemoglobin Standardization Program (NGSP) values, which showed the Swedish values to be 1.1% lower than NGSP values (24). The same was demonstrated in a study comparing HbA1c measured in 1994 at the Linkoping Hospital Laboratory with the DCCT laboratory (25). Since 2007, the HbA1c method is internationally standardized (26). All values in this report are converted by formulas to the new International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) reference values. The corresponding NGSP values are also stated, making it possible to compare the results with previous studies. The conversion formula is HbA1c (NGSP) (%) = 0.09153 HbA1c (IFCC) (mmol/mol) + 2.153. The normal range is 27–42 mmol/mol (IFCC) corresponding to 4.6–6.0% (NGSP). For many of the patients who moved, it has been possible to obtain their HbA1c values from their physicians and use conversion factors to the Equalis reference method done by the local laboratory.
Ninety percent of the HbA1c values came from laboratories in the catchment area and were, for this follow-up, collected directly from the central laboratories’ databases. As a measure of long-term glycemic control, all HbA1c values, from diabetes diagnosis until the date of diagnosis of PDR, the date for start of macroalbuminuria, or last follow-up time, were used for the calculations. In total, 36,550 HbA1c values were collected, 82 ± 27 (mean ± SD) values per patient and on average 2.6 ± 0.6 values per patient and year. We checked for gaps in the HbA1c series by making a pivot table. Gaps of up to 3 years were accepted. Six patients with complications had gaps in the HbA1c series between 4 and 9 years before the complication: one with PDR alone, four with both PDR and macroalbuminuria, and one with macroalbuminuria alone. All patients except one patient with wHbA1c 8.1% (65 mmol/mol) had wHbA1c values >8.6% (>70 mmol/mol). Since the values before and after the gaps were very consistent, we estimated that the missing data would not affect the results. Nine patients with HbA1c gaps ≥10 years were not included in the analysis of the impact of wHbA1c on complications. wHbA1c was calculated using the area under the graph of time versus HbA1c, divided by the total follow-up time for each patient, thereby compensating for the occasionally irregular intervals between measurements (14).
Statistical Analysis
For statistical analysis, wHbA1c was categorized into five levels. Differences between groups were tested with t test, ANOVA, or general linear model repeated measures with Bonferroni post hoc test. Life table analysis with Wilcoxon (Gehan) log-rank test was used for analysis of incidence of laser-treated proliferative retinopathy and macroalbuminuria.
Cox proportional hazard regression model was used to analyze the relative contribution of covariates to the risk of developing severe retinopathy or macroalbuminuria. Variables included in the model were age at diabetes diagnosis, sex, and categorized wHbA1c during the whole follow-up.
Data are shown as mean ± SD unless otherwise stated. The significance level was set as P < 0.05. SPSS version 28 was used for the analyses.
Results
Retinopathy
At follow-up in January 2019, diabetes duration was 32 (24, 34) (median [interquartile range]) years (range 6–36 years) to the last follow-up or to the year of first treatment with panretinal laser or injection therapy for PDR. Forty-one patients (9%) had no retinopathy, 281 (64%) had developed NPDR, and 118 (27%) had developed PDR. Fifty patients were treated with focal laser (n = 38) or injection therapy (n = 13) for macular edema; one of these patients received both treatments. Of the 50 patients treated for macular edema, 11 patients had NPDR and 39 had PDR. The prevalence of different severity levels of retinopathy in relation to categories of wHbA1c from diagnosis to the end of follow-up is shown in Table 1. Nine patients were not included due to lack of HbA1c data during several years, and for 9 patients, level of retinopathy was not available, leaving 422 patients with complete data (Table 1). In the lowest wHbA1c category, wHbA1c ≤6.7% (≤50 mmol/mol), none of the patients had developed PDR. In the wHbA1c category 6.8–7.6% (51–60 mmol/mol), five patients had developed PDR, with the lowest wHbA1c value being 7.3% (56 mmol/mol). The prevalence of PDR increased with increasing wHbA1c, being 74% in the highest category, wHbA1c >9.5% (>80 mmol/mol). The mean values for wHbA1c increased with increasing severity of retinopathy. The wHbA1c range for the different levels of retinopathy is illustrated in Supplementary Fig. 1A. The cumulative proportion of PDR started to increase after a diabetes duration of 6 years (Fig. 1A). Divided into wHbA1c categories, the cumulative proportion of PDR increased earlier and was higher with increasing wHbA1c levels (Fig. 1A). In all wHbA1c categories >6.7% (>50 mmol/mol), the cumulative proportion continued to increase up to at least 32 years’ diabetes duration.
Cumulative proportion of PDR (A) and macroalbuminuria (B) in an unselected population of patients with type 1 diabetes with different levels of long-term wHbA1c. P value denotes pairwise comparisons between the category ≤6.7% (≤50 mmol/mol) and higher categories (A) and between the category 6.8–7.6% (51–60 mmol/mol) and higher categories (B).
Cumulative proportion of PDR (A) and macroalbuminuria (B) in an unselected population of patients with type 1 diabetes with different levels of long-term wHbA1c. P value denotes pairwise comparisons between the category ≤6.7% (≤50 mmol/mol) and higher categories (A) and between the category 6.8–7.6% (51–60 mmol/mol) and higher categories (B).
Prevalence of microvascular complications in different HbA1c categories of long-term wHbA1c in the VISS cohort of patients with type 1 diabetes 32–36 years after diagnosis
. | wHbA1c categories NGSP % (IFCC mmol/mol) . | . | . | |||||
---|---|---|---|---|---|---|---|---|
≤6.7 (≤50) . | 6.8–7.6 (51–60) . | 7.7–8.6 (61–70) . | 8.7–9.5 (71–80) . | >9.5 (>80) . | All . | NGSP (%), mean ± SD . | IFCC (mmol/mol), mean ± SD . | |
None | 12/41.3 | 15/13.3 | 9/6.1 | 3/3.4 | 1/2.3 | 40/9.5 | 7.4 ± 1.1 | 57 ± 12 |
Mild NPDR | 17/58.6 | 56/49.6 | 53/35.6 | 8/9.2 | 1/2.3 | 135/31.5 | 7.6 ± 0.8 | 60 ± 9 |
Moderate NPDR | 0/0 | 33/29.2 | 42/28.2 | 24/27.6 | 5/11.4 | 104/22.0 | 8.2 ± 0.9 | 66 ± 10 |
Severe NPDR | 0/0 | 4/3.5 | 14/9.4 | 4/4.6 | 4/9.1 | 26/4.5 | 8.5 ± 0.9 | 69 ± 10 |
PDR | 0/0 | 5/4.4 | 31/20.8 | 48/55.2 | 33/75.0 | 117/32.5 | 9.1 ± 1.0 | 76 ± 11 |
All, n | 29 | 113 | 149 | 87 | 44 | 422 | 8.2 ± 1.1 | 66 ± 12 |
DME | 0/0 | 7/6.2 | 15/10.1 | 16/18.4 | 10/22.7 | 48/11.4 | 8.7 ± 0.9 | 72 ± 10 |
All, n | 29 | 113 | 149 | 87 | 44 | 422 | 8.2 ± 1.1 | 66 ± 12 |
Normoalbuminuria | 27/93.1 | 109/97.3 | 147/90.7 | 57/67.9 | 15/34.9 | 355/82.5 | 8.0 ± 0.9 | 64 ± 10 |
Microalbuminuria | 2/6.9 | 3/2.7 | 11/6.8 | 16/19.0 | 9/20.9 | 41/9.5 | 8.7 ± 1.0 | 72 ± 11 |
Macroalbuminuria | 0/0 | 0/0 | 5/3.0 | 10/11.9 | 19/44.2 | 34/7.9 | 9.7 ± 1.0 | 83 ± 11 |
All, n | 29 | 112 | 162 | 84 | 43 | 430 | 8.2 ± 1.1 | 66 ± 12 |
. | wHbA1c categories NGSP % (IFCC mmol/mol) . | . | . | |||||
---|---|---|---|---|---|---|---|---|
≤6.7 (≤50) . | 6.8–7.6 (51–60) . | 7.7–8.6 (61–70) . | 8.7–9.5 (71–80) . | >9.5 (>80) . | All . | NGSP (%), mean ± SD . | IFCC (mmol/mol), mean ± SD . | |
None | 12/41.3 | 15/13.3 | 9/6.1 | 3/3.4 | 1/2.3 | 40/9.5 | 7.4 ± 1.1 | 57 ± 12 |
Mild NPDR | 17/58.6 | 56/49.6 | 53/35.6 | 8/9.2 | 1/2.3 | 135/31.5 | 7.6 ± 0.8 | 60 ± 9 |
Moderate NPDR | 0/0 | 33/29.2 | 42/28.2 | 24/27.6 | 5/11.4 | 104/22.0 | 8.2 ± 0.9 | 66 ± 10 |
Severe NPDR | 0/0 | 4/3.5 | 14/9.4 | 4/4.6 | 4/9.1 | 26/4.5 | 8.5 ± 0.9 | 69 ± 10 |
PDR | 0/0 | 5/4.4 | 31/20.8 | 48/55.2 | 33/75.0 | 117/32.5 | 9.1 ± 1.0 | 76 ± 11 |
All, n | 29 | 113 | 149 | 87 | 44 | 422 | 8.2 ± 1.1 | 66 ± 12 |
DME | 0/0 | 7/6.2 | 15/10.1 | 16/18.4 | 10/22.7 | 48/11.4 | 8.7 ± 0.9 | 72 ± 10 |
All, n | 29 | 113 | 149 | 87 | 44 | 422 | 8.2 ± 1.1 | 66 ± 12 |
Normoalbuminuria | 27/93.1 | 109/97.3 | 147/90.7 | 57/67.9 | 15/34.9 | 355/82.5 | 8.0 ± 0.9 | 64 ± 10 |
Microalbuminuria | 2/6.9 | 3/2.7 | 11/6.8 | 16/19.0 | 9/20.9 | 41/9.5 | 8.7 ± 1.0 | 72 ± 11 |
Macroalbuminuria | 0/0 | 0/0 | 5/3.0 | 10/11.9 | 19/44.2 | 34/7.9 | 9.7 ± 1.0 | 83 ± 11 |
All, n | 29 | 112 | 162 | 84 | 43 | 430 | 8.2 ± 1.1 | 66 ± 12 |
Data are n/% unless otherwise indicated.
Differences in wHbA1c between different severity levels of retinopathy and DME (treated with focal laser or injection) in comparison with no retinopathy were highly significant (P < 0.001), except for the difference between no retinopathy and mild nonproliferative retinopathy (P = 1.0). Differences in wHbA1c between the different levels of albuminuria were highly significant (P < 0.001). Data were analyzed with one-way ANOVA and Bonferroni post hoc test. Only patients with complete wHbA1c data were included.
In a Cox proportional hazard regression model with sex, categorized wHbA1c, and age at onset, only wHbA1c was significant (Table 2, model A).
Cox regression models of risk factors for PDR model A and nephropathy (macroalbuminuria) model B
Model . | Risk factor . | P . | Hazard ratio . | 95% CI . |
---|---|---|---|---|
A, PDR | Sex | 0.791 | 0.951 | 0.656–1.379 |
Age at onset 0–4.9 years | 0.846 | 0.920 | 0.398–2.129 | |
Age at onset 5–9.9 years | 0.916 | 1.042 | 0.485–2.237 | |
Age at onset 10–14.9 years | 0.838 | 1.083 | 0.505–2.321 | |
Age at onset 15–19.9 years | 0.881 | 1.068 | 0.453–2.514 | |
Age at onset 20–24.9 years | 0.891 | 0.939 | 0.383–2.301 | |
Age at onset 25–29.9 years | 0.939 | 0.965 | 0.385–2.416 | |
Age at onset 30–34.9 years | Reference | |||
wHbA1c ≤6.7% (≤50 mmol/mol) | No event | |||
wHbA1c 6.8–7.6% (51–60 mmol/mol) | Reference | |||
wHbA1c 7.7–8.6% (61–70 mmol/mol) | 0.001 | 5.193 | 2.018–13.363 | |
wHbA1c 8.7–9.5% (71–80 mmol/mol) | 0.000 | 17.635 | 6.977–44.572 | |
wHbA1c >9.5% (>80 mmol/mol) | 0.000 | 38.133 | 14.779–98.392 | |
B, macroalbuminuria | Sex | 0.205 | 0.618 | 0.294–1.300 |
Age at onset 0–4.9 years | 0.043 | 0.274 | 0.078–0.961 | |
Age at onset 5–9.9 years | 0.001 | 0.115 | 0.030–0.433 | |
Age at onset 10–14.9 years | 0.017 | 0.241 | 0.075–0.775 | |
Age at onset 15–19.9 years | 0.096 | 0.332 | 0.091–1.216 | |
Age at onset 20–24.9 years | 0.945 | 1.044 | 0.309–3.524 | |
Age at onset 25–29.9 years | 0.091 | 0.272 | 0.060–1.231 | |
Age at onset 30–34.9 years | Reference | |||
wHbA1c ≤6.7% (≤50 mmol/mol) | No event | |||
wHbA1c 6.8–7.6% (51–60 mmol/mol) | No event | |||
wHbA1c 7.7–8.6% (61–70 mmol/mol) | Reference | |||
wHbA1c 8.7–9.5% (71–80 mmol/mol) | 0.001 | 7.275 | 2.254–23.486 | |
wHbA1c >9.5% (>80 mmol/mol) | 0.000 | 52.056 | 16.142–167.881 |
Model . | Risk factor . | P . | Hazard ratio . | 95% CI . |
---|---|---|---|---|
A, PDR | Sex | 0.791 | 0.951 | 0.656–1.379 |
Age at onset 0–4.9 years | 0.846 | 0.920 | 0.398–2.129 | |
Age at onset 5–9.9 years | 0.916 | 1.042 | 0.485–2.237 | |
Age at onset 10–14.9 years | 0.838 | 1.083 | 0.505–2.321 | |
Age at onset 15–19.9 years | 0.881 | 1.068 | 0.453–2.514 | |
Age at onset 20–24.9 years | 0.891 | 0.939 | 0.383–2.301 | |
Age at onset 25–29.9 years | 0.939 | 0.965 | 0.385–2.416 | |
Age at onset 30–34.9 years | Reference | |||
wHbA1c ≤6.7% (≤50 mmol/mol) | No event | |||
wHbA1c 6.8–7.6% (51–60 mmol/mol) | Reference | |||
wHbA1c 7.7–8.6% (61–70 mmol/mol) | 0.001 | 5.193 | 2.018–13.363 | |
wHbA1c 8.7–9.5% (71–80 mmol/mol) | 0.000 | 17.635 | 6.977–44.572 | |
wHbA1c >9.5% (>80 mmol/mol) | 0.000 | 38.133 | 14.779–98.392 | |
B, macroalbuminuria | Sex | 0.205 | 0.618 | 0.294–1.300 |
Age at onset 0–4.9 years | 0.043 | 0.274 | 0.078–0.961 | |
Age at onset 5–9.9 years | 0.001 | 0.115 | 0.030–0.433 | |
Age at onset 10–14.9 years | 0.017 | 0.241 | 0.075–0.775 | |
Age at onset 15–19.9 years | 0.096 | 0.332 | 0.091–1.216 | |
Age at onset 20–24.9 years | 0.945 | 1.044 | 0.309–3.524 | |
Age at onset 25–29.9 years | 0.091 | 0.272 | 0.060–1.231 | |
Age at onset 30–34.9 years | Reference | |||
wHbA1c ≤6.7% (≤50 mmol/mol) | No event | |||
wHbA1c 6.8–7.6% (51–60 mmol/mol) | No event | |||
wHbA1c 7.7–8.6% (61–70 mmol/mol) | Reference | |||
wHbA1c 8.7–9.5% (71–80 mmol/mol) | 0.001 | 7.275 | 2.254–23.486 | |
wHbA1c >9.5% (>80 mmol/mol) | 0.000 | 52.056 | 16.142–167.881 |
Nephropathy
Of the 439 patients evaluated for nephropathy, 355 (83%) had normoalbuminuria, 42 (9%) microalbuminuria, and 34 (8%) macroalbuminuria. Due to lack of HbA1c data during several years, 9 patients were not included in the evaluation of nephropathy in relation to wHbA1c, leaving 430 patients with a diabetes duration of 33 (31, 35) (median [interquartile range]) (range 8–36 years) to development of macroalbuminuria or to the last follow-up (Table 1). While cases with microalbuminuria were found in all wHbA1c categories, macroalbuminuria was only found in the categories with HbA1c >7.6% (>60 mmol/mol), starting at a wHbA1c of 8.1% (65 mmol/mol). There were highly significant differences in wHbA1c (P < 0.001) between the different levels of albuminuria (Table 1). The wHbA1c range for the different levels of albuminuria is illustrated in Supplementary Fig. 1B. The cumulative proportion of macroalbuminuria started to increase after a diabetes duration of 10 years and then increased for up to at least 32 years (Fig. 1B). There were marked differences between the wHbA1c categories, with the cumulative proportion in the highest category being twice as high as the next category (P < 0.001) (Fig. 1B). Ten of the patients with macroalbuminuria had developed end-stage renal failure, and eight of them were deceased.
For 432 patients reported to NDR, data about eGFR and albuminuria were available for 425 patients (Supplementary Table 2). In total, 39 patients had eGFR <60 mL/min/1.73 m2, 24 of them had macroalbuminuria, 7 microalbuminuria, and 8 no albuminuria. Four patients had HbA1c <7.7% (<61 mmol/mol) but no albuminuria. In comparison with the patients with eGFR ≥60 mL/min/1.73 m2, the patients with eGFR <60 mL/min/1.73 m2 were much more prevalent at high HbA1c categories, as shown in Supplementary Table 2 (P < 0.001 tested by χ2).
In a Cox proportional hazard regression model including sex, categorized wHbA1c, and categorized age at onset, wHbA1c and age at onset were significantly associated with macroalbuminuria (Table 2, model B).
PDR and Macroalbuminuria 32 Years After Diagnosis Compared With After 20 Years
In Figure 2A and B, the proportion of patients with PDR and nephropathy (macroalbuminuria) in different wHbA1c categories in this follow-up is compared with our result from the same cohort after 20–24 years’ duration (16). In the wHbA1c category ≤6.7% (≤50 mmol/mol), no one developed PDR, while in the wHbA1c categories 6.8–7.6% (51–60 mmol/mol) and higher, the proportion of PDR increased markedly with increased duration. No one developed macroalbuminuria in the two lowest wHbA1c categories, and in the wHbA1c category 7.7–8.6% (61–70 mmol/mol) and higher, the proportion of macroalbuminuria increased with increased duration. The cumulative proportion of macroalbuminuria was much lower than the proportion of PDR. The mean wHbA1c after 32–36 years was 8.2% (66 mmol/mol) (i.e., the same level as after 20–24 years). wHbA1c for patients developing PDR within 20–24 years after diagnosis was 9.5 ± 1.1% (80 ± 12 mmol/mol) and 8.7 ± 0.7% (72 ± 8 mmol/mol) for those developing PDR later (P < 0.001). The corresponding figures for macroalbuminuria were 10.0 ± 1.2% (86 ± 12 mmol/mol) and 9.4 ± 0.8% (79 ± 9 mmol/mol) (P = 0.065).
Severe retinopathy and nephropathy 32–36 years after diabetes diagnosis compared with 20–24 years after diabetes diagnosis. For retinopathy, the proportion of PDR in different wHbA1c categories is shown (A) and for nephropathy, the proportion of macroalbuminuria (B).
Severe retinopathy and nephropathy 32–36 years after diabetes diagnosis compared with 20–24 years after diabetes diagnosis. For retinopathy, the proportion of PDR in different wHbA1c categories is shown (A) and for nephropathy, the proportion of macroalbuminuria (B).
Conclusions
The major risk factors for development of diabetic microangiopathy are diabetes duration and hyperglycemia. In this observational population-based study, PDR had developed in 118 patients (i.e., 27% of the cohort) when followed up after 32–36 years, while after 20–24 years, the prevalence was 14% (16). Also, the prevalence of severe nephropathy, defined as macroalbuminuria, had increased from 4% after 20–24 years to 8% after 32–36 years. The mean wHbA1c had not increased and was 8.2% (66 mmol/mol) after both the 20–24- and 32–36-year diabetes durations.
The cumulative proportion of both PDR and macroalbuminuria started to increase after a diabetes duration of ∼10 years and then increased for up to at least 32 years, but with a much steeper curve for PDR. In comparison with previous observational studies, the cumulative proportion of PDR and nephropathy after 30 years was low compared with 50% and 30%, respectively, after 25 years in the Pittsburgh study (27). During the last few decades, there has been a decrease in the cumulative incidence of diabetic nephropathy in type 1 diabetes, probably related to improved glycemic control, as shown in several studies in Denmark and the Linköping Diabetes Complication study in Sweden (28,29). In the DCCT/Epidemiology of Diabetes Interventions and Complications (EDIC) study, which started at the same time as the VISS study, the prevalence of PDR and nephropathy was 30.7% and 7.9%, respectively, in the conventional group after 30 years’ diabetes duration (i.e., of similar magnitude as the overall prevalence of PDR [27%] and nephropathy [8%] after 30 years in VISS) (30). When wHbA1c was categorized into different levels, the prevalence of PDR and nephropathy was strongly associated with the level of wHbA1c, in agreement with differences between the conventionally and intensively treated groups in DCCT/EDIC (30). In the patients treated for diabetic macular edema (DME), this complication was associated with glycemic control as previously described (31).
Thirty-nine patients had impaired eGFR (eGFR <60 mL/min/1.73 m2), and this was related to high wHbA1c in line with the results of the DCCT/EDIC trial (32). Four patients had eGFR <60 mL/min/1.73 m2 and wHbA1c <7.7% (<61 mmol/mol), and none of these patients had albuminuria, suggesting other etiology than diabetic nephropathy.
With extended follow-up from 20–24 to 32–36 years, there was not only an increase in the number of patients who developed severe retinopathy and nephropathy, but also the wHbA1c threshold for severe retinopathy and nephropathy was lowered (16). In the wHbA1c category 6.8–7.6% (51–60 mmol/mol), there were five cases with PDR compared with one at the 20–24-year follow-up (16). The lowest value for appearance of PDR decreased from 7.6% (60 mmol/mol) to 7.3% (56 mmol/mol) and from 8.4% (68 mmol/mol) to 8.1% (65 mmol/mol) for macroalbuminuria. There may be some uncertainty about these limits due to lack of standardization of the HbA1c method during the first years of measurements, but our results and other studies suggest that there is a lowering of the HbA1c threshold for severe microvascular complications with increased duration (17).
A possible explanation for the lowered threshold for development of severe microangiopathy is the increase in “glycemic burden” with diabetes duration (11). The mechanisms by which hyperglycemia causes diabetic retinopathy and nephropathy are not clear. In this and other studies, there is a very strong association between HbA1c and development of severe microangiopathy (30). One possible explanation for this is that HbA1c reflects a pathophysiologic process of hyperglycemia, as it mirrors glycation of proteins (33). It should be noted that the relation between wHbA1c level and microangiopathy was not uniform. Some patients had high wHbA1c without having severe complications, but none had near-normal wHbA1c and severe microangiopathy. This suggests that the effect of hyperglycemia is modulated by other factors. Studies of long-standing of type 1 diabetes have identified putative protective mechanisms for diabetic complications (34). Beside diabetes duration and glycemic control age at onset, blood pressure, blood lipids, BMI, and genetic factors may be associated with development of microangiopathy (31,32,35). There is most probably also a variation in the glycation process between individuals that is illustrated by a rather large interindividual difference in level of HbA1c in relation to glucose measured by continuous glucose monitoring (36).
In a Cox proportional regression model, development of nephropathy was associated with age at onset, suggesting that patients with onset before puberty had a delayed onset of nephropathy. This is in line with the observation that young age at onset of diabetes prolongs the time until development of end-stage renal disease (37). Regarding PDR, we found no association with age of onset, in contrast to the results of the Finnish Diabetic Nephropathy (FinnDiane) Study (12).
The strength of our study is that we have prospectively followed HbA1c in an unselected population with type 1 diabetes in routine care with complete follow-up 32–36 years from diabetes onset. To our knowledge, there is no other study with such a long follow-up of HbA1c followed from diagnosis of type 1 diabetes. The HbA1c methods were, even in the first years, of high precision and standardized against national standards and also compared with NGSP values (24,25). It is therefore possible to compare our results with the HbA1c values of the DCCT study. An important limitation of our study is that it has not been possible to systematically register hypoglycemia. Our conclusions about HbA1c goals for treatment are therefore restricted to severe microangiopathy. Another limitation is that the possible effects of blood pressure, blood lipids, and BMI were not evaluated due to lack of longitudinal data for the whole cohort.
Guidelines from diabetes care organizations recommend different HbA1c targets, mostly between 6.5 and 8% (38,39). The American Diabetes Association and International Society for Pediatric and Adolescent Diabetes recommend HbA1c <7% (53 mmol/mol) as a treatment goal (4,5), while the National Institute for Health and Care Excellence recommends to aim for a target HbA1c level of 6.5% (48 mmol/mol) or lower in both children and adults with type 1 diabetes (3). Our data from the VISS study showed that up to a diabetes duration of 30 years, there was no case with severe retinopathy below a wHbA1c of 7.3% (56 mmol/mol) or macroalbuminuria below wHbA1c of 8.1% (65 mmol/mol). However, the threshold for severe microvascular complication tended to decrease with duration. The analytical performance specification (APS2010) for HbA1c in Sweden corresponds to a total allowable error of 0.33% NGSP or 3.6 mmol/mol at the HbA1c level of 6.7% (50 mmol/mol) (40). Adding this marginal to our lowest observed wHbA1c value for development of PDR in our study gives values identical to the American Diabetes Association’s goal for treatment of type 1 diabetes, <7.0% (53 mmol/mol). The observation that the threshold for complications probably decreases with increased duration makes it impossible to predict a safe HbA1c goal for avoiding PDR and macroalbuminuria for the coming decades. A reasonable statement from our data is to keep HbA1c <7.0% (53 mmol/mol) and as normal as possible if achievable without risk of severe hypoglycemia.
In conclusion, wHbA1c followed from diagnosis is a very strong biomarker for PDR and nephropathy, the prevalence of both still increasing 32 years after diagnosis. To avoid PDR and macroalbuminuria in patients with type 1 diabetes, HbA1c <7.0% (53 mmol/mol) and as normal as possible should be recommended when achievable without severe hypoglycemia and with good quality of life.
This article contains supplementary material online at https://doi.org/10.2337/figshare.20427942.
Article Information
Acknowledgments. The authors thank the colleagues in the VISS Study Group, southeast hospital region in Sweden for the help with medical data and hospital records: Karen Wahlin, Värnamo; Johan Blomgren, Eksjö; Oskar Lindholm, Jönköping; Marika Berg, Västervik; Edwin van Asseldonk, Oskarshamn; Herbert Krol, Kalmar; Anna-Maria Ottosson, Norrköping; Ulf Rosenqvist, Motala; and Christina Hedman, Linköping. The authors also thank colleagues all over Sweden who helped with medical data about patients who had moved out of the southeast hospital region.
Funding. This study was supported by Barndiabetesfonden (The Swedish Child Diabetes Foundation) and Region Östergötlands Stiftelsefonder (RÖ-760091).
Duality of Interest. No potential conflicts of interest relevant to this article were reported.
Author Contributions. H.J.A. and M.N. designed the study, performed the literature research and statistical analysis, and interpreted data. H.J.A. wrote the first draft of the manuscript. M.F. performed statistical analysis and interpreted data. J.L. designed the study, did literature research, and interpreted data. M.C.W. designed the study and evaluated the severity level of retinopathy. All authors reviewed and approved the final version of the report. H.J.A. is the guarantor of this work and, as such, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.