The NIDDK Takes on the Complications of Type 1 Diabetes: The Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) Study Free

The introduction of insulin therapy in the 1920s (1) transformed juvenile-onset diabetes, now called type 1 diabetes (T1D), from a uniformly fatal disease to a chronic degenerative one. During the first 50 years of the “insulin era,” 47% of people with T1D and a diabetes duration of >40 years had vision loss, 22% had impaired kidney function, 12% had amputations, and 10% and 12% had suffered a stroke or myocardial infarction (MI), respectively. Approximately 60% died by age 50 years (2).

In this setting and at the direction of the National Diabetes Commission and an act of Congress (3), the National Institute of Arthritis, Diabetes, and Digestive and Kidney Diseases (the forerunner of the National Institute of Diabetes and Digestive and Kidney Diseases [NIDDK]) initiated the planning of its first major clinical trial since the University Group Diabetes Program (4). Led by Dr. Oscar Crofford, planning for the Diabetes Control and Complications Trial (DCCT) occurred between 1982 and 1983 with 21 competitively selected investigators and the Coordinating Center (5) (Fig. 1). After the successful completion of a 2-year feasibility phase (6), the full-scale clinical trial began with the addition of 8 clinical sites to the original 21. Recruitment of the 1,441-member cohort was completed in early 1989, and the trial was completed in 1993. The main results, published in the most widely cited article in diabetes to date (7), demonstrated consistent salutary effects of intensive therapy, aimed at achieving near-normal glucose levels, compared with conventional therapy, on microvascular and neuropathic complications. The DCCT findings have been the basis of modern-day therapy for people with T1D ever since (8,9).

With an extraordinarily loyal group of participants (99% completed the mean 6.5 years [range 3–10] of DCCT) and a talented, dedicated group of investigators, study coordinators, and other staff, with a stable Data Coordinating Center, central laboratory, and reading centers, and with the addition of a Clinical Coordinating Center, an observational long-term follow-up of the DCCT cohort was proposed. The Epidemiology of Diabetes Interventions and Complications (EDIC) study was initiated in 1994 (10), with 96% of the DCCT cohort joining. EDIC continues to the current day (Fig. 1). EDIC has included exploration of the longer-term effects of the original intensive therapy compared with the original conventional therapy on more severe microvascular outcomes as well as on cardiovascular and a host of other diabetes-associated morbidities and on mortality. NIDDK has played an active role throughout in what has been a true partnership between the investigators and the institute. DCCT and EDIC have fulfilled the NIDDK’s diabetes mission “to improve people’s health and quality of life.”

In this article, on the occasion of the NIDDK 75th anniversary, we review the major results of DCCT and its long-term EDIC follow-up and examine the public health impact of DCCT/EDIC, including advances in T1D therapy that evolved in its wake.

The DCCT was designed as a randomized controlled clinical trial to test the glucose hypothesis, specifically, whether achieving lower glycemia would delay the development and slow the progression of long-term complications. For testing of the hypothesis, eligible volunteers (aged 13–39 years with diagnosed T1D duration 1–15 years and either no detectable microvascular complications at baseline in the primary prevention cohort or minimal but no severe complications in the secondary intervention cohort) (Table 1) were randomly assigned to an intensive (INT) or conventional (CONV) therapy regimen. In recognition of the demands of intensive therapy and the need for a highly motivated study cohort for this long-term study, a multicomponent informed consent procedure with emphasis on the need for random assignment and long-term adherence that included a quiz to validate participant understanding was implemented (11). Therapy in the INT group was aimed at achieving glycemic levels as close to the nondiabetic range as safely possible, while for CONV the standard treatment at the time was emulated (5) (Fig. 2). The HbA_1c goal of <6.05% (the mean HbA_1c ± 2 SD in a populaton without diabetes of similar age) was achieved by the majority of INT participants at least once; mean HbA_1c was ∼7% in the INT group vs. 9% in the CONV group during the DCCT (range 3–10 and mean 6.5 years) (7) (Fig. 2).

The primary outcome was retinopathy assessed with seven-field fundus photography every 6 months using the Early Treatment Diabetic Retinopathy Study (ETDRS) grading scale (12). Kidney disease was assessed with the measurement of albuminuria and creatinine clearance in timed urine collections. Peripheral neuropathy was determined with a standardized physical examination by a certified neurologist and required confirmation with electrophysiologic or abnormal autonomic test results.

The DCCT, ended by the independent monitoring board 1 year ahead of schedule owing to demonstrated efficacy, revealed consistent, significant 35%–76% reductions in the development and progression of all of the microvascular complications with INT versus CONV therapy (7). The reduction was nominally greater for the prevention of complications than for progression. Of note, severe microvascular complications and cardiovascular disease (CVD) were very uncommon in both treatment groups over the mean follow-up of 6.5 years. The adverse effects associated with INT therapy included a threefold increase in severe hypoglycemia, defined as episodes requiring assistance to treat, and weight gain (13).

Subsequent analyses revealed the central role of glycemia in the development and progression of complications with the difference in HbA_1c between INT and CONV therapy accounting for virtually all of the statistically significant differences in outcomes (14). Other major findings included similar beneficial effects of INT therapy in the adolescent subgroup (age <18 years, n = 195) (15) and similar neuropsychological outcomes in INT and CONV groups despite the higher frequency of severe hypoglycemia in the former (16). INT therapy also preserved residual β-cell function in participants with some endogenous insulin secretion at baseline (n = 303) for ∼2 years longer than CONV therapy, which translated into lower HbA_1c values with fewer episodes of severe hypoglycemia and fewer complications (17). INT therapy improved quality of life (18), and when the effects of INT therapy were applied across the estimated 120,000 people with T1D in the U.S. with characteristics similar to those of the DCCT cohort, INT therapy was projected to result in a gain of 920,000 years of sight, >690,000 years free from end-stage kidney disease, and >600,000 years of added life (19).

At the end of the DCCT, CONV participants were offered and instructed in intensive therapy, while INT participants were encouraged to continue it. All participants returned to community care and were followed annually in EDIC. The ∼2% separation in HbA_1c between the original treatment groups narrowed during EDIC and was no longer apparent by EDIC year 4 (Fig. 2).

At the beginning of EDIC, questions included whether the risk reductions in complications afforded by INT therapy would persist and whether longer-term, severe complications would be affected by the original therapy group assignment. During 4 years of EDIC follow-up, despite the convergence of HbA_1c levels, the cumulative incidence of retinopathy and kidney disease continued to separate between the original therapy groups, a phenomenon termed “metabolic memory” (20). Evidence for “metabolic amnesia,” the waning of metabolic memory, began to appear after ∼10 years of EDIC follow-up (21). Of note, while the adolescent subcohort had higher HbA_1c levels during DCCT, INT participants achieved a nearly 2% separation in HbA_1c versus CONV participants. While metabolic memory was demonstrated through EDIC year 4 in the adolescents, it waned more quickly than in the adults (22).

There are many putative mechanisms that may mediate the effects of hyperglycemia on complications. These include glycation, lipoxidation, inflammation, apoptosis, and other intracellular processes. Some of these genetic/epigenetic mechanisms could explain metabolic memory. Using stored DCCT/EDIC biological samples, we explored whether advanced glycation end products (AGEs), nonenzymatically formed and with long half-lives, might mediate metabolic memory (23). Measurement of circulating plasma AGEs, three times during DCCT and EDIC, demonstrated their independent associations beyond HbA_1c with the progression of microvascular disease (24). Additionally, epigenetic studies revealed enrichment of acetylation of H2K9Ac associated with mean HbA_1c and differentially methylated loci linked to thioredoxin-interacting protein and the nuclear factor-κB inflammatory pathway, each of which could contribute to metabolic memory (25–27).

Additional analyses were conducted to investigate the impact of different glycemic patterns over time on the development of microvascular disease and metabolic memory (28). Based on the original INT and CONV separation in HbA_1c during the DCCT, individuals who maintained HbA_1c of ∼7% over the following 20 years of EDIC had a 70% reduction in the cumulative incidence of impaired estimated glomerular filtration rate (eGFR) in comparison with individuals who maintained HbA_1c of 9%. For participants with HbA_1c of ∼9% during DCCT and the first 10 years of EDIC and who subsequently achieved HbA_1c of 7% for the subsequent 10 years of EDIC, a 20% reduction in the cumulative incidence of impaired eGFR was still possible. Thus, while early initiation of INT therapy was most effective in slowing the development and progression of microvascular disease, later achievement of target HbA_1c was also beneficial.

Finally, the increased rate of hypoglycemia observed among INT participants during the DCCT declined over time, with rates of severe hypoglycemia during EDIC equivalent among former DCCT treatment groups (29). Recurrent severe hypoglycemia was substantially lower in participants who retained β-cell function for >30 years (i.e., C-peptide >0.03 nmol/L); however, unlike in earlier assessments in DCCT/EDIC, long-term preservation of residual β-cell function did not result in fewer microvascular complications (29).

By 2024, the participants of DCCT/EDIC had been studied for 65% of their estimated lifespans and 95% of their diabetes durations, allowing exploration of long-term diabetes in the setting of advancing age. In addition to the effects of INT versus CONV therapy on relatively early-stage complications during DCCT, more severe complications were similarly reduced by prior INT versus CONV therapy during EDIC, as described below, due at least in part to metabolic memory (30–34).

After >30 years of follow-up, 65% of participants remained free of proliferative diabetic retinopathy and 60% free of clinically significant macular edema (an earlier diagnostic category now termed diabetic macular edema), and 70% had not required ocular surgery, with significantly fewer INT than CONV affected (30). For proliferative diabetic retinopathy and diabetic macular edema, the predominant risk factor was HbA_1c and nonmodifiable risk factors included duration of T1D and age (30). With regard to ocular surgeries, beyond HbA_1c, only systolic blood pressure was a modifiable risk factor.

Retinal fundus photography was performed every 6 months during DCCT and every 4 years during EDIC. This database allowed study of the transitions among retinopathy stages, progression and regression, and provided evidence necessary to personalize the frequency of assessments for retinopathy (35). Algorithms to optimize screening frequency resulted in a reduction in the screening burden for individuals with low levels of retinopathy and HbA_1c in the target range. Conversely, individuals with more severe retinopathy and high HbA_1c levels needed more frequent surveillance. Implementation of this individualized screening was projected to have large economic savings. In the DCCT adolescent cohort age <18 years, transition to advanced retinopathy requiring treatment was rare (36), supporting the recommendation that a single normal initial retinal exam for adolescents, coupled with intensive therapy, would suffice until 18 years of age.

The beneficial effects of INT versus CONV therapy on kidney disease included mean risk reductions of 50% and 20% for development of impaired eGFR and hypertension, respectively, with 59% and 84% risk reductions for moderate and severe albuminuria (37). The development of severe albuminuria and reductions in eGFR were both driven primarily by mean HbA_1c; however, triglyceride levels were also associated with eGFR over time. We also established evidence-based, personalized frequency for albuminuria screening based on prior albuminuria and HbA_1c (38). Screening intervals ranged from every 6 months for individuals at high risk (albumin excretion rate [AER] 21–30 mg/24 h or HbA_1c ≥9%) to every 2 years for those at low risk (AER ≤10 mg/24 h and HbA_1c ≤8%). Annual screening recommendations remained for those at moderate risk.

Of participants, 33% developed diabetic peripheral neuropathy (DPN) and 44% developed cardiovascular autonomic neuropathy (CAN) over 23 years of follow-up (39). As with other microvascular complications, mean HbA_1c was the most important modifiable risk factor for DPN and CAN; systolic blood pressure and cigarette smoking were additional modifiable risk factors for CAN. Foot ulcers were significantly reduced by ∼23% in the INT group, with a similar magnitude, albeit nonsignificant, reduction in amputations. DCCT/EDIC has also contributed novel clinical assessment tools for CAN. For example, indices of heart rate variability derived from electrocardiograms were compared with gold standard CAN testing with excellent reproducibility (40).

CVD events were adjudicated by the Morbidity and Mortality Committee. Adjudication was masked to former treatment assignment, HbA_1c, and glucose levels. CVD events were defined as major adverse cardiovascular events (MACE) (CVD death, nonfatal MI, or nonfatal stroke) or any-CVD (MACE plus subclinical MI on electrocardiography, angina with ischemia on exercise testing or from significant obstruction on coronary angiography, revascularization, or congestive heart failure).

Compared with the former CONV participants, former INT group participants had a 42% lower risk of any-CVD and a 57% lower risk of MACE by EDIC year 11 (2005) (41). The risk remained nominally, albeit no longer significantly, lower (23% lower for both any-CVD and MACE) 9 years later at EDIC year 20 (42) (Fig. 3). Adjusted for known and putative CVD risk factors, HbA_1c was the most significant modifiable risk factor among other vascular risk factors, second only to age, for both any-CVD and MACE (42).

In 2020, subsequent analyses with longer follow-up (median 29 years) and more events (239 participants with 421 any-CVD events and 120 participants with 149 MACE events) showed that age was the strongest risk factor for acute MI, subclinical MI, stroke, and revascularization, while HbA_1c was the strongest risk factor for CVD death, congestive heart failure, and angina and second strongest for revascularization (43). HbA_1c was the strongest modifiable risk factor for a first CVD event and also for subsequent (i.e., recurrent) CVD events.

Whether microvascular disease burden predicts future CVD events has been a long-standing question. Adjusted analyses demonstrated an increased risk for subsequent CVD events with the presence of any eye disease pathology (beyond three-step progression on retinal fundus photography), kidney disease, or nerve disease. Kidney disease, and specifically AER ≥30 mg/24 h, was the most significant risk factor, even after further adjustment for age and HbA_1c levels (44).

A previously identified polygenic risk score for CVD (45) and several single nucleotide polymorphisms were significantly associated with the risk of any-CVD and MACE independent of established risk factors (e.g., age, sex, lipids, blood pressure, and glycemia) (46). Sex differences in the cardiometabolic risk factors and the incidence of CVD were also investigated (47). Women generally had a more favorable cardiometabolic risk factor profile than the men. Despite this, and in contrast to the general population without diabetes, women were not at a significantly lower risk of CVD compared with men (47). While there was no statistically significant difference between sexes in the overall proportion of participants reporting treatment with ACE inhibitors or angiotensin receptor blockers or lipid-lowering medications, treatment with lipid-lowering medications was less frequent in the subset of women who did not achieve treatment targets for LDL cholesterol in comparison with men (30.9% vs. 44.6%). Collectively, these findings suggest the need for a recalibration of CVD risk factor stratification by sex in revised clinical care guidelines for individuals with T1D.

DCCT baseline levels of oxidized LDL (ox-LDL), MDA-modified LDL (MDA-LDL), and advanced glycosylation–modified LDL (AGE-LDL) in circulating immune complexes (ICs) were associated with CVD in unadjusted models and after adjustment for age and mean HbA_1c (48). After further adjustment for other CVD risk factors (including LDL cholesterol), oxLDL-IC and MDA-LDL-IC remained independently associated with risk of CVD and oxLDL-IC was independently associated with risk of MACE (48).

With adjustment for HbA_1c, higher paraoxonase activity, reflective of antioxidant activity, and 2,3 dinor‐8-iso prostaglandin F_2α, an oxidative marker, but not myeloperoxidase or F_2α isoprostanes, were significantly associated with lower risk of CVD (49). APOB, APOC3 and its subfractions (heparin precipitate, heparin-soluble [HS]) and apolipoprotein-defined lipoprotein subclasses–classified Lp-B were associated with any-CVD in unadjusted analyses, and association of APOC3-HS remained nominally significant in adjusted analyses. APOC3 and its subfractions and Lp-B:C were associated with risk of MACE in unadjusted analyses, and APOC3 and APOC3-HS associations remained significant in adjusted analyses (50).

In women, but not in men, serum urate was associated with subsequent risk of MACE in unadjusted analyses and after adjustment for age and HbA_1c but not after further adjustment for microalbuminuria (P = 0.3582) (51). Higher plasma kallikrein levels were associated with higher risk of any-CVD over the EDIC study period in unadjusted analyses, in analyses with adjustment for age and HbA_1c, and in the model fully adjusted for other CVD risk factors. For MACE, higher plasma kallikrein levels were associated with higher risk in the unadjusted and minimally adjusted models during EDIC (52).

Mediation analyses demonstrated that only a few factors considered individually (e.g., pulse, triglycerides, AER) explained >10% of the effect of glycemia on the risk of CVD (53). In consideration jointly in multivariable models, ∼50% of the effect of glycemia on the risk of CVD remained unexplained by these clinical factors (53). Importantly, while the proportion of the effect of HbA_1c on the risk of CVD mediated by these factors increased over time (54), HbA_1c remained strongly associated with the risk of CVD (53). Moderation analyses showed that higher pulse rate, higher triglyceride levels, use of calcium channel blockers, and presence of neuropathy individually enhanced the effect of glycemia on any-CVD, while higher pulse rate, triglyceride and AER levels, and hypertension enhanced the effect of glycemia on MAC (55).

Although there were no significant differences in intima-media thickness (IMT) between the former DCCT treatment groups at the start of EDIC (1994), by EDIC year 6 (2000), IMT was significantly higher (more atherosclerosis) in the former CONV than INT group and remained significantly higher 6 years later in EDIC year 12 (2006) (56). Common carotid IMT levels at EDIC year 1 were significantly associated with increased risk of CVD over the subsequent 17 years of follow-up in analyses adjusted for age and sex; however, these associations did not remain significant after further adjustment for traditional CVD risk factors and HbA_1c (57).

Coronary artery calcification (CAC) was only detectable in ∼20% of the relatively young EDIC cohort (mean age 43 years) at EDIC year 8 (2002), with <10% having scores >200 Agatston units (58). Participants in the former INT group had a significant 50% reduction in the odds of a score >200 compared with those in the CONV group. During 10–13 years of follow-up, CAC scores >100–300 and >300 were associated with increased risk of any-CVD and MACE in comparison with CAC scores of 0 (59).

MRI measures including left ventricular mass and end-diastolic volume were associated with older age, higher blood pressure, and other established CVD risk factors (60). There were no significant differences between the former DCCT treatments groups for any MRI outcomes. Importantly, the measures of cardiac structure and function were strongly associated with mean HbA_1c levels over time (60).

Analyses of factors associated with mortality were based on the vital status of 99.2% of the cohort as of 31 December 2012. Cause of death was assigned by the Morbidity and Mortality Committee based on reviews of death certificates, clinical center narratives, autopsy reports, and hospital records (61). There were 107 deaths (7.4%), 43 in the former INT group and 64 in the CONV group. Primary causes of death were CVD (22.4%), cancer (19.6%), acute diabetes complications (17.8%), and accidents or suicide (16.8%). All-cause mortality risk was 33% lower in the INT than in the CONV group (P < 0.05). Higher levels of HbA_1c and the development of albuminuria were associated with increased risk of all-cause mortality. Males had significantly higher all-cause mortality risk than females. While participants with a history of severe hypoglycemia did not have a significantly higher risk of all-cause mortality, the subset that had experienced past episodes of severe hypoglycemia with coma and/or seizure had a significantly higher mortality risk (61).

Compared with those among the general U.S. population, age-, sex-, and race-specific mortality rates were nominally but not significantly lower among the former INT group (standardized mortality rate [SMR] 0.88 [95% CI 0.67, 1.16]) but significantly higher among the CONV group (1.31 [1.05, 1.65]) (62). The SMR was slightly higher among female than among male participants (1.19 vs. 1.04), but the difference was not significant (relative mortality rate 1.14; P = 0.46). The SMR was significantly associated with HbA_1c, and the rate of increase in SMR among female participants was greater than that among male participants with HbA_1c ≥9% (62) (Fig. 4).

As of 30 April 2018, there were 173 deaths (12.0%). Primary causes of death were CVD (24.9%), cancer (22.0%), acute diabetes complications (16.2%), and accidents or suicide (13.9%). As of 30 April 2024, there were 280 deaths (19.4%).

During DCCT, there were no significant differences in cognitive function between INT and CONV despite the increased frequency of severe hypoglycemia in the former (16). Portions of the original DCCT cognitive battery were repeated during EDIC for evaluation of the cumulative impact of glycemia, age, and longer duration of T1D on cognitive function (63). Between 18 years (mean age 32–46 years) and 32 years (mean age 46–60 years) of combined DCCT/EDIC follow-up, the decline in psychomotor and mental efficiency was five times greater than the change from DCCT baseline to 18 years of combined follow-up. While there were no differences in cognition between the former DCCT treatment groups, cognitive decline was significantly associated with higher mean HbA_1c, more episodes of severe hypoglycemia, and elevated systolic blood pressure. The combined effect of increased mean HbA_1c and systolic blood pressure plus one or more episodes of severe hypoglycemia was equivalent to that of an additional 9.4 years of aging. In addition, 6% of those tested met the criteria for mild cognitive impairment based on the Montreal Cognitive Assessment (63).

Musculoskeletal disorders of the upper extremities (including cheiroarthropathy) have been shown to be associated with T1D and can lead to disability (64,65), yet these complications are often underappreciated in clinical research and in clinical care (66). Symptoms and/or diagnoses of adhesive capsulitis, carpal tunnel syndrome, flexor tenosynovitis, Dupuytren's contracture, and positive prayer sign were assessed in 2011–2012 in 1,217 participants (95% of active cohort, mean age 52 years, mean diabetes duration 31 years). These complications were present in 66% of participants and were associated with older age, female sex, longer duration of T1D, increased accumulation of AGEs, higher mean HbA_1c, and presence of neuropathy and/or retinopathy. Adhesive capsulitis in the shoulder, carpal tunnel syndrome, and flexor tenosynovitis were the most frequently reported conditions, occurring in 31%, 30%, and 28% of participants, respectively. Dupuytren's contractures were less common (9%). A positive prayer sign was present in 22% of participants. One-third of participants had two or more upper-limb complications; the most common combination was carpal tunnel syndrome plus flexor tenosynovitis (66). Upper-limb complications were associated with worse functional disability scores (66).

Higher glucose levels and the accumulation of AGEs have been proposed to contribute to lower bone mineral density and increased fracture risk observed in T1D (67). Between 2017 and 2019, 1,058 participants (n = 506 female, n = 552 male, mean age 59 years) had bone mineral density and vertebral fractures assessed with DXA (68). Osteoporosis was identified in nearly 10% of female and 3% of male participants, while vertebral fractures were more common among males (6%) than females (3%). In adjusted analyses, higher mean HbA_1c, increased accumulation of AGEs, and presence of kidney disease were associated with lower bone mineral density. A number of biomarkers of bone turnover, which were measured in a subset of participants (n = 232), differed from those in a subset of spousal control participants without diabetes (n = 104). Higher mean HbA_1c was associated with reduced bone formation, while lower eGFR was associated with increased bone resorption and bone formation.

Between 2020 and 2022, physical function was measured in 1,025 participants (mean age 60 years, mean diabetes duration 38 years) with the Short Physical Performance Battery (SPPB), which includes measures of balance, walking speed, and ability to rise unaided from a chair (69,70). SPPB scores <10 are associated with falls, frailty, and mortality (71) and were observed in 21% of participants, without sex differences (70). As a complement to the SPPB, participants (n = 1,087) self-reported limitations across five functional domains: lower-extremity mobility, leisure and social activity, activities of daily living, instrumental activities of daily living, and general physical function. Nearly half (48%) of the participants reported one or more functional limitations, with no difference by sex. In multivariable analyses, clinical factors associated with both the objective SPPB and self-report measures included higher mean HbA_1c, higher BMI, presence of psychological distress, and presence of CAN.

DCCT/EDIC evaluated the prevalence of hearing impairment in 1,150 individuals with T1D and 288 age-matched spousal control participants without diabetes (72). Results from the National Health and Nutrition Examination Survey (NHANES) previously reported a twofold greater prevalence of hearing loss in type 2 diabetes in comparison with those without diabetes (73). While the prevalence of hearing impairment did not differ between the EDIC cohort and spousal control participants without diabetes, older age, male sex, and noise exposure were all significantly associated with hearing impairment in both groups (72,74). While there were no significant differences in hearing impairment between the former DCCT treatment groups, mean HbA_1c was the most important risk factor. With adjustment for age and sex, for every 10% increase in mean HbA_1c, the risk of hearing impairment increased 32% for speech-frequency and 19% for high-frequency tones.

Collectively, DCCT/EDIC findings have demonstrated long-term glycemia as the single, consistent major risk factor for the development and progression of diabetes-specific complications and several age-related outcomes, including cognitive impairment, upper-limb complications, bone loss, functional limitations and reduced mobility, and hearing loss, which might also be considered complications of T1D. In concert, these findings reinforce the need for early intensive glycemic management.

The compelling long-term benefits of intensive therapy in DCCT/EDIC and the establishment of chronic glycemia, measured with HbA_1c, as the central theme in intensive therapy’s salutary effects led to the universal adoption of intensive therapy for T1D. DCCT/EDIC results reinforce the importance of achieving target glycemic control early in the course of T1D and maintaining it over time. Morbidity and premature mortality are not inevitable in T1D. The new standard of care for those with T1D provided the rationale for innovations to make intensive therapy as accessible and safe as possible (18).

Following the publication of the DCCT results, guidance in the American Diabetes Association (ADA) “Standards of Care in Diabetes” changed from vague guidance on the management of T1D to unequivocally clear, actionable glycemic targets and recommendations on the types and frequency of insulin administration (8,75). The new guidelines specifically recommended “physiologically based insulin regimens” made up of “multiple daily injections of short and longer acting insulins or continuous subcutaneous insulin infusion” (insulin pump therapy) and provided clear glycemic targets, setting forth a new era in clinical care (8). The glycemic goals in the ADA “Standards of Care in Diabetes” and in those of other international organizations cite the DCCT results (76,77). The DCCT investigators and, in particular, the diabetes educator/research coordinators have shared their experience and expertise regarding intensive insulin therapy, glycemic pattern recognition, mitigation of hypoglycemia risk, and implementation of technologies for intensive therapy through education networks (78,79).

The need for improved, safer intensive therapy spurred the development of novel insulin formulations to provide more physiologic replacement, replacement and new insulin delivery devices, and glucagon formulations. Continuous glucose monitoring technologies have allowed semiautomated pumps that have demonstrably reduced the risk of hypoglycemia while facilitating the achievement of recommended glycemic levels (80). The results of the DCCT created a “domino effect” of innovations that have transformed T1D care. Of note, ∼65% of DCCT/EDIC participants currently use insulin pumps and >80% use personal, unblinded continuous glucose monitoring.

DCCT/EDIC implemented traditional and new methods for complications screening. ETDRS procedures and grading system were adopted early in the DCCT/EDIC, and new methods, such as ultrawide-field retinal photography and optical coherence tomography, have been added. Gold standard, but laborious, electrophysiological methods for detecting DPN were ultimately changed to more clinically feasible measurements such as use of the Michigan Neuropathy Screening Instrument (81) and point of care devices to determine vibration perception thresholds. New methods were adapted and standardized to detect autonomic neuropathy, such as measuring heart rate variability (40).

With the extensive, frequent monitoring of complications required in DCCT/EDIC, the work informed and contributed to refinement of clinical practice guidelines. Specifically, recommendations for screening frequency for retinopathy and kidney disease can now be individualized, providing safe screening at lower overall costs (82,83). Finally, DCCT/EDIC data have formed the basis of simulation models for projecting long-term clinical outcomes of T1D such as the “Core Diabetes Model,” which has become a widely used tool for assessing the cost-effectiveness of diabetes interventions (84–86).

Compared with the effects of glycemic control on diabetes-specific, microvascular complications, the effects of glycemic control on CVD have been less clear, especially in type 2 diabetes where manifold nonglycemic risk factors are present. DCCT/EDIC was the first work to demonstrate the benefit of intensive insulin therapy and glucose control in reducing CVD risk in T1D (41). Considering the relatively young age and low CVD risk profiles at baseline, the beneficial effects of glycemic control may not be surprising. Nevertheless, the magnitude of the reduction in CVD and mortality attests to the importance of intensive glycemic therapy in T1D. The beneficial impact of intensive therapy and glycemic control on CVD and on mortality rates, which are similar (even nominally lower) in the DCCT/EDIC cohort in comparison with the general population, reinforces the importance of early and persistent implementation of intensive therapy. Finally, in DCCT/EDIC, CVD morbidity for women with T1D was comparable with that for the men, but women were treated less aggressively by their primary care providers. These findings support more aggressive treatment to reduce CVD risk in women with T1D.

A key finding of DCCT/EDIC is “metabolic memory,” the durable beneficial effect of glycemic separation, even after the separation has dissipated. Metabolic memory wanes to metabolic amnesia after ∼10 years. INT group participants who maintained lower HbA_1c levels throughout the study had the least complications (61,87,88) (Fig. 2 and Table 2). All of the DCCT/EDIC findings reinforce the recommendation that intensive therapy should be started as early in the course of T1D as possible and continued throughout life for optimal prevention of complications. Importantly, the findings also indicate that it is never too late to improve glycemic control to reduce the negative effect of antecedent hyperglycemia (89,90).

Over the past four decades, participant retention has remained remarkably high despite the introduction of dozens of ancillary studies and research assessments. While some participants reside close to EDIC clinical centers, many now travel significant distances to continue participation. Participant transfers to a closer EDIC location and the use of remote visits for participants unable to travel have ensured a high level of continued participation. Some of the reasons most commonly cited by participants for their continued involvement in DCCT/EDIC have included access to cutting-edge tests for assessment of diabetes complications, access to annual evaluations, and the desire to help others (91). The efforts of the study coordinators have been integral to the extraordinarily high participant retention and data completion rates. In addition, many important organizational features of DCCT/EDIC have allowed for successful long-term collaboration, including the cohesiveness of the study group, diversity of expertise in study-wide leadership, elevated role of the study coordinators, and the scientific engagement and collaboration of the NIDDK Program Officials and Project Scientists—serving as a model for other multicenter clinical studies.

With >40 years of follow-up, DCCT/EDIC is the largest and longest study of people with T1D. It provides unparalleled longitudinal phenotyping and genotyping. DCCT/EDIC could not have succeeded without the unwavering support of and partnership with NIDDK scientists and leaders, and as participants age, the study continues to evolve. The DCCT conclusively demonstrated the beneficial effects of INT therapy in delaying the onset and slowing the progression of early microvascular complications. These beneficial effects persisted long after the DCCT, a phenomenon termed metabolic memory. Longer-term follow-up of the original DCCT cohort during EDIC showed beneficial effects of the original INT therapy on more advanced microvascular complications, CVD, and mortality. Risk factors for traditional and newly described T1D-associated morbidities have been delineated during EDIC. The DCCT/EDIC cohort, now in their mid-60s with >40 years of diabetes duration, provides a unique opportunity to study age-related outcomes in long-duration T1D. The projected improvement in lifespan with INT therapy, now considered the standard of care, makes the study of T1D in aging more important than ever.

On the 75th anniversary of NIDDK, the DCCT/EDIC represents one of the longest continuous studies in NIDDK history. Its results have revolutionized the treatment of T1D and led to measurable improvements in health for people with T1D and represent the salutary impact that federal investment can and should have on the lives of people with diabetes. The continued longitudinal study of the precious DCCT/EDIC cohort, now for over 40 years, has resulted in paradigm-changing translational findings that transformed standards of care on a global level and is entirely consistent with the NIDDK mission. Such an immeasurable impact would not have been possible without the unwavering commitment and sustained, long-term support from NIDDK.

Acknowledgments. A complete list of members in the DCCT/EDIC Research Group can be found in Supplementary Material. The DCCT/EDIC Research Group owes its scientific success and public health contributions to the dedication and commitment of the DCCT/EDIC participants. Over the years, DCCT/EDIC has collaborated with outside investigators, resulting in dozens of independently funded ancillary studies, each initiative advancing our understanding of the multisystem effects of diabetes and guiding future scientific investigations.

Funding. The DCCT/EDIC study has been continuously funded through a series of cooperative agreements and contracts with the NIDDK. Support for this DCCT/EDIC collaborative study was provided by NIDDK grant DP3 DK114812. DCCT/EDIC has been supported by cooperative agreements (1982–1993, 2012–2017, 2017–2022, 2022–2027) and contracts (1982–2012) with the Division of Diabetes, Endocrinology, and Metabolic Diseases of the NIDDK (current grant nos. U01 DK094176 and U01 DK094157) and supported by the National Eye Institute, National Institute of Neurological Disorders and Stroke, General Clinical Research Centers Program (1993–2007), and Clinical Translational Science Center Program (2006–present), Bethesda, MD. The sponsor of this study is represented by the NIDDK Project Scientist who serves as part of the DCCT/EDIC Research Group and plays a part in the study design and conduct as well as the review and approval of manuscripts. The NIDDK Project Scientist was not a member of the writing group of this manuscript. Industry contributors have had no role in the DCCT/EDIC study but have provided free or discounted supplies or equipment to support participants’ adherence to the study: Abbott Diabetes Care (Alameda, CA), Animas (Westchester, PA), Bayer Diabetes Care (North America Headquarters, Tarrytown, NY), Becton, Dickinson and Company (Franklin Lakes, NJ), Eli Lilly (Indianapolis, IN), Extend Nutrition (St. Louis, MO), Insulet (Bedford, MA), LifeScan (Milpitas, CA), Medtronic Diabetes (Minneapolis, MN), Nipro Home Diagnostics (Ft. Lauderdale, FL), Nova Diabetes Care (Billerica, MA), Omron (Shelton, CT), Perrigo Diabetes Care (Allegan, MI), Roche Diabetes Care (Indianapolis, IN), and Sanofi (Bridgewater, NJ).

The opinions expressed are those of the investigators and do not necessarily reflect the views of the funding agencies.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. B.H.B., I.B., G.ML., C.L.M., B.A.P., R.G.-K., and D.M.N. all wrote sections of the manuscript and reviewed and edited the manuscript.

Data and Resource Availability. Data collected for the DCCT/EDIC study through 30 June 2017 are available to the public through the NIDDK Central Repository (https://repository.niddk.nih.gov/studies/edic/). Data collected in the last cycle (July 2017–June 2022) will be available July 2025. Data collected in the current cycle (July 2022–June 2027) will be available within 2 years after the end of the funding cycle.

Handling Editors. The journal editor responsible for overseeing the review of the manuscript was M. Sue Kirkman.