The Effect of Intensive Diabetes Therapy on the Development and Progression of Neuropathy

Methods

Study Design

The DCCT design and methods have been previously described [7]. We recruited separate primary prevention and secondary intervention cohorts to determine whether intensive diabetes therapy could prevent or retard the progression of diabetic retinopathy, the major study outcome. At entry, patients in the primary prevention cohort were free of retinopathy by stereo fundus photography and did not have microalbuminuria (albumin excretion rate < 40 mg/24 h); these patients had had insulin-dependent diabetes mellitus for 1 to 5 years. Patients in the secondary intervention cohort had mild to moderate nonproliferative retinopathy, had an albumin excretion rate of less than 200 mg/24 h, and had had insulin-dependent diabetes mellitus for 1 to 15 years [7]. Diabetic neuropathy was neither an inclusion nor an exclusion criterion unless the patient or investigator believed that the neuropathic symptoms were sufficiently severe to merit treatment. In such a case, neuropathy was grounds for exclusion. Patients were randomly assigned by cohort and clinical center to receive either conventional therapy or intensive therapy by methods that have been described previously [8]. Conventional therapy was designed to prevent symptoms of hyperglycemia and hypoglycemia with one or two insulin injections per day. Intensive treatment was designed to maintain blood glucose levels as close to the normal range as possible. Intensive management included three or more insulin injections per day or treatment with an insulin pump, with dose adjustment guided by frequent self-monitoring of blood glucose levels. Diabetic complications, including neuropathy, were assessed at baseline and periodically during the trial. Although investigators and patients could not be masked to treatment assignment, the reviewers and graders of outcome data were masked to assignment. Moreover, neither the investigators nor the patients were aware of outcome data unless predetermined criteria were met (for example, the development of retinopathy requiring photocoagulation), at which time appropriate disclosure and treatment were undertaken [7].

A total of 1441 patients was recruited into the two cohorts at 29 centers between 1983 and 1989. The study was terminated in June 1993 when its external monitoring committee decided that the main study questions had been answered [10]. Duration of follow-up at study termination was 3 to 9 years, depending on the date of recruitment (mean duration of follow-up, 6.5 years).

Clinical Neurologic Examination

A standardized neurologic history and physical examination was done by DCCT neurologists at baseline, 5 years, and study end (the first 278 patients recruited were also examined at years 1 and 2) [9]. The neurologists were masked to treatment assignment, levels of glycemia, and other medical information about the patients. During the neurologic history, peripheral, somatic, and autonomic neuropathic symptoms and other possible medical or environmental causes of peripheral neuropathy were solicited. The physical examination assessed peripheral sensation (light touch, position, temperature, and pin-prick) and deep-tendon reflexes. In the absence of other known causes of neuropathy, abnormal findings in two of these three categories (neuropathic symptoms, sensory deficits, or impaired reflexes attributable to a distal symmetric polyneuropathy) constituted a definite abnormal neurologic examination indicating clinical neuropathy [9].

Measurements of Nerve Conduction

After the clinical examinations were completed, median motor and sensory, peroneal motor, and sural sensory nerve conduction velocities; distal latencies and amplitudes; and median and peroneal motor F-wave latencies were measured according to a standard protocol [9]. Nerve conduction tracings and measurements were reviewed for completeness at the coordinating center and were compared with normative values generated from the respective DCCT clinical centers. A patient was considered to have abnormal nerve conduction when at least one measured attribute (nerve conduction velocity, F-wave latency, or amplitude) was abnormal in at least two anatomically distinct nerves. Abnormality of individual attributes was defined using the 50th percentile of the upper or lower limits of normal provided by participating electromyographers. These limits are consistent with the published 5th and 95th percentile or range of normal values for normal persons [11]. Although nerve conduction data were classified into normal and abnormal categories on the basis of this definition, the definition did not reflect the quantitative nature of these measures.

Measurements of the Autonomic Nervous System

Three autonomic nervous system tests were administered at baseline and biennially thereafter under standardized conditions [9, 12]. Beat-to-beat heart rate variation during deep breathing (R-R variation) and during a standardized Valsalva maneuver (Valsalva ratio) were recorded on magnetic tape and analyzed in a central autonomic nervous system reading laboratory. The R-R variation was computed as the dimensionless circular mean vector of R-R intervals, and the Valsalva ratio was computed as the longest R-R interval divided by the shortest interval during the Valsalva maneuver and the post-Valsalva reflex bradycardia [12]. Postural blood pressure testing consisted of two supine measurements done at least 6 minutes apart (before and after the R-R variation study) followed by repeated measurements 1, 2, 3, 4, 5, and 10 minutes after standing. Orthostatic hypotension caused by autonomic neuropathy was defined as a decrease of 10 mm Hg or more in diastolic blood pressure confirmed by an inadequate catecholamine response to standing; that is, abnormal autonomic nervous system testing was defined as an R-R variation less than 15 [13], a Valsalva ratio less than 1.5 with an R-R variation less than 20, or postural hypotension with a blunted catecholamine response [9].

Confirmed Clinical Neuropathy and Other Neurologic Outcomes

The main neurologic end point for DCCT was the development of confirmed clinical neuropathy, defined as clinical neuropathy (definite abnormal neurologic examination) confirmed by abnormal nerve conduction or autonomic nervous system testing or both (Appendix Table 5 [4]. Secondary neurologic outcomes included clinical neuropathy (definite abnormal examination without confirmation by nerve conduction or autonomic nervous system testing), subclinical neuropathy (both abnormal nerve conduction and autonomic nervous system testing but no clinical neuropathy), subclinical neuropathy (abnormal nerve conduction or abnormal autonomic nervous system test results without clinical neuropathy), abnormal nerve conduction, abnormal autonomic nervous system test results, and the individual continuous nerve conduction and autonomic nervous system measurements (Appendix Table 5.

Statistical Analysis

We compared the treatment groups using the Wilcoxon rank-sum test for continuous variables and the Pearson chi-square test for categorical variables. Binary variables were compared using the Fisher exact test. We assessed group differences in the distributions of repeated measures of continuous variables Figure 1 and cross-sectional multivariate differences using the multivariate nonparametric test of stochastic ordering proposed by Wei and Lachin, with equal weights applied to the corresponding univariate Mann-Whitney difference summary statistics [14].

For a binary measure at a single point in follow-up (for example, prevalence at 5 years), we calculated percentage risk reductions as 100% x (1 −RR), where RR is the relative risk of intensive therapy compared with conventional therapy. Confidence intervals for risk reductions were transformed from the equivalent intervals for the relative risks calculated on the log scale [15]. Because of this, we could not compute CIs for events with a prevalence of 0 in either group. “Combined” risk reductions and the associated confidence intervals are stratified-adjusted estimates, with each stratum inversely weighted to its estimated variance [15].

We evaluated changes in continuous variables from the baseline values using the Wilcoxon signed-rank test. Unless otherwise noted, tests were done at the 0.05 level of significance. A P value of 0.05 or less was considered statistically significant. All analyses were done on an intention-to-treat basis.

Results

Baseline Characteristics

The baseline demographic characteristics of the two study cohorts did not significantly differ (Table 1). Table 2 shows the distribution of neurologic abnormalities at baseline in the primary prevention and secondary intervention cohorts. Confirmed clinical neuropathy at baseline was limited to 3.5% of the primary prevention cohort and 9.4% of the secondary intervention cohort (with a slight baseline imbalance between treatment groups in the primary cohort; P = 0.04). Many patients in the primary prevention and secondary intervention cohorts showed various clinical signs or symptoms of peripheral neuropathy or abnormal nerve conduction. Relatively few patients in either cohort showed abnormal autonomic nervous system function.

Extent of Follow-up

The mean duration of follow-up for the entire cohort of 1441 patients was 6.5 years (range, 3.5 to 9 years). More than 2000 patient-years were accrued in each treatment group of both cohorts, yielding a total of approximately 9300 patient-years of observation. Because patients were recruited over a 6-year period (1983 to 1989), most patients were not studied for all 9 years of the study. Two hundred seventy-eight patients (19% of the total cohort), were studied for 9 years, and 1088 (76%) were studied for 5 years.

Final follow-up data were collected on 1422 patients (99%). Eleven patients died and 8 dropped out of the study. The vital status of the eight patients who dropped out was determined at study end; all were alive and healthy. More than 97% of expected visits for neurologic assessments took place during the entire 9-year period.

Adherence to Assigned Treatment

The patients' adherence to their randomly assigned treatment groups was high. Forty-nine patients switched from intensive to conventional therapy for some period of time during the study, and 106 patients switched from their assigned conventional treatment to intensive therapy. Ninety-five of the latter patients were women in whom intensive therapy was implemented in preparation for and during pregnancy, as per the study protocol. In all women assigned to it, conventional therapy was resumed when pregnancy ended. Overall, patients randomly assigned to intensive treatment received the assigned therapy during more than 98% of the study, and patients assigned to conventional treatment received the assigned therapy during more than 97% of the study. Mean hemoglobin A_1c levels in the intensive treatment and conventional treatment groups were separated by about 2 percentage points throughout the follow-up period (7.2% compared with 9.1%, respectively; P < 0.001).

Effect of Intensive Therapy on the Risk for Attaining Neurologic Outcomes at 5 Years

The percentage of patients in the primary prevention and secondary intervention cohorts who developed the neurologic end points at 5 years and the reduction in risk with intensive therapy compared with conventional therapy are shown in Table 3. These analyses exclude those patients (n = 92) who had the DCCT neurologic outcome of confirmed clinical neuropathy at study entry (Table 2). The risk reductions with intensive therapy ranged from 28% to 71%; all of these reductions were statistically significant except the 56% risk reduction for developing abnormal autonomic nervous system function in the primary prevention cohort and the risk reductions for developing subclinical neuropathy (abnormal autonomic nervous system function and nerve conduction study results) in either cohort (Table 3). The risk reduction of intensive therapy was evident in subgroups defined by baseline characteristics such as age, duration of disease, alcohol use, cigarette smoking, sex, renal status, and other risk factors for cardiovascular disease. Many of the subgroups defined by baseline characteristics were small, which precluded meaningful comparisons between treatment groups within these strata. However, both the level and direction of the effects of intensive therapy were similar for the subgroups.

Five-year results were also obtained from 84 of the 92 patients who had confirmed clinical neuropathy at baseline. Forty-one of these again met the definition for clinical neuropathy after 5 years. In the primary prevention group, 6 of 12 patients receiving intensive therapy (50%) and 2 of 6 patients receiving conventional therapy (33%) met the definition. In the secondary intervention group, 13 of 33 patients receiving intensive therapy (39%) and 20 of 33 patients receiving conventional therapy (61%) met the definition. The 43 patients whose status changed primarily had changes in clinical diagnosis: Only 2 of 26 patients in the intensive therapy group (8%) and 1 of 17 patients in the conventional therapy group (6%) were again graded “definite” on the basis of a clinical examination, whereas 54% of those in the intensive therapy group and 76% of those in the conventional therapy group showed improved diagnoses on the history and physical examination without improvement in nerve conduction or autonomic nervous system status.

Effects of Intensive Therapy on Somatic and Autonomic Nerve Function

Neurologic examinations were done more frequently and for a longer period of time in the 278 patients enrolled during the early phase of the DCCT in 1983-84 (examinations were done at baseline; years 1, 2, and 5; and study end) than in the overall study population. This allowed us to study measurements of sequential nerve conduction (Figure 1). The results are concordant with those of the larger patient group studied at baseline and at 5 years.

Patients receiving conventional therapy showed significant decreases (P < 0.01) in sensory-nerve conduction velocities during the first year and further gradual decreases thereafter. By contrast, sensory velocities of the median and sural nerves in the first year increased modestly in the patients in the primary prevention cohort who received intensive therapy. In the secondary intervention cohort, sensory nerve velocities decreased slightly in the first year in both treatment groups. In the intensive therapy group, median sensory nerve values returned to baseline by year 5 (Figure 1).

Conduction velocities in the motor nerves consistently (although statistically insignificantly) decreased in the conventional therapy group but remained steady or increased slightly in the intensive therapy group (Figure 1). Cross-sectional differences between the treatment groups were significant (P = 0.0038) from the first year in the primary prevention cohort and from 2 years onward (P = 0.0040) in the secondary intervention cohort. Amplitudes of the median and peroneal motor nerve and the median and sural sensory nerve in the two treatment groups did not significantly differ in either cohort. During follow-up, the intensive therapy group showed significantly less prolongation (P < 0.001) of F-wave latencies.

Similar nerve conduction results emerge from the larger patient group. As shown in Table 4, the median change in each of the four conduction velocities was negative in the conventional treatment group but was 0 or positive in the intensive treatment group. Patients receiving intensive therapy had a significantly smaller decrease in sensory-nerve and peroneal amplitudes and had less prolonged F-wave latencies. Overall, the changes in nerve-conduction attributes showed powerful differences in favor of intensive treatment (P < 0.001 in both strata).

All patients had biennial autonomic nervous system testing regardless of when they entered the study. Intensive therapy slowed the progressive decrease of the R-R variation in the primary prevention cohort (P = 0.035) but not in the secondary intervention cohort. Although intensive therapy significantly affected the R-R variation in the primary prevention cohort Figure 2, the Valsalva ratio did not differ significantly between treatment groups in either cohort.

Discussion

The precise pathogenetic mechanisms responsible for the progressive loss and damage of nerve fiber underlying clinical diabetic polyneuropathy [5] remain controversial and may involve direct metabolic and microvascular ischemic insult [16]. Abundant clinical, animal, and in vitro data strongly implicate hyperglycemia or other metabolic consequences of insulin deficiency in both components of this disease process [16, 17]. Thus, the salutary effects of intensive diabetes therapy aimed at achieving normal glucose levels on peripheral nerve function in patients with insulin-dependent diabetes mellitus, first recognized by Gregersen [18] more than 25 years ago, has a firm scientific underpinning. Nevertheless, the clinical benefit of improved metabolic control in preventing or delaying the signs or symptoms of diabetic peripheral polyneuropathy remained unproven, although surrogates for clinical neuropathy, such as nerve conduction or vibration perception threshold, had been noted to improve with intensified diabetic therapy [19, 20].

With this uncertainty in mind, the DCCT study group excluded patients with established symptomatic neuropathy severe enough to warrant pharmaceutical treatment [3]. Moreover, the study group selected as its principal neurologic outcome a conservative definition of clinically significant polyneuropathy. According to this definition, polyneuropathy must be detectable by clinical history and physical examination and confirmed by objective diagnostic tests. This was similar but not identical to the later recommendations of an American Diabetes Association Consensus Committee, which recommended that quantitative sensory testing be added to clinical examination, nerve conduction, and autonomic nervous system testing [5].

The results of the DCCT conclusively establish that intensive diabetes management in patients with insulin-dependent diabetes mellitus markedly reduces the risk for developing clinically overt, objectively confirmed diabetic polyneuropathy. Intensive therapy reduced the prevalence of both clinical and laboratory indications of neuropathy, including symptoms, sensory deficits, decreased deep-tendon reflexes, and abnormalities in nerve conduction and autonomic nervous system testing. Although the percentage of patients who developed the most conservative neurologic outcome was less than the percentage developing various components of that outcome, the risk reduction with intensive therapy was greatest for the most stringent, and presumably most relevant, outcome. The level of neuropathy encompassed by the definition of confirmed clinical neuropathy, if detected in a practitioner's office, would normally lead a physician to consider specific treatment and lifestyle recommendations for minimization of foot trauma (for example, cushioned shoes, restrictions on high-impact exercise, and intensified foot hygiene or consideration of palliative measures such as tricyclic antidepressant agents to diminish symptoms) [3].

Despite the different retinal and renal status and duration of diabetes in the two cohorts, the strikingly similar risk reductions for the development of confirmed clinical neuropathy and of the other subsidiary DCCT neurologic outcomes suggest that clinically overt diabetic neuropathy can be prevented by metabolic intervention well into the course of insulin-dependent diabetes mellitus, at least within the population included in the DCCT. Furthermore, the beneficial effects of intensive therapy were not restricted to subgroups of DCCT patients defined by age, sex, height, or renal status. The only apparent exception was that intensive therapy affected declining autonomic nervous system function only in the primary prevention cohort, perhaps suggesting that the course of autonomic neuropathy may be established earlier in the course of insulin-dependent diabetes mellitus. On the other hand, nerve conduction measures, especially nerve conduction velocity, remained relatively stable with intensive therapy throughout the study (as long as 9 years in patients recruited early in the trial), suggesting that the maintenance of nerve conduction velocity is more than merely a short-term response to improved glycemic control [18]. Reversibility of established clinical neuropathy by metabolic intervention was not well addressed in the trial because of the few patients who had clinical neuropathy at study entry and the relatively infrequent assessment of neuropathy during the course of the trial. The potent effect of intensive therapy to at least delay, and perhaps prevent, clinical neuropathy and presumably to lower the risk for foot ulcers, lower extremity amputation, and the attendant morbidity and costs [21] supports the growing consensus that intensive diabetes therapy is highly beneficial to patients with insulin-dependent diabetes mellitus.