journal of Internal Medicine 1991: 230: 101-108
ADONIS
0 9 5468209 10012 72
Intensified conventional insulin treatment retards the microvascular complications of insulin-dependent diabetes mellitus (IDDM): the Stockholm Diabetes Intervention Study (SDIS) after 5 years P. R E I C H A R D , B. BERGLUNDS, A. B R I T Z , I. CARS*, B. Y. N I L S S O N t & U. R O S E N Q V I S T S From the Department of Internal Medicine 11, the *Department of Ophthalmology and the tDepartment of Neurophysiology. Sodersjukhuset. the *Department of Internal Medicine, Karolinska sjukhuset. and the §Stockholm County Council Teaching Center for Diabetes (LUCD). Stockholm. Sweden
Abstract. Reichard P. Berglund B. Britz A, Cars I, Nilsson BY. Rosenqvist U (Department of Internal Medicine 11, Department of Ophthalmology and Department of Neurophysiology, Sodersjukhuset. Department of Internal Medicine, Karolinska sjukhuset. and Stockholm County Council Teaching Center for Diabetes (LUCD), Stockholm, Sweden). Intensified conventional insulin treatment retards the microvascular complications of insulin-dependent diabetes mellitus (IDDM): the Stockholm Diabetes Intervention Study (SDIS) after 5 years. Journal oflnternal Medicine 1 9 9 1 : 230: 101-108. Ninety-six patients with insulin-dependent diabetes mellitus (IDDM) and non-proliferative retinopathy were randomized to intensified conventional treatment (ICT) ( n = 44) or regular treatment (RT) ( n = 52), and followed up for 5 years. HbA,, decreased from 9 . 5 f 0 . 2 % (mean valuefSEM) to 7 . 2 f O . l % in the ICT group, and from 9 . 4 f 0 . 2 % to 8.7 f0.1% in the RT group (difference between the groups, P < 0.001). Retinopathy increased in both groups (P < 0.001), but after 5 years it was worse in the RT group ( P < 0.05). The urinary albumin excretion rate was higher in the RT group than in the ICT group after 5 years ( 2 3 9 . 9 f 1 2 9 . 7 Fgmin-' vs. 4 6 . 0 k 2 6 . 1 pgmin-'. P < 0.05). Eight RT patients developed manifest nephropathy, compared with none in the ICT group ( P < 0.01). After 5 years the conduction velocities of the s u r d (P < 0.05), peroneal ( P < 0.01)and tibia1 (P < 0.001)nerves were lower in the RT group. The respiratory sinus arrhythmia was 12.1 f1.2 beats min-' in the RT group and 16.7 f 1.4 beats min-' in the ICT group at the end of the study (P < 0.01). The increases in retinopathy (P < 0.01), nephropathy ( P < 0.01) and neuropathy ( P < 0.001) were all related to the mean HbAIc value during the study. Smoking habits only influenced the progression of retinopathy (P < 0.05). Serious hypoglycaemia occurred in 34 ICT patients and 29 RT patients (242 and 9 8 episodes, respectively) ( P < 0.05). Whereas weight was stable in the RT group, the body mass index increased by 5.8% in the ICT group ( P < 0.01). In conclusion, microvascular complications of diabetes were retarded by intensified conventional insulin treatment. However, such treatment increased the frequency of serious hypoglycaemia, and led to an increase in body weight. Keywords: hypoglycaemia. insulin dependent diabetes mellitus, metabolic control, nephropathy. neuropathy, retinopathy.
Introduction Abbreviations: IDDM = insulin-dependent diabetes mellitus. SDIS = Stockholm Diabetes Intervention Study, ICT = intensified conventional treatment, RT = regular treatment, OR z odds ratio, CI = 95% confidence interval.
The longmterm effects of improved metabolic control On the microvascular complications Of pendent diabetes are not firmly established [l].
101
102
P. REICHARD et al.
Treatment with CSII (insulin pumps) can reduce albumin excretion [ 2 4 ] , retard the progression from microalbuminuria to manifest nephropathy [51, and affect nerve conduction velocities, at least in the short term [6, 71. Retinopathy transiently increases when blood glucose levels are lowered [4, 81, and positive long-term effects on retinopathy have not been demonstrated [9, lo]. The side-effects of intensified insulin treatment are increased hypoglycaemia [ l l ] and weight gain [12]. The Stockholm Diabetes Intervention Study (SDIS) was started as a 5-year study in 1982. It has demonstrated some effect of intensified treatment on complications [ 13-1 51 and a n increased frequency of hypoglycaemic episodes [161. However, hypoglycaemia did not lead to any deterioration of cognitive function [16]. In this paper we report on the 5-year effects and side-effects of intensified conventional insulin treatment.
Methods Patients Initially 102 patients with insulin-dependent diabetes mellitus (IDDM) were recruited as described previously [13]. They all had non-proliferative retinopathy, normal serum creatinine levels and unsatisfactory blood glucose control [13]. After 5 years, 96 patients remained in the study, while five patients had died and one had moved to another area. The patients were initially randomized with closed identical envelopes to intensified conventional treatment (ICT) or regular treatment (RT) [13]. Data for the patients at the onset of the study are given in Table 1.The postpubertal diabetes duration (from 11 years in women and from 12 years in men) was 1 5 . 8 k 0 . 9 years in the ICT group and 1 5 . 1 f 0 . 7 years in the RT group (NS). The groups did not differ with regard to smoking habits or reported alcohol consumption. The ICT patients participated in a n educational
programme, and they were recommended home blood glucose monitoring and multiple injections [13, 151. The objective for the patients in the RT group was to reduce the blood glucose level without giving rise to serious hypoglycaemia [ 131. Eightytwo per cent of the ICT patients and 42 % of the RT patients used 3-6 insulin injections daily. A follow-up investigation of complications was conducted after 18, 36 and 6 0 months. All investigators (ophthalmologist, neurophysiologist. laboratory personnel) except the physician in charge of the study were unaware of the treatment group of the individual patients. Metabolic control Glycated haemoglobin, HbA,, (normal range 3.95.7 %), was measured as described previously [ 13, 151. During the last 2 years of the study two different HPLC methods were used (Auto-Alc-8110,l Kyoto Daiichi Kagaku Ltd, Kyoto, Japan and Pharmacia, Uppsala, Sweden). They yielded the same values, and both methods were used on the same number of occasions in the two groups. HbA,, was measured on entry to the study, after 6 months and then at approximately 4-month intervals. The average HbA,,, level during the study represents the mean of 1 4 values obtained during a period of 6-60 months. Retinopathy Visual acuity was assessed by means of MonoyerGranstrom charts. Fundus photographs were obtained, coded and read as described previously [13]. We used a modification of the ETDRS classification [17], with each eye graded from 1 (microaneurysms alone) to 6 (proliferative retinopathy) [13]. The mean value for both eyes was used in the analysis. Nephropa t hy The urinary albumin excretion rate (UAER) in 24-h
Table 1 . Data for the intensified conventional treatment (ICT) and regular treatment (RT) groups at entry to the study
Group
Number (male/female)
Diabetes duration (years)
Age (years)
Insulin dose (IU kg-')
BMI (kg m-')
ICT
44 (22/22)
18.Ok1.0
29.5k1.1
0.73k0.03
22.5k0.3
RT
52 (27125)
16.1 f 0 . 7
31.6f1.0
0.75f0.03
22.8k0.4
Mean values fSEM are shown.
INTENSIFIED CONVENTIONAL INSULIN TKEATMENT
collections, and the glomerular filtration rate (GFR) as measured by 51Cr-EDTA clearance [18] were obtained as described previously [13-151. Nephropathy was diagnosed when the UAER was 2 200 pg min-', and microalbuminuria was defined as a UAER of 20-200 pg min-' [19]. Patients with values below the detection limit of the method used (10 mg-') were assigned an excretion rate of 10 pg min-'. At baseline only one urine sample was collected by each patient, but at the follow-up three collections were obtained over a period of approximately 1 year [14, 201. For classification and direct comparisons the median value was used, as the results were skewed. At each follow-up examination, and whenever the UAEK exceeded 2 0 pg min-' in a patient with previously lower levels, the urine was analysed for the presence of bacteriuria. If the latter was present, a new 24-h collection of urine was obtained after treatment. Blood pressure was measured in the supine position as described previously [14, 151. and was treated whenever it exceeded 140/90 mmHg. The protein content of the dietary intake was analysed by two dietitians at baseline and after 3 years. It was not found to change or differ between the two groups [14]. Peripheral neuropathy The nerve conduction velocities of the tibial, peroneal and sural nerves (lower normal limit = 4 1 m s-'), the amplitude of the action potential of the sural nerve (lower normal limit = 3 pV), the vibration threshold (upper normal limit = 1.75-11 pm, depending on age) and the thermal threshold (upper normal limit 8 "C) were analysed in the legs as described previously [13, 2 1-23]. The examination was performed on the dominant side, with skin temperature well controlled. Neuropathy was defined as at least two independent measurements that were abnormal [24]. Autonomic neuropathy The valsalva ratio [2 51, respiratory sinus arrhythmia (KSA) [26], dive test [ 2 7 ] . blood flow increase in the arm on contraction of the contralateral hand [28] and orthostatic blood pressure changes [2 91 were assessed at entry to the study and after 5 years.
103
Hypoglycaemia Every instance of serious hypoglycaemia (defined as hypoglycaemia requiring assistance by someone else) was noted and reported by the patients [13, 161. Body weight The patients were weighed at entry to the study and at every follow-up examination. Body mass index (BMI) was calculated as weight divided by height squared (kg m-*). Mortality The causes of death were analysed by means of autopsy reports, police reports and interviews with relatives or friends of the deceased patients [lS]. Statistical analyses Student's t-test was used for comparison of normally distributed interval data. For ordinal or skewed data the Wilcoxon sign-rank (paired data) test or the Mann-Whitney (unpaired data) test were applied. All the tests were two-tailed/two-sided. Contingency tables were analysed using the X2-test,and confidence intervals were calculated [30]. Correlations are expressed as Spearman's r-value. For multivariate analyses we used logistic regression [31]. The odds ratio (OR) and its 9 5 % confidence interval (CI) were calculated. The progression of retinopathy (deterioration of the mean retinopathy level by more than 0.5 units), nephropathy (progression from normoalbuminuria to microalbuminuria or manifest nephropathy, or from microalbuminuria to nephropathy) and peripheral neuropathy (the development of at least two pathological neurophysiological values) were analysed against the following set of independent variables : duration of diabetes, initial diastolic blood pressure, smoking habits, HbA,, at baseline and mean HbA,, during the study. P-values of > 0.05 were regarded as nonsignificant (NS).Except when otherwise stated, mean values fSEM are given. The study protocol was approved by the Ethics Committee of the Karolinska Institute, Stockholm.
P. REICHARD et al.
104
Visual acuity decreased by at least two lines in one eye in seven ICT patients and 11 RT patients.
Results Metabolic control
Nep hropa thy
The mean HbA,, values for the two treatment groups are shown in Fig. 1. The HbA,, value on entry to the study was 9.5 f 0 . 2 % in the ICT group and 9.4+0.2% in the RT group. During the study the mean HbA,, was 7.2 f0.1 % in the ICT group and 8.7 + 0. 1% in the RT group (P < 0.001).
The distribution between normal UAER. microalbuminuria and nephropathy at baseline did not differ between the two groups. The mean UAER increased from 74.3 k 31.0 pg min-' to 239.9 k 129.7 pg min-' in the RT group (P < 0.05), but remained lower in the ICT group [5 5 .7 f2 6 .7 pg min-l at entry to the study, and 46.0 k 2 6 .1 pg min-' after 5 years (NS)]. The difference between the groups was significant (P < 0.05) only after 5 years. Eight RT patients, but none of the ICT patients, developed UAER values defined as manifest nephropathy (P < 0.01). Four of these RT patients had normal UAER values at entry to the study. There was no significant correlation between the change in UAER values and the UAER value at baseline (r = -0.11). There was a significant reduction in GFR in both treatment groups (Table 2). One ICT patient and five RT patients with elevated albumin excretion rates had a GFR of < 90 ml min-' after 5 years.
Retinopathy
The relationship between the individual mean retinopathy levels at entry to the study and after 5 years is shown in Fig. 2. The mean retinopathy level increased from 2.4 f0.1 to 3.5 k 0.2 in the ICT group (P < 0.001) and from 2 .6 f0 .1 to 4.1 k 0 . 2 in the RT group (P < 0.001). The retinopathy level was significantly higher in the RT group than in the ICT group after 5 years (P < 0.05). There was no significant correlation between the change in retinopathy level and the level at entry to the study (r = 0.1). Proliferative retinopathy appeared in at least one eye in ten ICT patients and 1 5 RT patients (NS).
Fig. 1 . HbA,, values over a 60-
I
A
Ib
Ib
;1
ii'
213 I; i 6 410 4b Time from entry (months)
d8
-
d2
.56
month period in the intensified and conventional treatment (-) regular treatment (- - -) groups. Mean values f SD are shown. Hatched area denotes the normal range.
$0
0..
!2L-u1
2
3
Entry
4
5
0..
6 1
2
3 Entry
4
5
6
Fig. 2. Mean retinopathy level after 5 years compared to the level at entry to the study in (a) the intensified conventional treatment (ICT)group and (b) the regular =5 treatment (RT) group: (A) patients: ( 0 )= 1 patient.
INTENSIFIED CONVENTIONAL INSULIN TREATMENT
105
Table 2 . Glomerular filtration rate (GFR) and blood pressure in the intensified conventional treatment (ICT) and regular treatment (RT) groups at entry to the study and after 5 years RT
ICT
-
Entry
5 years
Entry
5 years
GFR (ml min-')
122k3
112+3*
126k3
115*4*
Systolic blood pressure (mmHg)
129k2
126k2t
133k2
133+2t
Diastolic blood pressure (mmHg)
77k1
77+lt
79k1
78klt
* P < 0.05, fNS. Mean values k SEM are shown.
The groups did not differ significantly with regard to systolic or diastolic blood pressure at entry to the study, and none of the groups showed significant changes during the study (Table 2). Seven ICT patients and eleven RT patients received treatment for hypertension, six ICT and ten RT patients with an angiotensin-converting-enzyme (ACE) inhibitor. One normotensive ICT patient received treatment with a selective beta-blocking agent after a myocardial infarction.
Neuropathy After 5 years the conduction velocities were lower in the RT group than in the ICT group ( P < 0.001 for the tibial nerve, P < 0.01 for the peroneal nerve and P < 0.05 for the sural nerve) (Table 3). The amplitude of the action potential for the sural nerve decreased only in the RT group (Table 3).
The differences between the groups with regard to vibration and thermal thresholds were not significant (Table 3). At entry to the study, 1 3 ICT patients and 1 7 RT patients exhibited neuropathy. After 5 years neuropathy was found in 1 6 and 34 patients, respectively ( P < 0.01). There were negative correlations between the changes in nerve conduction velocities and the conduction velocities at entry to the study ( r = -0.39 for the tibial nerve, -0.31 for the peroneal nerve and -0.40 for the sural nerve, P < 0.05), indicating that deterioration tended to be more extensive when the initial values were higher. The correlation between changes in the amplitude of the nerve action potential of the sural nerve and the value at baseline was r = -0.26 ( P < 0.05). The changes in vibration threshold correlated with the baseline value (r = 0.24, P < 0.05), while there was
Table 3. Nerve conduction velocities in the tibial. peroneal and sural nerves, amplitude of the nerve action potential of the sural nerve, vibration threshold and thermal threshold (neutral zone interval between threshold for perception of warmth and cold) in the intensified conventional treatment (ICT) and regular treatment (RT) groups at entry to the study and after 5 years RT
ICT Entry
5 years
Entry
5 years
Nerve conduction velocity (ms-') Tibia1 Peroneal Sural
41.3k0.8 43.0+0.7 44.2f1.3
42.1 k 0 . 6 t 42.8k0.6t 40.3+ 1.8'
40.4k0.7 42.1 k0.7 42.6k1.7
37.7k0.8' 39.3k0.7' 36.5+2.0*
Amplitude of the nerve action potential of the sural nerve (pV)
9.9k0.7
*
9.0 1 .Ot
+
9.1 k0.8
7.8
Vibration threshold (pm)
1.2 0.2
2.7k1.0'
1.5k0.3
2.8 *0.5*
Temperature threshold ("C)
5.9 0.3
+
7.5k0.6.'
6.7+0.5
8.4k0.8**
*P < 0.01, **P < 0.05. tNS. Mean values k SEM are shown.
1.1..
106
P. REICHARD et al.
Table 4. Autonomic nerve function in the intensified conventional treatment (ICT) and regular treatment (RT) groups at entry to the study and after 5 years RT
ICT
Entry
5 years
Entry
Valsalva ratio (heart rate ratio)
2.5k0.5
1.9kO.lt
1.9 0.0
1.8 f0.1*
RSA ( A heart rate)
19.0f1.2
16.721.4;
17.0k1.3
12.1+1.2**
Dive test ( A heart rate)
-28f2
-2Of2"
-24k2
-23+2t
Blood flow (% blood flow)
158+6
131*5**
162+7
139 f4.'
Orthostatic test ( A blood pressure)
0.7k 2.3
-
-4.9f2.4
-2.4k 3.2t
1.4k3 . l t
+
5 years
*P < 0.05, **P < 0.01. tNS. Mean values fSEM are shown.
a negative correlation between the changes in thermal threshold and the value at entry to the study ( r = -0.33, P < 0.05).
Autonomic neuropathy Both groups tended to show a deterioration in autonomic nerve function (Table 4). The RSA was significantly lower in the RT group than in the ICT group after 5 years ( P < 0.01). There was a negative correlation between the change in RSA and the initial RSA value ( r = - 0.3 7, P < 0.05).
Hypogl y caem ia During the entire study a total of 34 ICT patients (77%, 9 5 % confidence interval 6 5 4 9 % ) and 29 RT patients (56%, CI 43-69%) exhibited at least one serious hypoglycaemic episode (P < 0.05). There were 242 episodes of serious hypoglycaemia altogether in the ICT group, and 98 episodes in the RT group.
Mortality Four ICT patients died during the first 36 months, as previously reported [15]. One KT patient died of a myocardial infarction that occurred between 36 and 60 months after the start of the study. The patients who died did not differ from the rest with regard to duration of diabetes or progression of microvascular complications during the first 18 months.
Multivariate analyses The increases in retinopathy (OR = 2.2. CI = 1.33.7, P < 0.01). nephropathy (OR = 2.3, CT = 1.33.4, P < 0.01) and neuropathy (OR = 3.5, CI = 1.7-7.0. P < 0.001) were all related to the mean HbA,, value during the study, but not to the initial HbA,, level. the duration of diabetes or the diastolic blood pressure at entry to the study. Smoking habits were related to the progession of retinopathy (OR = 3.0, CI = 1.2-7.9, P < 0.05), but not to the other complications.
Discussion Body weight The mean BMI increased from 2 2 . 5 f 0 . 3 to 23.8 f 0 . 4 kg m-' in the ICT group (P < 0.01), but there was no increase in the RT group (22.8 0.3 kg m-' at entry and 22.8 f0.3 kg m-' after 5 years). The mean weight gain in the ICT group was 4.1 kg per patient.
The results of the retrospective, cross-sectional and non-randomized prospective studies [32-3 51 suggest that there is a relationship between metabolic control and late complications. We have shown that IDDM patients with unsatisfactory metabolic control can maintain a lower blood glucose level for a long period without insulin
INTENSIFIED CONVENTIONAL INSULIN TREATMENT
pumps. For the first time we can demonstrate that lower blood glucose levels lead to a decrease in the progression of retinopathy. nephropathy and peripheral neuropathy. Retinopathy increased to slightly higher levels in the RT group compared to the ICT group. This has not been demonstrated previously [3, 4, 9, 101, but the present study is larger than the previous ones, and it is also of longer duration. Most of the differences between the groups were within the nonproliferative range. The urinary albumin excretion (UAER) is influenced by dietary protein intake [36] and blood pressure [37]. In our treatment groups there were no significant differences at baseline with regard to these factors. None of the groups showed significant changes during the study (for protein intake the analysis was made at the time of entry and after 3 years 1141. More RT patients than ICT patients developed neuropathy, defined as at least two pathological neurophysiological values. Such use of two abnormal measurements when defining neuropathy was suggested by Dyck et al. [24], who, however, employed a somewhat different set of investigations. Mortality was not correlated with treatment regimen as far as could be established when the individual patients were analysed [ 151. The number of patients in this study is too small to allow a detailed statistical analysis. The deceased ICT patients did not differ from the others with regard to the progression of complications, and thus their exclusion did not affect the comparison between the two groups. The frequency of serious hypoglycaemia was higher in the ICT group. As described previously 1151, hypoglycaemia could not be excluded as the cause of death in one patient. This individual had experienced frequent serious hypoglycaemic episodes before entry to the study, mostly after alcohol consumption. The patient’s HbA,, levels had not reached < 8 % ; thus treatment had not been very intensive. Body weight increased in the ICT group but not in the RT group. We have already shown after 3 years that this was related to the total energy intake [15]. When making multiple comparisons between two treatment groups there is always a risk of statistical type 1 errors. However, in this study not one comparison, but most of them, pointed toward a favourable effect of intensified treatment. This tendency was also consistent throughout the study. These effects could not be due merely to chance.
107
At baseline there were no significant differences with regard to microvascular complications. However, there were small differences, and they generally favoured the ICT group. We have found no explanation for this observation. Post-pubertal diabetes duration has been shown to correlate with the development of retinopathy and nephropathy [38], but it did not differ between the groups. In general there were no significant positive correlations between progression of complications and baseline values. Thus the small differences between the groups on entry to the study cannot explain the long-term differences with regard to microvascular complications. The multivariate analyses clearly indicate that the mean HbA,, value during the study was related to the development of each of the three main microvascular complications. Among the patients who participated in this study the risks of deterioration with regard to microvascular complications increased approximately 2- to 3.5-fold for each oneunit increase in the mean HbA,,, value. Smoking clearly influenced the progression of retinopathy. The fact that subclinical retinopathy, nephropathy and neuropathy were affected by the lower blood glucose levels indicates that the long-term prognosis will probably be improved as well, but this has not yet been confirmed. In particular, the findings with regard to nephropathy, where the natural progression from microalbuminuria to manifest nephropathy and then renal insufficiency is well known, are promising.
Acknowledgements This study was supported by grants from the Swedish division of NOVO-Nordisk Inc., the Boehringer Mannheim Scand. Inc., and the Swedish Medical Research Council (06615).
References 1 The DCCT Research Group. Are continuing studies of metabolic control and microvascular complications in insulindependent diabetes mellitus justified? N Engl / Med 1988: 31 8 : 246-50. 2 Dahl-Jorgensen K. Hanssen KF. Keirulf P. Bjoro T. Sandvik L, Aagenaes 0. Reduction of urinary albumin excretion after 4 years of continuous subcutaneous insulin infusion in insulindependent diabetes mellitus. The Oslo Study. Acta Endocrinol (Copenh) 1988: 117: 19-25. 3 Lauritzen T. Frost-Larsen K. Larsen H-W. Deckert T. the Steno Study Group. Efkct of 1 year of near-normal blood glucose levels on retinopathy in insulin-dependent diabetes. I ~ n c e t 1983; 1 : 2 0 0 4 .
108
P. R E I C H A R D
et al.
4 The KROC Collaborative Study Group. Blood glucose control and the evolution of diabetic retinopathy and albuminuria. N Engl ] Med 1984: 2 9 0 : 365-72. 5 Feldt-Rasmussen B. Mathiesen ER. Deckert T. Etrect of two years of strict metabolic control on progression of incipient nephropathy in insulin-dependent diabetes. Lancet 1 9 8 6 : 11 : 13004. 6 Dahl-Jorgensen K, Brinchmann-Hansen 0. Hanssen KF et a!. Effect of near normoglycemia for two years on progression of early diabetic retinopathy. nephropathy and neuropathy : the Oslo study. BM] 1 9 8 6 : 293: 1195-9. 7 Service FJ. Rizza RA. Daube JR. O'Brien PCO. Uyck PJ. Near normoglycaemia improved nerve conduction and vibration sensation in diabetic neuropathy. Diabetologia 1985: 2 8 : 722-7. 8 Dahl-Jorgensen K, Brinchmann-Hansen 0. Hanssen KF. Sandvik L. Aagenaes 0, Aker Diabetes group. Rapid tightening of blood glucose control leads to transient deterioration of retinopathy in insulin-dependent diabetes mellitus: the Oslo study. BM] 1 9 8 5 ; 2 9 0 : 811-5. 9 Brinchrnann-Hansen 0, Dahl-Jorgensen K. Hanssen KF. Sandvik L. The response of diabetic retinopathy to 41 months of multiple insulin injections, insulin pumps and conventional insulin therapy. Arch Ophthalmol 1 9 8 8 : 106: 1242-6. 10 Olsen T, Richelsen B, Ehlers N. Beck-Nielsen H. Diabetic retinopathy after 3 years' treatment with continuous subcutaneous insulin infusion (CSII). Acta Ophthalmol (Copenh) 1 9 8 7 : 65: 185-9. 11 The DCCT Research Group. Diabetes Control and Complications Trial (DCCT): results of feasibility study. Diabetes Care 1 9 8 7 : 10: 1-19. 1 2 The DCCT Research Group. Weight gain associated with intensive therapy in the Diabetes Control and Complications Trial. Diabetes Care 1 9 8 8 ; 1 1 : 567-73. 13 Reichard P. Britz A, Cars I. Nilsson BY, Sobocinsky-Olsson B. Rosenqvist U. The Stockholm Diabetes Intervention Study (SDIS): 18 months' results. Acta Med Scand 1988: 224: 1 1 5-22. 14 Reichard P. Kosenqvist U. Nephropathy is delayed by intensitied insulin treatment in patients with insulin-dependent diabetes mellitus and retinopathy. / Intern Med 1 9 8 9 : 226: 81-7. 1 5 Reichard P. Britz A. Carlsson P et al. Metabolic control and complications over 3 years in patients with insulin-dependent diabetes mellitus (IDDM):the Stockholm Diabetes Intervention Study (SDIS). I Intern Med 1 9 9 0 ; 228: 511-7. 1 6 Reichard P. Berglund B. Britz A, kvander S. Rosenqvist U. Hypoglycemic episodes during three years of intensified conventional insulin treatment: increased frequency but no effects on cognitive function. Intern Med 1991 : 229: 9-16. 17 Early Treatment Diabetic Retinopathy Study. Manual of Operations. Bethesda. Maryland : Public Health Service, 1980. 1 8 Chantler C. Garnett ES. Parsons V. Veal1 N. Glomerular filtration rate measurement in man by the single injection method using "Cr-EDTA. Clin Sci 1 9 6 9 : 37: 169-80. 1 9 Mogensen CE, Chachati A, Christensen CK et al. Microalbuminuria: a n early marker of renal involvement in diabetes. Uremia Znvest 1985-86: 9 : 85-95. 2 0 Feldt-Rasmussen B. Mathiesen ER. Variability of urinary albumin excretion in incipient diabetic nephropathy. Diab Nephropathy 1 9 8 4 : 3: 101-3. 21 Ludin H-P. Electrornyography in Practice. Stuttgart. New York: Georg Thieme Verlag. Thieme-Stratton Inc., 1 9 8 0 ; 3 1 4 2 .
2 2 Goldberg ]M. Lindblom U. Standardised method of determining vibratory perception thresholds for diagnosis and screening in neurological investigation. ] Neurol Neurosurg Psychiatry 1 9 7 9 ; 4 2 : 793-803. 2 3 Fruhstorfer H. Lindblom U. Schmidt WG. Method for quantitative estimation of thermal thresholds in patients. / Neurol Neurosurg Psychiatry 1 9 7 6 : 39: 1071-5. 2 4 Dyck PJ. Karnes J. O'Brien PC. Diagnosis, staging, and classification of diabetic neuropathy and associations with other complications. In: Dyck PJ. Thomas PK. Asbury AK. Winegrad AI. Porte D. eds. Diabetic Neuropathy. Philadelphia : W.B. Saunders Company, 1987: 36-44. 25 Levin AV. A simple test of cardiac function based upon heart rate changes induced by the Valsalva manoeuvre. A m ] Cardiol 1966: 1 8 : 90-9. 2 6 Freyschuss U. Melcher A. Sinus arrhythmia in man: influence of tidal volume and oesophageal pressure. Scand / Clin I*lb Invest 1 9 7 5 : 35: 487-96. 2 7 Bennett T. Hosking DJ, Hampton JR. Cardiovascular responses to apnoic face immersion and mental stress in diabetic subjects. Curdiovasc Res 1976: 1 0 : 192-9. 2 8 Eklund B. Kaijser L. Knutsson E. Blood flow in resting (contralateral) arm and leg during isometric contraction. ] Physiol (Lond) 1 9 7 4 : 240: 11 1-24. 29 Burke D. Sundlof G. Wallin BG. Postural effects on muscle nerve sympathetic activity in man. 1 Physiol (Lond) 1977: 272: 3 9 9 4 1 4 . 30 Conover WJ. Practical Non-l~arametricstatistics. 2nd edn. New York: John Wiley and Sons. 1980: 99-105. 31 Breslow NE. Day NE. Statistical Methods in Cancer Research. Vol. 1. Lyon: IARC Scientific Publications No. 32. 1 9 8 0 : 196-246. 32 Pirart 1. Diabetes mellitus and its degenerative complications: a prospective study of 4 4 0 0 patients observed between 1947 and 1973. Diabetes Cure 1 9 7 8 : 1 : 168-88. 3 3 Klein R. Klein BEK. Moss SE. Davis MU. DeMets D. The Wisconsin epidemiologic study of diabetic retinopathy. 11. Prevalence and risk of diabetic retinopathy when age at diagnosis is less than 3 0 years. Arch Ophthalmol 1 9 8 4 : 102: 520-6. 34 Rosenstock J. Friberg T.Raskin P. Effect of glycemic control on microvascular complications in patients with type 1 diabetes mellitus. Am 1 Med 1986: 81: 1012-8. 35 McCance DR. Atkinson AB, Hadden DR. Archer DB. Kennedy L. Long-term glycaemic control and diabetic retinopathy. Lancet 1989: 1 1 : 824-8. 36 Cohen D, Dodds R, Viberti G. Effect of protein restriction in insulin dependent diabetics at risk of nephropathy. BM] 1987; 2 9 4 : 795-8. 37 Mogensen CE. Long-term antihypertensive treatment inhibiting progression of diabetic nephropathy. BM] 1982 : 2 8 5 : 685-8. 38 Norris Kostraba J. Dorman JS. Orchard TJ et al. Contribution of diabetes duration before puberty to development of microvascular complications in IDDM subjects. Diabetes Care 1989 ; 1 2 : 686-93. Received 2 7 November 1990, accepted 8 January 1991.
Correspondence: Dr Per Reichard. Department of Internal Medicine 11. Sodersjukhuset, S-100 6 4 Stockholm, Sweden.
ADONIS
0 9 5468209 10012 72
Intensified conventional insulin treatment retards the microvascular complications of insulin-dependent diabetes mellitus (IDDM): the Stockholm Diabetes Intervention Study (SDIS) after 5 years P. R E I C H A R D , B. BERGLUNDS, A. B R I T Z , I. CARS*, B. Y. N I L S S O N t & U. R O S E N Q V I S T S From the Department of Internal Medicine 11, the *Department of Ophthalmology and the tDepartment of Neurophysiology. Sodersjukhuset. the *Department of Internal Medicine, Karolinska sjukhuset. and the §Stockholm County Council Teaching Center for Diabetes (LUCD). Stockholm. Sweden
Abstract. Reichard P. Berglund B. Britz A, Cars I, Nilsson BY. Rosenqvist U (Department of Internal Medicine 11, Department of Ophthalmology and Department of Neurophysiology, Sodersjukhuset. Department of Internal Medicine, Karolinska sjukhuset. and Stockholm County Council Teaching Center for Diabetes (LUCD), Stockholm, Sweden). Intensified conventional insulin treatment retards the microvascular complications of insulin-dependent diabetes mellitus (IDDM): the Stockholm Diabetes Intervention Study (SDIS) after 5 years. Journal oflnternal Medicine 1 9 9 1 : 230: 101-108. Ninety-six patients with insulin-dependent diabetes mellitus (IDDM) and non-proliferative retinopathy were randomized to intensified conventional treatment (ICT) ( n = 44) or regular treatment (RT) ( n = 52), and followed up for 5 years. HbA,, decreased from 9 . 5 f 0 . 2 % (mean valuefSEM) to 7 . 2 f O . l % in the ICT group, and from 9 . 4 f 0 . 2 % to 8.7 f0.1% in the RT group (difference between the groups, P < 0.001). Retinopathy increased in both groups (P < 0.001), but after 5 years it was worse in the RT group ( P < 0.05). The urinary albumin excretion rate was higher in the RT group than in the ICT group after 5 years ( 2 3 9 . 9 f 1 2 9 . 7 Fgmin-' vs. 4 6 . 0 k 2 6 . 1 pgmin-'. P < 0.05). Eight RT patients developed manifest nephropathy, compared with none in the ICT group ( P < 0.01). After 5 years the conduction velocities of the s u r d (P < 0.05), peroneal ( P < 0.01)and tibia1 (P < 0.001)nerves were lower in the RT group. The respiratory sinus arrhythmia was 12.1 f1.2 beats min-' in the RT group and 16.7 f 1.4 beats min-' in the ICT group at the end of the study (P < 0.01). The increases in retinopathy (P < 0.01), nephropathy ( P < 0.01) and neuropathy ( P < 0.001) were all related to the mean HbAIc value during the study. Smoking habits only influenced the progression of retinopathy (P < 0.05). Serious hypoglycaemia occurred in 34 ICT patients and 29 RT patients (242 and 9 8 episodes, respectively) ( P < 0.05). Whereas weight was stable in the RT group, the body mass index increased by 5.8% in the ICT group ( P < 0.01). In conclusion, microvascular complications of diabetes were retarded by intensified conventional insulin treatment. However, such treatment increased the frequency of serious hypoglycaemia, and led to an increase in body weight. Keywords: hypoglycaemia. insulin dependent diabetes mellitus, metabolic control, nephropathy. neuropathy, retinopathy.
Introduction Abbreviations: IDDM = insulin-dependent diabetes mellitus. SDIS = Stockholm Diabetes Intervention Study, ICT = intensified conventional treatment, RT = regular treatment, OR z odds ratio, CI = 95% confidence interval.
The longmterm effects of improved metabolic control On the microvascular complications Of pendent diabetes are not firmly established [l].
101
102
P. REICHARD et al.
Treatment with CSII (insulin pumps) can reduce albumin excretion [ 2 4 ] , retard the progression from microalbuminuria to manifest nephropathy [51, and affect nerve conduction velocities, at least in the short term [6, 71. Retinopathy transiently increases when blood glucose levels are lowered [4, 81, and positive long-term effects on retinopathy have not been demonstrated [9, lo]. The side-effects of intensified insulin treatment are increased hypoglycaemia [ l l ] and weight gain [12]. The Stockholm Diabetes Intervention Study (SDIS) was started as a 5-year study in 1982. It has demonstrated some effect of intensified treatment on complications [ 13-1 51 and a n increased frequency of hypoglycaemic episodes [161. However, hypoglycaemia did not lead to any deterioration of cognitive function [16]. In this paper we report on the 5-year effects and side-effects of intensified conventional insulin treatment.
Methods Patients Initially 102 patients with insulin-dependent diabetes mellitus (IDDM) were recruited as described previously [13]. They all had non-proliferative retinopathy, normal serum creatinine levels and unsatisfactory blood glucose control [13]. After 5 years, 96 patients remained in the study, while five patients had died and one had moved to another area. The patients were initially randomized with closed identical envelopes to intensified conventional treatment (ICT) or regular treatment (RT) [13]. Data for the patients at the onset of the study are given in Table 1.The postpubertal diabetes duration (from 11 years in women and from 12 years in men) was 1 5 . 8 k 0 . 9 years in the ICT group and 1 5 . 1 f 0 . 7 years in the RT group (NS). The groups did not differ with regard to smoking habits or reported alcohol consumption. The ICT patients participated in a n educational
programme, and they were recommended home blood glucose monitoring and multiple injections [13, 151. The objective for the patients in the RT group was to reduce the blood glucose level without giving rise to serious hypoglycaemia [ 131. Eightytwo per cent of the ICT patients and 42 % of the RT patients used 3-6 insulin injections daily. A follow-up investigation of complications was conducted after 18, 36 and 6 0 months. All investigators (ophthalmologist, neurophysiologist. laboratory personnel) except the physician in charge of the study were unaware of the treatment group of the individual patients. Metabolic control Glycated haemoglobin, HbA,, (normal range 3.95.7 %), was measured as described previously [ 13, 151. During the last 2 years of the study two different HPLC methods were used (Auto-Alc-8110,l Kyoto Daiichi Kagaku Ltd, Kyoto, Japan and Pharmacia, Uppsala, Sweden). They yielded the same values, and both methods were used on the same number of occasions in the two groups. HbA,, was measured on entry to the study, after 6 months and then at approximately 4-month intervals. The average HbA,,, level during the study represents the mean of 1 4 values obtained during a period of 6-60 months. Retinopathy Visual acuity was assessed by means of MonoyerGranstrom charts. Fundus photographs were obtained, coded and read as described previously [13]. We used a modification of the ETDRS classification [17], with each eye graded from 1 (microaneurysms alone) to 6 (proliferative retinopathy) [13]. The mean value for both eyes was used in the analysis. Nephropa t hy The urinary albumin excretion rate (UAER) in 24-h
Table 1 . Data for the intensified conventional treatment (ICT) and regular treatment (RT) groups at entry to the study
Group
Number (male/female)
Diabetes duration (years)
Age (years)
Insulin dose (IU kg-')
BMI (kg m-')
ICT
44 (22/22)
18.Ok1.0
29.5k1.1
0.73k0.03
22.5k0.3
RT
52 (27125)
16.1 f 0 . 7
31.6f1.0
0.75f0.03
22.8k0.4
Mean values fSEM are shown.
INTENSIFIED CONVENTIONAL INSULIN TKEATMENT
collections, and the glomerular filtration rate (GFR) as measured by 51Cr-EDTA clearance [18] were obtained as described previously [13-151. Nephropathy was diagnosed when the UAER was 2 200 pg min-', and microalbuminuria was defined as a UAER of 20-200 pg min-' [19]. Patients with values below the detection limit of the method used (10 mg-') were assigned an excretion rate of 10 pg min-'. At baseline only one urine sample was collected by each patient, but at the follow-up three collections were obtained over a period of approximately 1 year [14, 201. For classification and direct comparisons the median value was used, as the results were skewed. At each follow-up examination, and whenever the UAEK exceeded 2 0 pg min-' in a patient with previously lower levels, the urine was analysed for the presence of bacteriuria. If the latter was present, a new 24-h collection of urine was obtained after treatment. Blood pressure was measured in the supine position as described previously [14, 151. and was treated whenever it exceeded 140/90 mmHg. The protein content of the dietary intake was analysed by two dietitians at baseline and after 3 years. It was not found to change or differ between the two groups [14]. Peripheral neuropathy The nerve conduction velocities of the tibial, peroneal and sural nerves (lower normal limit = 4 1 m s-'), the amplitude of the action potential of the sural nerve (lower normal limit = 3 pV), the vibration threshold (upper normal limit = 1.75-11 pm, depending on age) and the thermal threshold (upper normal limit 8 "C) were analysed in the legs as described previously [13, 2 1-23]. The examination was performed on the dominant side, with skin temperature well controlled. Neuropathy was defined as at least two independent measurements that were abnormal [24]. Autonomic neuropathy The valsalva ratio [2 51, respiratory sinus arrhythmia (KSA) [26], dive test [ 2 7 ] . blood flow increase in the arm on contraction of the contralateral hand [28] and orthostatic blood pressure changes [2 91 were assessed at entry to the study and after 5 years.
103
Hypoglycaemia Every instance of serious hypoglycaemia (defined as hypoglycaemia requiring assistance by someone else) was noted and reported by the patients [13, 161. Body weight The patients were weighed at entry to the study and at every follow-up examination. Body mass index (BMI) was calculated as weight divided by height squared (kg m-*). Mortality The causes of death were analysed by means of autopsy reports, police reports and interviews with relatives or friends of the deceased patients [lS]. Statistical analyses Student's t-test was used for comparison of normally distributed interval data. For ordinal or skewed data the Wilcoxon sign-rank (paired data) test or the Mann-Whitney (unpaired data) test were applied. All the tests were two-tailed/two-sided. Contingency tables were analysed using the X2-test,and confidence intervals were calculated [30]. Correlations are expressed as Spearman's r-value. For multivariate analyses we used logistic regression [31]. The odds ratio (OR) and its 9 5 % confidence interval (CI) were calculated. The progression of retinopathy (deterioration of the mean retinopathy level by more than 0.5 units), nephropathy (progression from normoalbuminuria to microalbuminuria or manifest nephropathy, or from microalbuminuria to nephropathy) and peripheral neuropathy (the development of at least two pathological neurophysiological values) were analysed against the following set of independent variables : duration of diabetes, initial diastolic blood pressure, smoking habits, HbA,, at baseline and mean HbA,, during the study. P-values of > 0.05 were regarded as nonsignificant (NS).Except when otherwise stated, mean values fSEM are given. The study protocol was approved by the Ethics Committee of the Karolinska Institute, Stockholm.
P. REICHARD et al.
104
Visual acuity decreased by at least two lines in one eye in seven ICT patients and 11 RT patients.
Results Metabolic control
Nep hropa thy
The mean HbA,, values for the two treatment groups are shown in Fig. 1. The HbA,, value on entry to the study was 9.5 f 0 . 2 % in the ICT group and 9.4+0.2% in the RT group. During the study the mean HbA,, was 7.2 f0.1 % in the ICT group and 8.7 + 0. 1% in the RT group (P < 0.001).
The distribution between normal UAER. microalbuminuria and nephropathy at baseline did not differ between the two groups. The mean UAER increased from 74.3 k 31.0 pg min-' to 239.9 k 129.7 pg min-' in the RT group (P < 0.05), but remained lower in the ICT group [5 5 .7 f2 6 .7 pg min-l at entry to the study, and 46.0 k 2 6 .1 pg min-' after 5 years (NS)]. The difference between the groups was significant (P < 0.05) only after 5 years. Eight RT patients, but none of the ICT patients, developed UAER values defined as manifest nephropathy (P < 0.01). Four of these RT patients had normal UAER values at entry to the study. There was no significant correlation between the change in UAER values and the UAER value at baseline (r = -0.11). There was a significant reduction in GFR in both treatment groups (Table 2). One ICT patient and five RT patients with elevated albumin excretion rates had a GFR of < 90 ml min-' after 5 years.
Retinopathy
The relationship between the individual mean retinopathy levels at entry to the study and after 5 years is shown in Fig. 2. The mean retinopathy level increased from 2.4 f0.1 to 3.5 k 0.2 in the ICT group (P < 0.001) and from 2 .6 f0 .1 to 4.1 k 0 . 2 in the RT group (P < 0.001). The retinopathy level was significantly higher in the RT group than in the ICT group after 5 years (P < 0.05). There was no significant correlation between the change in retinopathy level and the level at entry to the study (r = 0.1). Proliferative retinopathy appeared in at least one eye in ten ICT patients and 1 5 RT patients (NS).
Fig. 1 . HbA,, values over a 60-
I
A
Ib
Ib
;1
ii'
213 I; i 6 410 4b Time from entry (months)
d8
-
d2
.56
month period in the intensified and conventional treatment (-) regular treatment (- - -) groups. Mean values f SD are shown. Hatched area denotes the normal range.
$0
0..
!2L-u1
2
3
Entry
4
5
0..
6 1
2
3 Entry
4
5
6
Fig. 2. Mean retinopathy level after 5 years compared to the level at entry to the study in (a) the intensified conventional treatment (ICT)group and (b) the regular =5 treatment (RT) group: (A) patients: ( 0 )= 1 patient.
INTENSIFIED CONVENTIONAL INSULIN TREATMENT
105
Table 2 . Glomerular filtration rate (GFR) and blood pressure in the intensified conventional treatment (ICT) and regular treatment (RT) groups at entry to the study and after 5 years RT
ICT
-
Entry
5 years
Entry
5 years
GFR (ml min-')
122k3
112+3*
126k3
115*4*
Systolic blood pressure (mmHg)
129k2
126k2t
133k2
133+2t
Diastolic blood pressure (mmHg)
77k1
77+lt
79k1
78klt
* P < 0.05, fNS. Mean values k SEM are shown.
The groups did not differ significantly with regard to systolic or diastolic blood pressure at entry to the study, and none of the groups showed significant changes during the study (Table 2). Seven ICT patients and eleven RT patients received treatment for hypertension, six ICT and ten RT patients with an angiotensin-converting-enzyme (ACE) inhibitor. One normotensive ICT patient received treatment with a selective beta-blocking agent after a myocardial infarction.
Neuropathy After 5 years the conduction velocities were lower in the RT group than in the ICT group ( P < 0.001 for the tibial nerve, P < 0.01 for the peroneal nerve and P < 0.05 for the sural nerve) (Table 3). The amplitude of the action potential for the sural nerve decreased only in the RT group (Table 3).
The differences between the groups with regard to vibration and thermal thresholds were not significant (Table 3). At entry to the study, 1 3 ICT patients and 1 7 RT patients exhibited neuropathy. After 5 years neuropathy was found in 1 6 and 34 patients, respectively ( P < 0.01). There were negative correlations between the changes in nerve conduction velocities and the conduction velocities at entry to the study ( r = -0.39 for the tibial nerve, -0.31 for the peroneal nerve and -0.40 for the sural nerve, P < 0.05), indicating that deterioration tended to be more extensive when the initial values were higher. The correlation between changes in the amplitude of the nerve action potential of the sural nerve and the value at baseline was r = -0.26 ( P < 0.05). The changes in vibration threshold correlated with the baseline value (r = 0.24, P < 0.05), while there was
Table 3. Nerve conduction velocities in the tibial. peroneal and sural nerves, amplitude of the nerve action potential of the sural nerve, vibration threshold and thermal threshold (neutral zone interval between threshold for perception of warmth and cold) in the intensified conventional treatment (ICT) and regular treatment (RT) groups at entry to the study and after 5 years RT
ICT Entry
5 years
Entry
5 years
Nerve conduction velocity (ms-') Tibia1 Peroneal Sural
41.3k0.8 43.0+0.7 44.2f1.3
42.1 k 0 . 6 t 42.8k0.6t 40.3+ 1.8'
40.4k0.7 42.1 k0.7 42.6k1.7
37.7k0.8' 39.3k0.7' 36.5+2.0*
Amplitude of the nerve action potential of the sural nerve (pV)
9.9k0.7
*
9.0 1 .Ot
+
9.1 k0.8
7.8
Vibration threshold (pm)
1.2 0.2
2.7k1.0'
1.5k0.3
2.8 *0.5*
Temperature threshold ("C)
5.9 0.3
+
7.5k0.6.'
6.7+0.5
8.4k0.8**
*P < 0.01, **P < 0.05. tNS. Mean values k SEM are shown.
1.1..
106
P. REICHARD et al.
Table 4. Autonomic nerve function in the intensified conventional treatment (ICT) and regular treatment (RT) groups at entry to the study and after 5 years RT
ICT
Entry
5 years
Entry
Valsalva ratio (heart rate ratio)
2.5k0.5
1.9kO.lt
1.9 0.0
1.8 f0.1*
RSA ( A heart rate)
19.0f1.2
16.721.4;
17.0k1.3
12.1+1.2**
Dive test ( A heart rate)
-28f2
-2Of2"
-24k2
-23+2t
Blood flow (% blood flow)
158+6
131*5**
162+7
139 f4.'
Orthostatic test ( A blood pressure)
0.7k 2.3
-
-4.9f2.4
-2.4k 3.2t
1.4k3 . l t
+
5 years
*P < 0.05, **P < 0.01. tNS. Mean values fSEM are shown.
a negative correlation between the changes in thermal threshold and the value at entry to the study ( r = -0.33, P < 0.05).
Autonomic neuropathy Both groups tended to show a deterioration in autonomic nerve function (Table 4). The RSA was significantly lower in the RT group than in the ICT group after 5 years ( P < 0.01). There was a negative correlation between the change in RSA and the initial RSA value ( r = - 0.3 7, P < 0.05).
Hypogl y caem ia During the entire study a total of 34 ICT patients (77%, 9 5 % confidence interval 6 5 4 9 % ) and 29 RT patients (56%, CI 43-69%) exhibited at least one serious hypoglycaemic episode (P < 0.05). There were 242 episodes of serious hypoglycaemia altogether in the ICT group, and 98 episodes in the RT group.
Mortality Four ICT patients died during the first 36 months, as previously reported [15]. One KT patient died of a myocardial infarction that occurred between 36 and 60 months after the start of the study. The patients who died did not differ from the rest with regard to duration of diabetes or progression of microvascular complications during the first 18 months.
Multivariate analyses The increases in retinopathy (OR = 2.2. CI = 1.33.7, P < 0.01). nephropathy (OR = 2.3, CT = 1.33.4, P < 0.01) and neuropathy (OR = 3.5, CI = 1.7-7.0. P < 0.001) were all related to the mean HbA,, value during the study, but not to the initial HbA,, level. the duration of diabetes or the diastolic blood pressure at entry to the study. Smoking habits were related to the progession of retinopathy (OR = 3.0, CI = 1.2-7.9, P < 0.05), but not to the other complications.
Discussion Body weight The mean BMI increased from 2 2 . 5 f 0 . 3 to 23.8 f 0 . 4 kg m-' in the ICT group (P < 0.01), but there was no increase in the RT group (22.8 0.3 kg m-' at entry and 22.8 f0.3 kg m-' after 5 years). The mean weight gain in the ICT group was 4.1 kg per patient.
The results of the retrospective, cross-sectional and non-randomized prospective studies [32-3 51 suggest that there is a relationship between metabolic control and late complications. We have shown that IDDM patients with unsatisfactory metabolic control can maintain a lower blood glucose level for a long period without insulin
INTENSIFIED CONVENTIONAL INSULIN TREATMENT
pumps. For the first time we can demonstrate that lower blood glucose levels lead to a decrease in the progression of retinopathy. nephropathy and peripheral neuropathy. Retinopathy increased to slightly higher levels in the RT group compared to the ICT group. This has not been demonstrated previously [3, 4, 9, 101, but the present study is larger than the previous ones, and it is also of longer duration. Most of the differences between the groups were within the nonproliferative range. The urinary albumin excretion (UAER) is influenced by dietary protein intake [36] and blood pressure [37]. In our treatment groups there were no significant differences at baseline with regard to these factors. None of the groups showed significant changes during the study (for protein intake the analysis was made at the time of entry and after 3 years 1141. More RT patients than ICT patients developed neuropathy, defined as at least two pathological neurophysiological values. Such use of two abnormal measurements when defining neuropathy was suggested by Dyck et al. [24], who, however, employed a somewhat different set of investigations. Mortality was not correlated with treatment regimen as far as could be established when the individual patients were analysed [ 151. The number of patients in this study is too small to allow a detailed statistical analysis. The deceased ICT patients did not differ from the others with regard to the progression of complications, and thus their exclusion did not affect the comparison between the two groups. The frequency of serious hypoglycaemia was higher in the ICT group. As described previously 1151, hypoglycaemia could not be excluded as the cause of death in one patient. This individual had experienced frequent serious hypoglycaemic episodes before entry to the study, mostly after alcohol consumption. The patient’s HbA,, levels had not reached < 8 % ; thus treatment had not been very intensive. Body weight increased in the ICT group but not in the RT group. We have already shown after 3 years that this was related to the total energy intake [15]. When making multiple comparisons between two treatment groups there is always a risk of statistical type 1 errors. However, in this study not one comparison, but most of them, pointed toward a favourable effect of intensified treatment. This tendency was also consistent throughout the study. These effects could not be due merely to chance.
107
At baseline there were no significant differences with regard to microvascular complications. However, there were small differences, and they generally favoured the ICT group. We have found no explanation for this observation. Post-pubertal diabetes duration has been shown to correlate with the development of retinopathy and nephropathy [38], but it did not differ between the groups. In general there were no significant positive correlations between progression of complications and baseline values. Thus the small differences between the groups on entry to the study cannot explain the long-term differences with regard to microvascular complications. The multivariate analyses clearly indicate that the mean HbA,, value during the study was related to the development of each of the three main microvascular complications. Among the patients who participated in this study the risks of deterioration with regard to microvascular complications increased approximately 2- to 3.5-fold for each oneunit increase in the mean HbA,,, value. Smoking clearly influenced the progression of retinopathy. The fact that subclinical retinopathy, nephropathy and neuropathy were affected by the lower blood glucose levels indicates that the long-term prognosis will probably be improved as well, but this has not yet been confirmed. In particular, the findings with regard to nephropathy, where the natural progression from microalbuminuria to manifest nephropathy and then renal insufficiency is well known, are promising.
Acknowledgements This study was supported by grants from the Swedish division of NOVO-Nordisk Inc., the Boehringer Mannheim Scand. Inc., and the Swedish Medical Research Council (06615).
References 1 The DCCT Research Group. Are continuing studies of metabolic control and microvascular complications in insulindependent diabetes mellitus justified? N Engl / Med 1988: 31 8 : 246-50. 2 Dahl-Jorgensen K. Hanssen KF. Keirulf P. Bjoro T. Sandvik L, Aagenaes 0. Reduction of urinary albumin excretion after 4 years of continuous subcutaneous insulin infusion in insulindependent diabetes mellitus. The Oslo Study. Acta Endocrinol (Copenh) 1988: 117: 19-25. 3 Lauritzen T. Frost-Larsen K. Larsen H-W. Deckert T. the Steno Study Group. Efkct of 1 year of near-normal blood glucose levels on retinopathy in insulin-dependent diabetes. I ~ n c e t 1983; 1 : 2 0 0 4 .
108
P. R E I C H A R D
et al.
4 The KROC Collaborative Study Group. Blood glucose control and the evolution of diabetic retinopathy and albuminuria. N Engl ] Med 1984: 2 9 0 : 365-72. 5 Feldt-Rasmussen B. Mathiesen ER. Deckert T. Etrect of two years of strict metabolic control on progression of incipient nephropathy in insulin-dependent diabetes. Lancet 1 9 8 6 : 11 : 13004. 6 Dahl-Jorgensen K, Brinchmann-Hansen 0. Hanssen KF et a!. Effect of near normoglycemia for two years on progression of early diabetic retinopathy. nephropathy and neuropathy : the Oslo study. BM] 1 9 8 6 : 293: 1195-9. 7 Service FJ. Rizza RA. Daube JR. O'Brien PCO. Uyck PJ. Near normoglycaemia improved nerve conduction and vibration sensation in diabetic neuropathy. Diabetologia 1985: 2 8 : 722-7. 8 Dahl-Jorgensen K, Brinchmann-Hansen 0. Hanssen KF. Sandvik L. Aagenaes 0, Aker Diabetes group. Rapid tightening of blood glucose control leads to transient deterioration of retinopathy in insulin-dependent diabetes mellitus: the Oslo study. BM] 1 9 8 5 ; 2 9 0 : 811-5. 9 Brinchrnann-Hansen 0, Dahl-Jorgensen K. Hanssen KF. Sandvik L. The response of diabetic retinopathy to 41 months of multiple insulin injections, insulin pumps and conventional insulin therapy. Arch Ophthalmol 1 9 8 8 : 106: 1242-6. 10 Olsen T, Richelsen B, Ehlers N. Beck-Nielsen H. Diabetic retinopathy after 3 years' treatment with continuous subcutaneous insulin infusion (CSII). Acta Ophthalmol (Copenh) 1 9 8 7 : 65: 185-9. 11 The DCCT Research Group. Diabetes Control and Complications Trial (DCCT): results of feasibility study. Diabetes Care 1 9 8 7 : 10: 1-19. 1 2 The DCCT Research Group. Weight gain associated with intensive therapy in the Diabetes Control and Complications Trial. Diabetes Care 1 9 8 8 ; 1 1 : 567-73. 13 Reichard P. Britz A, Cars I. Nilsson BY, Sobocinsky-Olsson B. Rosenqvist U. The Stockholm Diabetes Intervention Study (SDIS): 18 months' results. Acta Med Scand 1988: 224: 1 1 5-22. 14 Reichard P. Kosenqvist U. Nephropathy is delayed by intensitied insulin treatment in patients with insulin-dependent diabetes mellitus and retinopathy. / Intern Med 1 9 8 9 : 226: 81-7. 1 5 Reichard P. Britz A. Carlsson P et al. Metabolic control and complications over 3 years in patients with insulin-dependent diabetes mellitus (IDDM):the Stockholm Diabetes Intervention Study (SDIS). I Intern Med 1 9 9 0 ; 228: 511-7. 1 6 Reichard P. Berglund B. Britz A, kvander S. Rosenqvist U. Hypoglycemic episodes during three years of intensified conventional insulin treatment: increased frequency but no effects on cognitive function. Intern Med 1991 : 229: 9-16. 17 Early Treatment Diabetic Retinopathy Study. Manual of Operations. Bethesda. Maryland : Public Health Service, 1980. 1 8 Chantler C. Garnett ES. Parsons V. Veal1 N. Glomerular filtration rate measurement in man by the single injection method using "Cr-EDTA. Clin Sci 1 9 6 9 : 37: 169-80. 1 9 Mogensen CE, Chachati A, Christensen CK et al. Microalbuminuria: a n early marker of renal involvement in diabetes. Uremia Znvest 1985-86: 9 : 85-95. 2 0 Feldt-Rasmussen B. Mathiesen ER. Variability of urinary albumin excretion in incipient diabetic nephropathy. Diab Nephropathy 1 9 8 4 : 3: 101-3. 21 Ludin H-P. Electrornyography in Practice. Stuttgart. New York: Georg Thieme Verlag. Thieme-Stratton Inc., 1 9 8 0 ; 3 1 4 2 .
2 2 Goldberg ]M. Lindblom U. Standardised method of determining vibratory perception thresholds for diagnosis and screening in neurological investigation. ] Neurol Neurosurg Psychiatry 1 9 7 9 ; 4 2 : 793-803. 2 3 Fruhstorfer H. Lindblom U. Schmidt WG. Method for quantitative estimation of thermal thresholds in patients. / Neurol Neurosurg Psychiatry 1 9 7 6 : 39: 1071-5. 2 4 Dyck PJ. Karnes J. O'Brien PC. Diagnosis, staging, and classification of diabetic neuropathy and associations with other complications. In: Dyck PJ. Thomas PK. Asbury AK. Winegrad AI. Porte D. eds. Diabetic Neuropathy. Philadelphia : W.B. Saunders Company, 1987: 36-44. 25 Levin AV. A simple test of cardiac function based upon heart rate changes induced by the Valsalva manoeuvre. A m ] Cardiol 1966: 1 8 : 90-9. 2 6 Freyschuss U. Melcher A. Sinus arrhythmia in man: influence of tidal volume and oesophageal pressure. Scand / Clin I*lb Invest 1 9 7 5 : 35: 487-96. 2 7 Bennett T. Hosking DJ, Hampton JR. Cardiovascular responses to apnoic face immersion and mental stress in diabetic subjects. Curdiovasc Res 1976: 1 0 : 192-9. 2 8 Eklund B. Kaijser L. Knutsson E. Blood flow in resting (contralateral) arm and leg during isometric contraction. ] Physiol (Lond) 1 9 7 4 : 240: 11 1-24. 29 Burke D. Sundlof G. Wallin BG. Postural effects on muscle nerve sympathetic activity in man. 1 Physiol (Lond) 1977: 272: 3 9 9 4 1 4 . 30 Conover WJ. Practical Non-l~arametricstatistics. 2nd edn. New York: John Wiley and Sons. 1980: 99-105. 31 Breslow NE. Day NE. Statistical Methods in Cancer Research. Vol. 1. Lyon: IARC Scientific Publications No. 32. 1 9 8 0 : 196-246. 32 Pirart 1. Diabetes mellitus and its degenerative complications: a prospective study of 4 4 0 0 patients observed between 1947 and 1973. Diabetes Cure 1 9 7 8 : 1 : 168-88. 3 3 Klein R. Klein BEK. Moss SE. Davis MU. DeMets D. The Wisconsin epidemiologic study of diabetic retinopathy. 11. Prevalence and risk of diabetic retinopathy when age at diagnosis is less than 3 0 years. Arch Ophthalmol 1 9 8 4 : 102: 520-6. 34 Rosenstock J. Friberg T.Raskin P. Effect of glycemic control on microvascular complications in patients with type 1 diabetes mellitus. Am 1 Med 1986: 81: 1012-8. 35 McCance DR. Atkinson AB, Hadden DR. Archer DB. Kennedy L. Long-term glycaemic control and diabetic retinopathy. Lancet 1989: 1 1 : 824-8. 36 Cohen D, Dodds R, Viberti G. Effect of protein restriction in insulin dependent diabetics at risk of nephropathy. BM] 1987; 2 9 4 : 795-8. 37 Mogensen CE. Long-term antihypertensive treatment inhibiting progression of diabetic nephropathy. BM] 1982 : 2 8 5 : 685-8. 38 Norris Kostraba J. Dorman JS. Orchard TJ et al. Contribution of diabetes duration before puberty to development of microvascular complications in IDDM subjects. Diabetes Care 1989 ; 1 2 : 686-93. Received 2 7 November 1990, accepted 8 January 1991.
Correspondence: Dr Per Reichard. Department of Internal Medicine 11. Sodersjukhuset, S-100 6 4 Stockholm, Sweden.
