These pages are best viewed with Netscape version 3.0 or higher or Internet Explorer version 3.0 or higher. When viewed with other browsers, some characters or attributes may not be rendered correctly.

ORIGINAL ARTICLE

Effect of Intensive Therapy on Early Macrovascular Disease in Young Individuals With Type 1 Diabetes

A systematic review and meta-analysis

Margaret L. Lawson, MD, MHSC
Hertzel C. Gerstein, MD, MSC
Elaine Tsui, MD
Bernard Zinman, MD

OBJECTIVE— We conducted a systematic review of randomized controlled trials (RCTs) of intensive insulin therapy (IIT) in type 1 diabetes to determine the effect on macrovascular complications.

RESEARCH DESIGN AND METHODS— MEDLINE (1966–1996), Citation Index, reference lists, and personal files were used to identify RCTs of >2 years' duration comparing IIT to conventional therapy (CT) in type 1 diabetes. Two independent reviewers applied selection criteria and identified 11 studies; 5 were subsequently excluded because no data were available for macrovascular complications. Data were extracted on macrovascular disease and cardiovascular risk factors. Macrovascular disease was defined as angina, myocardial infarction, angioplasty, coronary artery bypass graft, stroke, claudication, or peripheral bypass. The first event of each type was counted.

RESULTS— IIT decreased the number of macrovascular events (odds ratio [OR] 0.55, [95% CI 0.35–0.88], P = 0.015) but had no significant effect on the number of patients developing macrovascular disease (OR 0.72, [95% CI 0.44–1.17], P = 0.22) or on macrovascular mortality (OR 0.91, [95% CI 0.31–2.65], P = 0.93).

CONCLUSIONS— IIT decreases the extent of early macrovascular disease in young individuals with type 1 diabetes but has no effect on the number of patients affected or on macrovascular mortality. These data suggest that IIT may stabilize macrovascular disease or prevent progression in those at risk.

Diabetes Care 22 (Suppl. 2):B35–B39, 1999

The relationship between metabolic control and the complications of type 1 diabetes has been debated for many years. The early randomized controlled trials that examined this issue did not find significant effects because of their small sample size and short duration of follow-up. However, several studies, including the Diabetes Control and Complications Trial (DCCT), have now clearly demonstrated that intensive insulin therapy (IIT) decreases the risk of developing the microvascular complications of type 1 diabetes, namely retinopathy, nephropathy, and neuropathy (1–5).

However, the macrovascular complications of diabetes (cardiovascular, cerebrovascular, and peripheral vascular disease) contribute most to the excess morbidity and early mortality associated with type 1 diabetes. In a 20-year prospective cohort study of 292 subjects with type 1 diabetes, Krolewski et al. (6) demonstrated that 46% of all deaths were due to coronary artery disease or sudden death. Similarly, a 40-year follow-up study of individuals diagnosed with type 1 diabetes before 1933 found that 50% died before age 50 at a time when only 10% of the nondiabetic population had died, with 33% of the diabetes deaths secondary to macrovascular disease (7).

Although neither the DCCT nor previous smaller trials were designed to have the power to detect changes in the risk of developing macrovascular complications, the DCCT noted a statistically insignificant 44% reduction in relative risk in the total number of macrovascular events. Subsequent analysis found that the number of combined major macrovascular events was almost twice as high in the conventionally treated group of the DCCT (P = 0.08) (8). Because the DCCT has established IIT as the goal of care for most adolescents and adults with type 1 diabetes (9), further randomized trials would be unethical, thus leaving unanswered the question regarding the relationship between IIT and macrovascular complications. Our objective was to critically review and analyze all randomized controlled trials of IIT in type 1 diabetes and to use meta-analytical techniques to estimate the effect of IIT on the risk of developing the macrovascular complications of type 1 diabetes: cardiovascular, cerebrovascular, and peripheral vascular disease.

RESEARCH DESIGN AND METHODS

Literature search
A comprehensive literature search was conducted using MEDLINE (January 1966 to January 1996) and the following MESH terms: Insulin-Dependent Diabetes Mellitus; Tight or Intensive or Multiple or Pump; and Random or Randomized or Random Allocation or Randomized Controlled Trial. Citation Index, personal files, and bibliographies of all retrieved articles were used to identify further articles. There were no restrictions on language of publication.

Eligibility criteria
Abstracts were reviewed, unblinded (M.L.), and articles were excluded if they were not randomized controlled trials, involved subjects with type 2 diabetes, or were early reports of studies if later reports were available. Randomized trials of <2 years duration were excluded to minimize inclusion of subjects with pre-existing subclinical macrovascular disease. Crossover trials were considered for inclusion, with the intention of using data from the first phase only, but no crossover trial was longer than 2 years' duration. Studies meeting inclusion criteria were subjected to full unblinded independent review (M.L. and E.T.). IIT was defined as a method of intensifying diabetes management with the goal of improving metabolic control over that achieved by conventional therapy (CT). IIT could be achieved through multiple daily injections (3–4/day) or subcutaneous insulin pump, whereas CT was defined as one or two insulin injections per day. To test the hypothesis that improved metabolic control decreases the risk of developing macrovascular disease, studies had to have shown a statistically significant difference in glycosylated hemoglobin between the IIT and CT groups over the duration of the trial (P< 0.05). Studies were initially included regardless of whether data were provided on macrovascular complications, as all authors were to be contacted directly to confirm outcomes and/or obtain unpublished data. Agreement regarding study inclusion was evaluated by weighted with quadratic weights (10). A minimum acceptable agreement was set a priori at a of 0.65. Disagreements were resolved by a third independent reviewer (B.Z.).

Data collection
Data were extracted, independently and in duplicate (M.L. and E.T.), from the text, tables, and figures, and through personal communication with the authors of each study. Attempts were made to contact all corresponding authors of the published trials to confirm or obtain missing data.

Definition of outcome measures
The primary outcome measure was the number of major macrovascular events that were defined as in the DCCT (8) and included: 1) cardiovascular disease (angina, myocardial infarction, angioplasty, CABG); 2) cerebrovascular disease (cerebrovascular accident); 3) peripheral vascular disease (intermittent claudication, peripheral artery bypass); and 4) macrovascular death (fatal myocardial infarction, fatal cerebrovascular accident, sudden death). As in the DCCT, two or more events of the same type were counted as one; if a patient had different types of events, they were counted separately, even if they were within the same class (e.g., angina and myocardial infarction). In addition to abstracting data regarding the number of events, we contacted each of the authors to obtain data on the number of patients having one or more macrovascular event. Data were also collected, when available, on cardiovascular risk factors: hypertension, lipids, smoking, BMI, and insulin dose.

Quality assessment
Quality assessment was performed on each study, independently and in duplicate (M.L. and E.T.). A five-point scale was developed a priori, with one point given for each of the following: quality of randomization (method of randomization and allocation concealment), specific inclusion and exclusion criteria, similar distribution of baseline characteristics, <20% drop-outs or lost to follow-up, and intention-to-treat analysis or analyzable as such.

Statistical analysis
Intention-to-treat analysis was used. Individual trial results were pooled, and odds ratios (ORs) were calculated using a fixed-effects model with adjustment for low event rates (addition of 0.5 to each cell) (11). For comparison purposes, the analysis was also performed using a random effects model. Mantel-Haenszel individual and common ORs were calculated with 95% CIs (12). Studies were assessed for heterogeneity using the Breslow-Day method (13).

RESULTS— The initial search identified 30 studies. After review of the abstracts, 16 were excluded (12 were earlier reports of selected studies; 4 were of <2 years' duration). A total of 14 studies were assessed in full and selection criteria applied; 1 study met the selection criteria but was excluded, by mutual agreement, because it involved type 1 diabetic subjects after renal transplant (14). Two studies were excluded because they did not, in fact, meet inclusion criteria (in one, subjects started on insulin within 1 year of diagnosis, and it was therefore likely that not all subjects had type 1 diabetes [15]; the other was a cohort study [16]). After this review, 11 of the 14 studies were selected for inclusion ( 0.75). Attempts were made to contact each of the primary authors to obtain and/or confirm macrovascular outcomes. Four studies were excluded at this point because no data had been collected on macrovascular outcomes (17–20). We were unable to make contact with the authors of one additional study and therefore excluded it from the analysis (21). The published report of this study made no mention of macrovascular events. Sensitivity analysis that included this study as well as the four studies that did not collect data on macrovascular outcomes (17–20) and assumed a zero event rate in both groups did not change our results. The excluded studies ranged in duration from 2 to 5 years, with sample sizes of 24 to 70.

Description of studies
We report the results of the six studies for which we were able to obtain confirmation of macrovascular outcomes, two of which are reported together from the same group (2–4,8,22). The studies are described in Table 1. Each study had follow-up of >90%. The studies ranged in duration from 2 to 9 years. Four studies included subjects with evidence of early microvascular complications (secondary intervention), whereas two (4,8) included both primary prevention (no evidence of complications) and secondary intervention subjects. However, only the DCCT analyzed the primary prevention and secondary intervention subjects separately, although only for baseline characteristics and cardiovascular risk factors. Patient age at entry was relatively young (mean age <30 years) in two of the studies, whereas the other four (2–4) included patients with a higher baseline risk of macrovascular disease (>30 years of age). Each of the studies demonstrated that their IIT and CT groups had similar baseline characteristics at entry. All studies achieved a statistically significant difference in GHb between the IIT and CT groups, with a mean GHb of 122% above the upper limit of normal for the IIT group and 140% for the CT group. The six studies were similar in quality; all scored full marks for explicit entry criteria, similarity of baseline characteristics, adequacy of follow-up, and intention-to-treat analysis; the method of randomization was unclear in three of the studies (3,4).

Figure 1—Effect of IIT on number of macrovascular events. *Accurate determination of CI not possible due to zero-event rate in both groups.

Effect of IIT on macrovascular events
IIT significantly reduced the total number of first major macrovascularevents of each type (OR 0.55, [95% CI 0.35–0.88], P = 0.015, Fig. 1). In three of the studies, some subjects had more than one type of event. In the DCCT, one IIT subject had 2 types of events and four CT subjects had a total of 14 events. Two subjects in the CT group of the Stockholm Diabetes Intervention Study (SDIS) each had two different types of events, and one subject in the CT group of Holman et al.'s (2) study had two types of events (Table 2). When events within the same class (e.g., angina and myocardial infarction) were counted only once, the result became nonsignificant. Although there was a trend with IIT toward a decrease in the number of patients having one or more macrovascular event, it was not statistically significant (OR 0.72, [95% CI 0.44–1.17], P = 0.22, Fig. 2). Macrovascular mortality was not affected by IIT (OR 0.91, [95% CI 0.31–2.65], P = 0.93, Table 2). There was no significant statistical heterogeneity for all of the above analyses (P> 0.76). The results were similar when the analysis was repeated using a random-effects model.

Figure 2—Effect of IIT on number of patients with macrovascular disease. *Accurate determination of CI not possible due to zero-event rate in both groups.

Effect of IIT on cardiovascular risk factors
Minimal data were available on cardiovascular risk factors. One study did not report on blood pressure (4). Of the remaining five studies, only the Steno 1 and 2 studies (3) identified a significant increase in blood pressure in the CT group with no change in the IIT group, although there was no difference between the groups at follow-up. Two studies reported data on blood lipids, both showing a beneficial effect on serum LDL cholesterol but no effect on HDL (2,8). Only three studies examined the effect of IIT on weight: the DCCT found that IIT resulted in greater weight gain than CT, whereas the other two studies showed no difference (2,22). Again, only the DCCT found that IIT was associated with increased insulin doses compared to CT. In Steno 1, Steno 2, and SDIS, there was no difference between the groups. The two studies that examined smoking prevalence found no difference between the IIT and CT groups (8,22).

Sensitivity analysis
Publication bias was assessed through funnel plot analysis and did not appear to be present. The failsafe number was 3, indicating that three negative trials would be required for the effect of IIT on the number of macrovascular events to become nonsignificant. Sensitivity analysis based on quality assessment was not performed, nor was subgroup analysis done (e.g., duration of diabetes, age, type of macrovascular disease) because of the inordinate contribution that the DCCT would have had to these analyses, potentially biasing the analysis against subgroups represented in the smaller studies. In addition, it was not possible to analyze the results based on macrovascular risk factors because outcome events were not reported in this way.

CONCLUSIONS— The rise in macrovascular disease in type 1 diabetes does not occur until late in the third decade of life. In contrast, most of the randomized controlled trials of IIT in type 1 diabetes have included relatively young subjects (<30 years of age at entry) who were at very low risk for macrovascular disease during the 2–9 years of follow-up. Despite this bias against detecting a beneficial effect of IIT on macrovascular disease, meta-analysis of these randomized controlled trials demonstrated a decrease in the total number of macrovascular events but no significant effect on the number of patients developing macrovascular disease, although the trend was toward a benefit.

It is well known that the risk of a macrovascular event is highest in those who have already had one event (23). Our finding of a benefit of IIT over CT on the number of macrovascular events, but not on the number of patients, suggests that IIT decreases the likelihood of a patient having multiple types of events. It is important to note that different types of events within the same class (e.g., angina and myocardial infarction) were counted as separate events, as was done in the DCCT, even though events within the same class are not independent events. Nevertheless, there was a significant difference between the IIT and CT groups. These findings may indicate that IIT decreases the progression of macrovascular disease among those at high risk, which is supported by a recent study that demonstrated a reduction in 1-year mortality after acute myocardial infarction in individuals with type 1 or type 2 diabetes who were treated with an insulin-glucose infusion followed by a multidose insulin regimen (24).

Subgroup analysis would have been interesting, but it was not attempted because of potential confounding. Because of its large sample size, any subgroup represented in the DCCT would have appeared to benefit more than subgroups represented by the other, much smaller trials, regardless of whether a true effect existed. Furthermore, the DCCT Research Group has already analyzed and published data on the subgroups represented in the study (1,8).

Our examination of cardiovascular risk factors failed to explain the cause of the observed reduction in the number of macrovascular events with IIT. Studies that reported blood pressure showed conflicting results with IIT. We did not find any evidence linking IIT to macrovascular disease through weight gain and/or hyperinsulinemia. The only risk factor that was significantly and consistently reduced by IIT was serum LDL cholesterol, although only two studies examined this factor. Although not examined in these studies, it has previously been suggested that IIT may benefit macrovascular disease through improvements in platelet and fibrinolytic function (24).

The effect of long-term IIT on early atherosclerosis was recently examined using high-frequency ultrasound to assess endothelial function, carotid intima-media thickness, and arterial stiffness (25). Of the original 102 SDIS subjects, 59 participated in this study ~12 years after randomization to IIT or CT. The authors found significantly better endothelial function and less-stiff arteries in the IIT group, which their analyses attributed to lower HbA_1c and blood pressure in the intensively treated subjects. These ultrasound findings could not be linked to clinical events because there were only three macrovascular events among their 59 subjects who were part of the SDIS cohort included in our meta-analysis. However, their finding that IIT slows atherosclerosis supports the conclusions of our meta-analysis that IIT decreases the number of macrovascular events.

The long-term effects of IIT on macrovascular disease remain unknown. Because the longest follow-up to date is only 9 years, there is concern that IIT may only delay the onset and progression of microvascular disease (1). The beneficial effect of IIT on macrovascular disease might similarly wane over time. Alternatively, as patients move into the age-groups at highest risk for macrovascular disease, the beneficial effect of IIT on macrovascular disease may become more evident.

This meta-analysis has several limitations. First, as in the DCCT (8),our primary analysis focused on the total number of first events of each type. However, these macrovascular events are not independent, especially events within the same class. Second, the DCCT results had an inordinate contribution to the power of our analysis. However, the DCCT on its own did not have sufficient power to detect an effect on macrovascular complications. Third, the analysis was done after the individual studies were completed and reported, and thus our meta-analysis is subject to the limitations of retrospective studies. As a result, our findings need confirmation through prospective longitudinal studies. Finally, we only included published trials.

The exclusion of unpublished reports is controversial (26). Clearly, unpublished reports are not subject to the methodological review required for publication. However, small studies with positive results are more likely to be published than small negative studies. Nevertheless, many of the published studies identified through our literature review were small negative studies with microvascular disease as their primary outcome. Macrovascular complications were reported in these original studies as a secondary outcome or under adverse events. It is unlikely that other studies would have remained unpublished, regardless of their effect on microvascular disease, had there been macrovascular events associated with IIT. As such, publication bias is unlikely to have had an effect on our analysis.

Future trials of IIT versus CT would seem to be unethical in patients with type 1 diabetes given the positive effects on microvascular complications. However, this meta-analysis supports the DCCT findings that IIT also benefits macrovascular disease in type 1 diabetes. Furthermore, our analysis shows a benefit of IIT in decreasing the extent of macrovascular disease, expressed as the number of macrovascular events, but no significant effect on the number of patients affected or on macrovascular mortality. These findings suggest that IIT may stabilize macrovascular disease or prevent progression in those already at risk. Long-term prospective studies are needed to confirm these data and fully understand the effect of IIT on macrovascular disease in individuals with type 1 diabetes.

Acknowledgments— We wish to acknowledge the support and assistance of the authors whose primary studies were included in this meta-analysis. In particular, we wish to thank Knut Dahl-Jørgensen (Oslo), Bo Feldt-Rasmussen (Steno), Rury Holman (Oxford), David Kenny (DCCT Biostatistics Centre, George Washington University), and Per Reichard (SDIS). We greatly appreciate their efforts in providing the data and information required for our analysis. In addition, we thank David Moher, MSc (Clinical Epidemiology Unit, Thomas C. Craymers Centre for Systematic Reviews, Children's Hospital of Eastern Ontario Research Institute, Ottawa) for his epidemiological advice and critical review of the manuscript.

References
1. The DCCT Research Group: The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 329:977–986, 1993

2. Holman RR, Mayon-White V, Orde-Peckar C, Steemson J, Smith B, McPherson K, Rizza C, Knight AH, Dornan TL, Howard-Williams J, Jenkins L, Rolfe R, Barbour D, Poon P, Mann JI, Bron AJ: Prevention of deterioration of renal and sensory-nerve function by more intensive management of insulin-dependent diabetic patients: a two-year randomised prospective study. Lancet i:204–208, 1983

3. Feldt-Rasmussen B, Mathiesen ER, Jensen T, Lauritzen T, Deckert T: Effect of improved metabolic control on loss of kidney function in type 1 (insulin-dependent) diabetic patients: an update of the Steno studies. Diabetologia 34:164–170, 1991

4. Brinchmann-Hansen O, Dahl-Jørgensen K, Sankvik L, Hanssen KF: Blood glucose concentrations and progression of diabetic retinopathy: the seven year results of the Oslo study. BMJ 304:19–22, 1992

5. Reichard P, Bengt-Yngve N, Rosenqvist U: The effect of long-term intensified insulin treatment on the development of microvascular complications of diabetes mellitus. N Engl J Med 329:304–309, 1993

6. Krolewski AS, Kosinski EJ, Warram JH, Leland OS, Busick EJ, Asmal AC, Rand LI, Christlieb Arm Bradley RF, Kahn CR: Magnitude and determinants of coronary artery disease in juvenile-onset, insulin-dependent diabetes mellitus. Am J Cardiol 59:750–755, 1987

7. Deckert T, Poulsen JE, Larsen M: Prognosis of diabetics with diabetes onset before the age of thirtyone: 1. Survival, causes of death, and complications. Diabetologia 14:363–370, 1978

8. The DCCT Research Group: Effect of intensive diabetes management on macrovascular events and risk factors in the Diabetes Control and Complications Trial. Am J Cardiol 75:894–903, 1995

9. American Diabetes Association: Medical Management of Insulin-Dependent (Type 1) Diabetes. 2nd ed. Alexandria, VA, American Diabetes Association, 1994

10. Cicchetti DV, Fleiss JL: A comparison of the null distributions of weighted kappa and the ordinal statistic. Appl Psychol Measurement 1:195–201, 1977

11. Haldane JBS: The estimation and significance of the logarithm of a ratio of frequencies. Ann Hum Genet 20:309–311, 1955

12. Emerson JD: Combining estimates of the odds ratio: the state of the art. Stat Methods Med Res 3:157–178, 1994

13. Breslow NE, Day NE: Statistical Methods in Cancer Research. Vol 1. The Analysis of Case Cancer Studies. Lyon, France, LARC Scientific Publications, 1980

14. Barbosa J, Steffes MW, Sutherland DER, Connett JE, Rao KV, Mauer SM: Effect of glycemic control on early diabetic renal lesions: a 5-year randomized controlled clinical trial of insulin-dependent diabetic kidney transplant recipients. JAMA 272:600–606, 1994

15. Eschwege E, Job D, Guyot-Argenton C, Aubry JP, Tchobroutsky G: Delayed progression of diabetic retinopathy by divided insulin administration: a further follow-up. Diabetologia 16:13–15, 1979

16. Rosenstock J, Friberg T, Raskin P: Effect of glycemic control on microvascular complications in patients with type 1 diabetes mellitus. Am J Med 81:1012–1018, 1986

17. Olsen T, Richelsen B, Ehlers N, Beck-Nielsen H: Diabetic retinopathy after 3 years' treatment with continuous subcutaneous insulin infusion (CSII). Acta Ophthalmol 65:185–189, 1987

18. Christensen CK, Christiansen JS, Schmitz A, Christensen T, Hermansen K, Mogensen CE: Effect of continuous subcutaneous insulin infusion on kidney function and size in IDDM patients: a 2 year controlled study. J Diabetes Comp 1:91–95, 1987

19. The Kroc Collaborative Study Group: Diabetic retinopathy after two years of intensified insulin treatment. JAMA 260:37–41, 1988

20. Microalbuminuria Collaborative Study Group: Intensive therapy and progression to clinical microalbuminuria in patients with insulin dependent diabetes mellitus. BMJ 311:973–977, 1995

21. Verrillo A, De Teresa A, Martino C, Verrillo L, Di Chiara G: Long-term correction of hypergylcemia and progression of retinopathy in insulin-dependent diabetes: a five-year randomized prospective study. Diabetes Res Clin Pract 8:71–76, 1988

22. Reichard P, Pihl M: Mortality and treatment side-effects during long-term intensified conventional insulin treatment in the Stockholm Diabetes Intervention Study. Diabetes 43:313–317, 1994

23. Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults: Summary of the second report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (adult treatment panel II). JAMA 269:3015–3023, 1993

24. Malmberg K, Ryden L, Efendic S, Herlitz J, Nicol P, Walderstrom A, Wedel H, Welin L: Randomized trial of insulin-glucose infusion followed by subcutaneous insulin treatment in diabetic patients with acute myocardial infarction (DIGAMI Study): effects on mortality at 1 year. J Am Coll Cardiol 26:57–65, 1995

25. Jensen-Urstad KJ, Reichard PG, Rosfors JS, Lindblad LEL, Jensen-Urstad MT: Early atherosclerosis is retarded by improved long-term blood glucose control in patients with IDDM. Diabetes 45:1253–1258, 1996

26. Cook DJ, Guyatt GJ, Ryan G, Clifton J, Buckingham L, Willan A, McIlroy W, Oxman AD: Should unpublished data be included in meta-analyses? Current convictions and controversies. JAMA 269:2749–2753, 1993

Copyright © 1999 American Diabetes Association
Last updated: 3/99
For ADA Related Issues contact CustomerService@diabetes.org

For Technical Issues contact webmaster@diabetes.org