- Email: [email protected]
Effect of glycemic control on microvascular complications in patients with type I diabetes mellitus
Effect of Glycemic Control on Microvascular Complications in Patients with Type I Diabetes Mellitus
JULIO ROSENSTOCK, M.D. THOMAS FRIBERG, M.D. PHILIP RASKIN, M.D.
The relation between the control of blood glucose levels and the progression of early diabetic retinopathy and the width of skeletal muscle capillary basement membrane was studied in 54 insulindependent diabetic patients. After initial ophthalmologic evaluation including seven-field fundus photography and fluorescein angiography and measurement of levels of glycosylated hemoglobin and width of skeletal muscle capillary basement membrane, the patients were divided into two groups: an experimental group of 30 patients who were treated with continuous subcutaneous insulin infusion and a control group of 24 patients who continued to receive conventional treatment-usually two injections of insulin daily. After a mean followup period of 31.4 months, the experimental group had a significant decrease in glycosylated hemoglobin levels as compared with baseline values (mean f SEM, 7.2 f 0.3 percent versus 10.1 f 0.4 percent), reflecting improved control of blood glucose levels. The conventional treatment group had no change in glycosylated hemoglobin levels after a mean of 33.5 months of follow-up. Wiih use of either a modified Early Treatment Diabetic Retinopathy Study grading system or macular microaneurysm counts, the experimental treatment group showed significantly less progression of retinopathy (p <0.05). The skeletal muscle capillary basement membrane width was significantly reduced only in the experimental treatment group with stable or improved retinopathy and was unchanged In the control group. There was a tendency for skeletal muscle capillary basement membrane width to increase in thickness over time in those patients whose retinopathy worsened irrespective of treatment. It is concluded that meticulous diabetic control may slow the progression of early diabetic retinopathy. Changes in skeletal muscle capillary basement membrane width may reflect the course of diabetic retinopathy.
Dallas, Texas
From the University of Texas Health Science Center at Dallas, Southwestern Medical School, Dallas, Texas. This work was supported in part by grants (l-MOl-RR-0063 and AM-07307) from the National Institutes of Health and from the Juvenile Diabetes Foundation, and was presented in part at the 45th Annual Meeting of the American Diabetes Association, Baltimore, Maryland, June 1985. Requests for reprints should be addressed to Dr. Philip Raskin, Department of Internal Medicine, University of Texas Health Science Center at Dallas, 5323 Harry Hines Boulevard, Dallas, Texas 75235. Manuscript submitted December 13, 1985, and accepted February 11, 1986.
1012
December
1986
The American
Journal
The relationship between the microvascular complications of diabetes mellitus and blood glucose control is unknown [I]. Although this issue has been the subject of a long-standing controversy [2, 31, the failure to resolve this issue has been due to the inability to achieve long-term normalization of blood glucose levels in most diabetic patients. The recent development of more effective treatment modalities [4] as well as the rapidly expanding use of self-monitoring of blood glucose levels [E] has now made it possible to achieve near-normoglycemia for long periods of time. Thus, it is now feasible to study whether or not long-term euglycemia can diminish the frequency or severity of diabetic microvascular complications in patients with insulin-dependent diabetes. This study was designed to evaluate the relationship between the level
of Medicine
Volume
81
EFFECT
OF GLYCEMIC
CONTROL
ON MICROVASCULAR
of diabetes control and its effect on the progression of early diabetic retinopathy and on the width of skeletal capillary basement membrane in patients with type I diabetes. Our data show that in insulin-dependent diabetic patients followed for up to five years in a prospective fashion, maintenance of near-normal levels of glycemia is associated with significantly less progression of early diabetic retinopathy. Furthermore, a significant reduction of the capillary basement membrane width was found only in patients with near-normoglycemia who showed improved or stable retinopathy. PATIENTS
AND
TABLE
COMPLICATIONS
I
Clinical
Characteristics
Conventional Treatment Group (n = 24) Age (years) Duration (wars) Follow-up (months)
27.1
f 1.7
11.9 f 1.3 33.5
IN DIABETES-ROSENSTOCK
f 3.7
of Study Groups
Experimental Treatment Group (n = 30) 27.6
ET AL
f 1.5
Nondiabetic Group (n= 26) f 1.1
28.1
13.8 f 1.2
-
f 2.5
-
31.4
METHODS
Fifty-four patients with insulin-dependent type I diabetes mellitus were asked to participate in an intensive diabetes treatment program. Patients with a prior history of laser photocoagulation treatment were excluded. Twenty-four of the patients decided to remain in their existing conventional treatment program. The clinical characteristics of these 24 patients who continued in the conventional treatment group are given in Table 1. In most circumstances, diabetes treatment in this group consisted of no more than two daily injections of insulin. Some patients received a mixture of intermediateand short-acting insulin, others only intermediate-acting insulin. All were given instruction on the principles of the dietary treatment of diabetes and placed on a diet as recommended by the American Diabetes Association [6] with sufficient calories to maintain their ideal body weight. A small number of this conventionally treated group chose to occasionally monitor their blood glucose levels as part of their overall diabetes treatment group. Many preferred urine testing for glucose and acetone, however. The average age (ZIZSEM) of this group was 27 f 1.7 years and they had diabetes for an average of 11.9 f 1.3 years. Thirty patients participated in the experimental treatment program that included intensive diabetes and dietary education. The dietary management was based on principles recommended by the American Diabetes Association [6]. Self-monitoring of blood glucose levels was performed systematically several times during each day, and patients were taught to use the test results with individual algorithms for self-insulin adjustments. Insulin was given as a continuous subcutaneous insulin infusion via a portable insulin infusion device. (The devices used during the duration of this study were the Autosyringe ASPC, ASGC, ASGC-UlOO, ASGCMP, and CPI-9100, and the Pacesetter Miniseries). The clinical characteristics of the patients in the experimental treatment group are listed in Table I. This group was identical to the conventional treatment group with respect to age (27.6 f 1.5 years) and duration of diabetes (13.6 f 1.2 years). Procedures for the orientation to our program of treatment with portable infusion devices and for follow-up have been previously described [7,6]. There were no differences between the two groups in terms of the level of arterial blood pressure and the number of patients who smoked. All patients were studied in the General Clinical Research Unit at Parkland Memorial Hospital, Dallas, Texas, at baseline and at six-month intervals thereafter. Initial and
follow-up studies included determinations of levels of total glycosylated hemoglobin, a complete ophthalmologic evaluation, and measurement of the width of skeletal muscle capillary basement membrane. This latter assesment was made at 12-month intervals. Results for the diabetic patients were compared with those of 43 nondiabetic control subjects in whom levels of total glycosylated hemoglobin were measured. Twenty-six of these nondiabetic subjects also volunteered for skeletal muscle biopsy for measurement of the width of their skeletal muscle capillary basement membrane. None had a family history of diabetes. The clinical characteristics of these subjects are given in Table I. Their age averaged 26.1 f 1.1 years. The ophthalmologic evaluation included assessment of best-corrected visual acuity, motility, slit-lamp examinations, and direct and indirect ophthalmoscopy performed after pupillary dilation. Seven-field fundus photography and fluorescein angiography were performed at each six-month visit using a Zeiss fundus camera, which had a field of view of 30 degrees. Prior to grading, the fundus photographs and fluorescein angiograms for a particular visit from each patient were placed in a numbered folder. Photographs were arranged at random and not chronologic sequence. A numeric code was used and the patient’s names were omitted or masked on the photographs. Grading of all photographs was performed separately by two retinal specialists. At the time of grading, neither grader knew the patient’s treatment group assignment. Grading was accomplished by viewing all of the color photographs and fluorescein angiograms with magnification and utilizing stereo pairs wherever possible. In a few cases, color photographs alone were used to evaluate the retinal status if the patient could not tolerate fluorescein because of allergy or untoward reactions, or if the quality of the fluorescein angiograms was inadequate. Microaneurysms were counted only on fluorescein angiograms except in three isolated cases in which fluorescein angiograms were not obtained (two experimental treatment group patients and one conventional treatment group patient). The data were then entered into a computer and the grades for each eye were plotted against time for each patient. Two grading systems were used to categorize the retinopathy of each eye of each patient at each visit. First we used the modified Early Treatment Diabetic Retinopathy Study (ETDRS) system with combined assessment of fun-
December 1986
The
American Journal of Medicine
Volume
81
1013
EFFECT
OF GLYCEMIC
CONTROL
ON MICROVASCULAR
EXPERIMENTAL Ff(
8
::E 2 o
7
IN DIABETES---R~SENST~GK
the identification of the sample would be unknown to the personnel involved with final processing and measurements. The basement membrane width was measured according to the method of Shannon et al [ 131, which is a computer-assisted method of measuring biologic struc-
CONVENTIONAL l$
The total glycosylated hemoglobin level (Ala+~+J was determined by ion exchange chromatography with the use of prepacked microcolumns (Quik-column, Helena Laboratories, Beaumont, Texas) after a thrice washing of the erythrocytes and incubation overnight at 23OC to remove the labile fraction. All total glycosylated hemoglobin levels obtained at visits subsequent to the baseline visit were averaged and the mean glycosylated hemoglobin value (or follow-up value) was compared with that of the baseline. Likewise, all skeletal muscle capillary basement membrane measurements made at visits subsequent to the baseline visit were averaged and this mean value (follow-up value) was compared with the baseline one. The glycosylated hemoglobin levels and capillary basement membrane measurements were compared using the Student t test for paired observations for comparison within groups and for unpaired observations for comparison between groups [ 141. Results of the microaneurysm counts and the modified ETDRS grades were compared by log-likehood chisquared analyses and Fisher’s exact test [ 141. Tests were considered significant at the 0.05 level.
**=p<~ool~vs~
6 Baseline Followup
Non Diabetics
Baseline Followup
Figure 1. Giycosylated hemoglobin levels at baseline and follow-up in patients in the experimental treatment group and in the conventional treatment group. Values in nondiabetic volunteers are included for reference.
dus photographs and fluorescein angiograms. We used the Puklin [9] modification of this grading system that creates an additional category (grade II) for minimal retinopathy. We studied the fluorescein angiograms in conjunction with the color photographs to determine areas of neovascularization, intraretinal microvascular abnormalities, and macular edema. Using this modified ETDRS system, each eye at each particular time point was further categorized by indexing the worst eye and describing the better eye according to the system advanced by Klein et al [IO]. The order of progression within the retinopathy categories was thus (1,i) (2,<2) (2,2) (3,<3) (3,3), etc. A patient was considered worse or better if the eyes changed by two or more of these categories over time. Statistical calculations were performed on the last reading that always coincided with the general trend observed in the plots over time. The second grading system used was the macular microaneurysm count. From the fluorescein angiograms, the ETDRS standard 30-degree macular field was used for counting microaneurysms. The patient’s retinopathy was arbitrarily considered worse or better if the total number of microaneurysms seen in both eyes added together increased or decreased by 20 or more over the duration of the study. If the results of the final grading between the two graders conflicted as to whether a particular patient’s condition worsened or remained stable, the photographs of these patients were assembled by a helper and re-evaluated by both readers together until a consensus was reached. This was also done without the ophthalmologists’ knowledge to which treatment group the patient belonged. The skeletal muscle biopsy specimen was taken from the right quadriceps muscle group under local anesthesia according to the method originally described by Siperstein et al [ 1 l]. The muscle samples were fixed in Palade’s [ 121 fixative and processed by the methods of Siperstein et al [ 1 I]. The samples were sent to the electron microscopy laboratory with consecutive numeric designations so that
1014
ET AL
tures.
2 11 2 0” 10 2E 9 g
COMPLICATIONS
December
1986
The American
Journal
of Medicine
RESULTS The average glycosylated hemoglobin value for the 43 nondiabetic subjects was 6.1 f 0.4 percent. This value was significantly different from those of both treatment groups at all time points. The average width of the skeletal muscle capillary basement membrane in the 26 nondiabetic subjects, inOwhom this measurement was made, was 1,094 f 41 A. In the nondiabetic control group, only a single determination of glycosylated hemoglobin and basement membrane width was made. At baseline, the mean glycosylated hemoglobin level for the patients in the conventional treatment group was 10.6 f 0.3 percent (Figure 1). It remained essentially unchanged over the average 33.5 f 3.7 months of follow-up and averaged 10.4 f 0.3 percent. The baseline glycosylated hemoglobin level in the experimental treat-
ment group was IO. 1 f 0.4 percent, a value that was not different from that in those patients treated by conventional therapy (Figure 1). Following a mean of 31.4 f 2.5 months of the experimental treatment, the glyCOSylated hemoglobin level fell to an average value of 7.2 f 0.3
percent, p
value
in the experimental
treatment
group
attests to the effectiveness of the experimental program. The severity of the retinopathy at entry in both study groups was the same (Table II). Using the modified ETDRS grading system, there was a preponderance of patients with very minimal or no retinopathy. In addition, although there were no statistical differences between the
Volume
81
EFFECT
OF GLYCEMIC
CONTROL
ON MICROVASCULAR
two groups at entry, the experimental treatment group had a few more patients with more severe retinopathy. This was true irrespective of whether the ETDRS grading system or the macular microaneurysm counts were used (data not shown). The results of the effect of diabetes control on the progression of retinopathy are shown in Table III. In terms of the modified ETDRS grading system, in the conventional treatment group, the condition of 18 patients remained stable, the condition of six progressed, and none had improvement. In the experimental treatment group, the condition of 23 patients remained stable, the condition of three progressed, and four showed improvement. With respect to the macular microaneurysm counts in the conventional treatment group, the condition of 19 remained stable, the condition of five worsened, and none showed improvement. In the experimental treatment group, the condition of 28 remained stable, the condition of one worsened, and one showed improvement. Using either grading systems, there were significant differences between the results from the two treatment groups. The, ETDRS grading showed a significant relationship between diabetes control and the progression of diabetic retinopathy as defined by the categories better, stable, or worse (p <0.05). In terms of macular microaneurysm counts (since only one patient in the experimental group improved), only the categories stable and worse were analyzed and again both treatment groups were significantly different (p <0.05). The width of skeletal muscle capillary basement membrane in the conyentional treatment group at baseline was 1,675 f 750 A, a value identical to that of the experimental treatment group, 1,617 f 91 A (Figure 2). Both baseline values were greater than in the nondiabetic group (p
COMPLICATIONS
TABLE II
342
Figure 2. Skeletal muscle capil/aIy basement membrane width at baseline and follow-up in patients in the experimental treatment group and in the co* ventional treatment group. l l =p < 0.05 versus baseline.
ET AL
Degree of Retinopathy at Study Entry (Modified Early Treatment Diabetic Retinopathy Study Grading) Conventional Treatment Group
(l,l)to (4,X4) (6,<6) (88) to
TABLE Ill
Effect of Diabetes of Retinopathy Modilied Early Treatment Diabetic Retlnopathy Study Experimental” Conventional Treatment Treatment
Better Stable Worse
Experimental Treatment Group
22 2 0 0
(3,3) to (575) to (8,<8) (10,lO)
4 23 3
21 6 3 0
Control
on Progression
Macular Microaneurysm Count Experimental* Conventional Treatment Treatment
0 18 6
* p <0.05, experimental (log-likelihood chi-square).
1 28 1
treatment
versus
0 19 5
conventional
treatment
brane wktth in the conventional treatment group after 33.5 f 3.7 months of follow-up was 1,794 f 100 A, a value that was identical to the baseline value in this treatment group. However, in the experimental treatment group, the average follow-up basement membrane width was 1,417 f 50 A, a value that was significantly less (p (0.005) than the baseline value and significantly different from the follow-up value in the conventional treatment group (p
ExDerimental
-
IN DIABETES-ROSENSTOCK
Conventional
Rx
Rx
2000
2000 1800
“5
1800
.$
1600
g
1600
$
1400
3Q 5 1400
2
1200 1000
i
2E 3E
800
24 1000 E 2 800
z
600
3 Baseline
December
1986
1200
600 Baseline
Followup
The American
Journal
of Medicine
Volume
Followup
81
1015
EFFECT
OF GLYCEMIC
CONTROL
ON MICROVASCULAR
COMPLICATIONS
EXPERIMENTAL F$ RETINOPATHY Better Stable (4) / (23)
IN DIABETES-ROSENSTOCK
CONVENTIONAL s, RETINOPATHY
Worse (3)
Better (0) / Baseline Followup +p
Stable (18)
Worse (8)
0
i f 2000 P 2 1800 22. 1600 g 1400 +g 1200 L+i1000 m”
q
q
VSH
Figure 3. Skeletal muscle capillary basement membrane width at baseline and follow-up in patients in the experEmental treatment group and in the colt ventional treatment group as a function of the status of their diabetic retinopathy.
800
given in Figure 3. This figure shows that the 27 patients in the experimental treatment group whose retinopathy remained stable or improved had a significant reduction of basement membrane width from 1,636 f 99 A at baseline to 1,404 f 54 A at follow-up. Those 18 patients in the conventional treatment group whose retinopathy remained stable (none improved) had identical basement membrane width at baseline and at follow-up (1,627 f 120 A and 1,704 f 93 A, respectively). In both treatment groups, in those patients whose retinopathy worsened, there was a tendency for the skeletal muscle capillary basement membrane width to increase in thickness, although no statistically significant relationship was appreciated. This was probably due to the small sample size.
These data, although exciting, contrast to some extent with other studies performed to date. Specifically, we make reference to the Steno [ 141 and KROC [ 151 studies. In these prospective, randomized, but shorter-duration clinical trials, there was actually a tendency towards progression of diabetic retinopathy in those patients who had the best diabetes control. The progression of retinopathy in these groups was usually the result of the appearance of cotton-wool spots. With longer follow-up, these cotton wool spots tend to regress [ 161. The precise reason why our data differ from these is not entirely clear, but the longer duration of follow-up at six-month rather than shorter intervals and the fact that all of our patients had early diabetic retinopathy are two explanations that seem readily apparent. Our finding that long-term normoglycemia causes a reduction in skeletal muscle capillary basement membrane widths confirms and extends our previous report [ 171. These data show that the reduction in skeletal capillary basement membrane width that occurs after two years of therapy [ 171 continues with prolonged follow-up. Of interest in this study is the association between the improvement (i.e., reduction in thickness) seen in skeletal muscle capillary basement membrane width and clinical microangiopathy. One of the major issues regarding the relevance of measurements of skeletal capillary basement membrane width is the fact that it has been difficult in the past to correlate those measurements with clinical microangiopathy [ 18,191. In this study, those patients in the experimental treatment group whose retinopathy improved or remained stable had a significant reduction in their skeletal muscle basement membrane width. Patients whose diabetic retinopathy progressed, irrespective of treatment, tended to have an increase in the width of the basement membrane in the capillaries of their skeletal muscle.
COMMENTS Our data show that in patients with type I diabetes mellitus, there appears to be a significant effect of improved glycemic control, using an experimental treatment program, on the small blood vessel complications of diabetes. Clearly, patients in our experimental treatment group had long-term improvement of blood glucose control at normal or near-normal levels as evidenced by significant falls in levels of their total glycosylated hemoglobin. In addition, it should be emphasized that in our experimental treatment group, the diabetic retinopathy exhibited by these patients showed significantly less deterioration over time, i.e., the retinopathy was either stable or improved more often in the experimental treatment group than in the conventional treatment group, where the retinopathy tended to worsen. Thus, in our study, there appears to be a significant advantage in terms of the slowing of the progression of early diabetic retinopathy in patients undergoing an experimental treatment program that resulted in near-normal glycemia for periods of up to five years.
1016
December
ET AL
1986
The American
Journal
of Medicine
Volume
81
EFFECT
OF GLYCEMIC
CONTROL
ON MICROVASCULAR
In this regard, it is important to note that others have been unable to correlate skeletal muscle capillary basement membrane width to the presence or absence of diabetic nephropathy [20]. There are several possible explanations. First of all, skeletal muscle capillary basement membrane thickness may be more indicative of retinopathy than of nephropathy. In addition, the studies on nephropathy were cross-sectional in nature and not longitudinal as is ours. Furthermore, they used a very selective and heterogeneous group of patients, those being evaluated for possible pancreas transplantation, with different stages of microvascular complications. Finally, they fixed their muscle biopsy specimens in gluteraldehyde, a fixative that has been shown to produce better photographs, but to be less specific in documenting diabetic microangiopathy [2 11. Several factors relating to the design of this study and interpretation of the results should be discussed. First and most imortant, this was a prospective study, although patients were not randomly assigned to a treatment group. All patients in the study were offered enrollment in the experimental treatment program. It is possible that those patients who were more interested in improving their level of diabetic control and who believed that this could be important to their overall well-being would be most likely to choose the experimental treatment. It is also possible that there were some inherent differences between the two groups that were not apparent in biologic measures. Clearly, these subjects were similar in age, duration of diabetes, arterial blood pressure, and the number who
COMPLICATIONS
IN DIABETES-ROSENSTOCK
ET AL
smoked. At entry into the study, both treatment groups had identical values for the width of skeletal muscle capillary basement membranes and for glycosylated hemoglobin. The results of this study must be viewed in the context of the entire clinical spectrum of insulin-dependent diabetes mellitus. Although our data, for the first time, indicate an advantage of improved diabetic control in terms of the progression of diabetic retinopathy, we cannot at present recommend this type of intensive treatment to all patients with insulin-dependent diabetes. This study certainly provides substantial data that need to be verified by a larger prospective randomized clinical trial such as the Diabetes Control and Complications Trial [22] that is presently underway. ACKNOWLEDGMENT We wish to thank Patricia Challis, R.N., B.S.N., Suzanne Strowig, B.S.N., M.S.N., and Susan Cercone, RD., M.S. for their invaluable help; George Sanborn, M.D., for helping with the grading of the retinal photographs and fluorescein angiograms; Lisa Gagliano, Arthur Ojirika, and Kevin Sullivan for their technical assistance; Debra Elkins and William Anderson for the retinal photographs; Bette Newton for her assistance in the preparation of the manuscript; and the staff and nurses of the General Clinical Research Center for the care of our patients. Finally, our heartfelt thanks to our patients, without whose loyalty and dedication this study could not have been possible.
REFERENCES 1. 2. 3.
4.
5. 6.
7.
8.
9.
Raskin P: Diabetic regulation and its relationship to microangiopathy. Metabolism 1978; 2: 235-252. Cahill GF Jr, Etzwiler DD, Freinkel N: “Control” and diabetes. ?I Engl J Med 1976; 294: 1004-1005. Siperstein MD, Foster DW, Knowles HC Jr, Levine R, Madison LL, Roth J: Control of blood glucose and diabetic vascular disease. N Engl J Med 1977; 296: 1060-1063. Raskin P: Treatment of type I diabetes mellitus with portable insulin infusion devices. Diabetes Care 1982; 1: (suppl 5) 48-52. Skyler JS: Patient self-monitoring of blood glucose. Clin Diabetes 1983; 1: 12-17. American Diabetes Association: Principles of nutrition and dietary recommendations for ‘individuals with diabetes mellitus. Diabetes 1979; 28: 1027-1030. Raskin P, Pietri A, Unger R: Changes in glucagon levels after four to five weeks of glucoregulation by portable insulin infusion pumps. Diabetes 1979; 28: 1033-1035. Raskin P: Treatment of insulin-dependent diabetes mellitus with portable insulin infusion devices. Med Clinic North Am 1982; 66: 1269-1283. Puklin JE, Tamborlane WV, Felig P, Gene1 M, Sherwin RS: Influence of long-term insulin infusion pump treatment of
December
10.
11.
12. 13.
14.
15.
16.
1986
type I diabetes on diabetic retinopathy. Ophthalmology 1982; 89: 735-747. Klein, BE, Davis MD, Segal P, et al: Diabetic retinopathy. Assessment of severity and progression. Ophthalmology 1984; 91: 10-17. Siperstein MD, Unger RH, Madison LL: Studies of muscle capillary basement membranes in normal subjects, diabetic, and prediabetic patients. J Clin Invest 1968; 47: 1973-1999. Palade GE: A study of fixation for electron microscopy. J Exp Med 1952; 95: 285-298. Shannon WA Jr, Rockholt DL, Bates SB: Computer-assisted measurement of the thickness of biological structures. Comput Biol Med 1982; 12: 149-155. Lauritzen T, Frost-Larsen K, Deckert T, et al: Steno Study Group. Effect of 1 year of study of near-normal blood glucose levels on retinopathy in insulin-dependent diabetics. Lancet 1983; I: 200-204. KROC Collaborative Study Group: Blood glucose control and the evolution of diabetic retinopathy and albuminuria. A preliminary multicenter trial. N Engl J Med 1984; 311: 365-372. Lauritzen T, Frost-Larsen K, Larsen HW, et al: Steno Study
The American
Journal
of Medicine
Volume
81
1017
EFFECT
17.
18.
19.
1018
OF GLYCEMIC
CONTROL
ON MICROVASCULAR
COMPLICATIONS
Group. Continuous subcutaneous insulin infusion. Lancet 1983; I: 1445-1446. Raskin P, Pietri A, Unger. R, et al: The effect of diabetic control on skeletal muscle capillary basement membrane width in patients with type I diabetes mellitus. N Engl J Med 1983; 300: 1546-1560. Kilo 6, Volger N, Williamson JR: Muscle capillary basement membrane changes related to aging’and to diabetes mellitus. Diabetes 1972; 21: 881-905. Siperstein MD: Capillary basement membranes and diabetic microanglopathy. Adv-Intern Med 1972; 18: 325-344.
December
1986
The American
Journal
of Medicine
IN DIABETES-ROSENSTOCK
20.
Steffes MW, Sutherland DER, Goertz FC, Rich SS, Mauer SM: Studies of kidney andmuscle biopsy specimens from identical twins discordant for-type I diabetes mellitus. N Engl J Med 1985; 312: 1282-1286. Siperstein MD, Raskin P, Bum .H: Electron microscopic quantification of diabetic retinopathy. Diabetes 7973; 22: 514-527. . Proposed protocol for the clinical trial to assess the relationship between metabolic control and the early vascular complications of insulin-dependent diabetes mellitus. Diabetes ‘1982; 31: 1132-l 133.
21.
22.
Volume
ET AL
61
JULIO ROSENSTOCK, M.D. THOMAS FRIBERG, M.D. PHILIP RASKIN, M.D.
The relation between the control of blood glucose levels and the progression of early diabetic retinopathy and the width of skeletal muscle capillary basement membrane was studied in 54 insulindependent diabetic patients. After initial ophthalmologic evaluation including seven-field fundus photography and fluorescein angiography and measurement of levels of glycosylated hemoglobin and width of skeletal muscle capillary basement membrane, the patients were divided into two groups: an experimental group of 30 patients who were treated with continuous subcutaneous insulin infusion and a control group of 24 patients who continued to receive conventional treatment-usually two injections of insulin daily. After a mean followup period of 31.4 months, the experimental group had a significant decrease in glycosylated hemoglobin levels as compared with baseline values (mean f SEM, 7.2 f 0.3 percent versus 10.1 f 0.4 percent), reflecting improved control of blood glucose levels. The conventional treatment group had no change in glycosylated hemoglobin levels after a mean of 33.5 months of follow-up. Wiih use of either a modified Early Treatment Diabetic Retinopathy Study grading system or macular microaneurysm counts, the experimental treatment group showed significantly less progression of retinopathy (p <0.05). The skeletal muscle capillary basement membrane width was significantly reduced only in the experimental treatment group with stable or improved retinopathy and was unchanged In the control group. There was a tendency for skeletal muscle capillary basement membrane width to increase in thickness over time in those patients whose retinopathy worsened irrespective of treatment. It is concluded that meticulous diabetic control may slow the progression of early diabetic retinopathy. Changes in skeletal muscle capillary basement membrane width may reflect the course of diabetic retinopathy.
Dallas, Texas
From the University of Texas Health Science Center at Dallas, Southwestern Medical School, Dallas, Texas. This work was supported in part by grants (l-MOl-RR-0063 and AM-07307) from the National Institutes of Health and from the Juvenile Diabetes Foundation, and was presented in part at the 45th Annual Meeting of the American Diabetes Association, Baltimore, Maryland, June 1985. Requests for reprints should be addressed to Dr. Philip Raskin, Department of Internal Medicine, University of Texas Health Science Center at Dallas, 5323 Harry Hines Boulevard, Dallas, Texas 75235. Manuscript submitted December 13, 1985, and accepted February 11, 1986.
1012
December
1986
The American
Journal
The relationship between the microvascular complications of diabetes mellitus and blood glucose control is unknown [I]. Although this issue has been the subject of a long-standing controversy [2, 31, the failure to resolve this issue has been due to the inability to achieve long-term normalization of blood glucose levels in most diabetic patients. The recent development of more effective treatment modalities [4] as well as the rapidly expanding use of self-monitoring of blood glucose levels [E] has now made it possible to achieve near-normoglycemia for long periods of time. Thus, it is now feasible to study whether or not long-term euglycemia can diminish the frequency or severity of diabetic microvascular complications in patients with insulin-dependent diabetes. This study was designed to evaluate the relationship between the level
of Medicine
Volume
81
EFFECT
OF GLYCEMIC
CONTROL
ON MICROVASCULAR
of diabetes control and its effect on the progression of early diabetic retinopathy and on the width of skeletal capillary basement membrane in patients with type I diabetes. Our data show that in insulin-dependent diabetic patients followed for up to five years in a prospective fashion, maintenance of near-normal levels of glycemia is associated with significantly less progression of early diabetic retinopathy. Furthermore, a significant reduction of the capillary basement membrane width was found only in patients with near-normoglycemia who showed improved or stable retinopathy. PATIENTS
AND
TABLE
COMPLICATIONS
I
Clinical
Characteristics
Conventional Treatment Group (n = 24) Age (years) Duration (wars) Follow-up (months)
27.1
f 1.7
11.9 f 1.3 33.5
IN DIABETES-ROSENSTOCK
f 3.7
of Study Groups
Experimental Treatment Group (n = 30) 27.6
ET AL
f 1.5
Nondiabetic Group (n= 26) f 1.1
28.1
13.8 f 1.2
-
f 2.5
-
31.4
METHODS
Fifty-four patients with insulin-dependent type I diabetes mellitus were asked to participate in an intensive diabetes treatment program. Patients with a prior history of laser photocoagulation treatment were excluded. Twenty-four of the patients decided to remain in their existing conventional treatment program. The clinical characteristics of these 24 patients who continued in the conventional treatment group are given in Table 1. In most circumstances, diabetes treatment in this group consisted of no more than two daily injections of insulin. Some patients received a mixture of intermediateand short-acting insulin, others only intermediate-acting insulin. All were given instruction on the principles of the dietary treatment of diabetes and placed on a diet as recommended by the American Diabetes Association [6] with sufficient calories to maintain their ideal body weight. A small number of this conventionally treated group chose to occasionally monitor their blood glucose levels as part of their overall diabetes treatment group. Many preferred urine testing for glucose and acetone, however. The average age (ZIZSEM) of this group was 27 f 1.7 years and they had diabetes for an average of 11.9 f 1.3 years. Thirty patients participated in the experimental treatment program that included intensive diabetes and dietary education. The dietary management was based on principles recommended by the American Diabetes Association [6]. Self-monitoring of blood glucose levels was performed systematically several times during each day, and patients were taught to use the test results with individual algorithms for self-insulin adjustments. Insulin was given as a continuous subcutaneous insulin infusion via a portable insulin infusion device. (The devices used during the duration of this study were the Autosyringe ASPC, ASGC, ASGC-UlOO, ASGCMP, and CPI-9100, and the Pacesetter Miniseries). The clinical characteristics of the patients in the experimental treatment group are listed in Table I. This group was identical to the conventional treatment group with respect to age (27.6 f 1.5 years) and duration of diabetes (13.6 f 1.2 years). Procedures for the orientation to our program of treatment with portable infusion devices and for follow-up have been previously described [7,6]. There were no differences between the two groups in terms of the level of arterial blood pressure and the number of patients who smoked. All patients were studied in the General Clinical Research Unit at Parkland Memorial Hospital, Dallas, Texas, at baseline and at six-month intervals thereafter. Initial and
follow-up studies included determinations of levels of total glycosylated hemoglobin, a complete ophthalmologic evaluation, and measurement of the width of skeletal muscle capillary basement membrane. This latter assesment was made at 12-month intervals. Results for the diabetic patients were compared with those of 43 nondiabetic control subjects in whom levels of total glycosylated hemoglobin were measured. Twenty-six of these nondiabetic subjects also volunteered for skeletal muscle biopsy for measurement of the width of their skeletal muscle capillary basement membrane. None had a family history of diabetes. The clinical characteristics of these subjects are given in Table I. Their age averaged 26.1 f 1.1 years. The ophthalmologic evaluation included assessment of best-corrected visual acuity, motility, slit-lamp examinations, and direct and indirect ophthalmoscopy performed after pupillary dilation. Seven-field fundus photography and fluorescein angiography were performed at each six-month visit using a Zeiss fundus camera, which had a field of view of 30 degrees. Prior to grading, the fundus photographs and fluorescein angiograms for a particular visit from each patient were placed in a numbered folder. Photographs were arranged at random and not chronologic sequence. A numeric code was used and the patient’s names were omitted or masked on the photographs. Grading of all photographs was performed separately by two retinal specialists. At the time of grading, neither grader knew the patient’s treatment group assignment. Grading was accomplished by viewing all of the color photographs and fluorescein angiograms with magnification and utilizing stereo pairs wherever possible. In a few cases, color photographs alone were used to evaluate the retinal status if the patient could not tolerate fluorescein because of allergy or untoward reactions, or if the quality of the fluorescein angiograms was inadequate. Microaneurysms were counted only on fluorescein angiograms except in three isolated cases in which fluorescein angiograms were not obtained (two experimental treatment group patients and one conventional treatment group patient). The data were then entered into a computer and the grades for each eye were plotted against time for each patient. Two grading systems were used to categorize the retinopathy of each eye of each patient at each visit. First we used the modified Early Treatment Diabetic Retinopathy Study (ETDRS) system with combined assessment of fun-
December 1986
The
American Journal of Medicine
Volume
81
1013
EFFECT
OF GLYCEMIC
CONTROL
ON MICROVASCULAR
EXPERIMENTAL Ff(
8
::E 2 o
7
IN DIABETES---R~SENST~GK
the identification of the sample would be unknown to the personnel involved with final processing and measurements. The basement membrane width was measured according to the method of Shannon et al [ 131, which is a computer-assisted method of measuring biologic struc-
CONVENTIONAL l$
The total glycosylated hemoglobin level (Ala+~+J was determined by ion exchange chromatography with the use of prepacked microcolumns (Quik-column, Helena Laboratories, Beaumont, Texas) after a thrice washing of the erythrocytes and incubation overnight at 23OC to remove the labile fraction. All total glycosylated hemoglobin levels obtained at visits subsequent to the baseline visit were averaged and the mean glycosylated hemoglobin value (or follow-up value) was compared with that of the baseline. Likewise, all skeletal muscle capillary basement membrane measurements made at visits subsequent to the baseline visit were averaged and this mean value (follow-up value) was compared with the baseline one. The glycosylated hemoglobin levels and capillary basement membrane measurements were compared using the Student t test for paired observations for comparison within groups and for unpaired observations for comparison between groups [ 141. Results of the microaneurysm counts and the modified ETDRS grades were compared by log-likehood chisquared analyses and Fisher’s exact test [ 141. Tests were considered significant at the 0.05 level.
**=p<~ool~vs~
6 Baseline Followup
Non Diabetics
Baseline Followup
Figure 1. Giycosylated hemoglobin levels at baseline and follow-up in patients in the experimental treatment group and in the conventional treatment group. Values in nondiabetic volunteers are included for reference.
dus photographs and fluorescein angiograms. We used the Puklin [9] modification of this grading system that creates an additional category (grade II) for minimal retinopathy. We studied the fluorescein angiograms in conjunction with the color photographs to determine areas of neovascularization, intraretinal microvascular abnormalities, and macular edema. Using this modified ETDRS system, each eye at each particular time point was further categorized by indexing the worst eye and describing the better eye according to the system advanced by Klein et al [IO]. The order of progression within the retinopathy categories was thus (1,i) (2,<2) (2,2) (3,<3) (3,3), etc. A patient was considered worse or better if the eyes changed by two or more of these categories over time. Statistical calculations were performed on the last reading that always coincided with the general trend observed in the plots over time. The second grading system used was the macular microaneurysm count. From the fluorescein angiograms, the ETDRS standard 30-degree macular field was used for counting microaneurysms. The patient’s retinopathy was arbitrarily considered worse or better if the total number of microaneurysms seen in both eyes added together increased or decreased by 20 or more over the duration of the study. If the results of the final grading between the two graders conflicted as to whether a particular patient’s condition worsened or remained stable, the photographs of these patients were assembled by a helper and re-evaluated by both readers together until a consensus was reached. This was also done without the ophthalmologists’ knowledge to which treatment group the patient belonged. The skeletal muscle biopsy specimen was taken from the right quadriceps muscle group under local anesthesia according to the method originally described by Siperstein et al [ 1 l]. The muscle samples were fixed in Palade’s [ 121 fixative and processed by the methods of Siperstein et al [ 1 I]. The samples were sent to the electron microscopy laboratory with consecutive numeric designations so that
1014
ET AL
tures.
2 11 2 0” 10 2E 9 g
COMPLICATIONS
December
1986
The American
Journal
of Medicine
RESULTS The average glycosylated hemoglobin value for the 43 nondiabetic subjects was 6.1 f 0.4 percent. This value was significantly different from those of both treatment groups at all time points. The average width of the skeletal muscle capillary basement membrane in the 26 nondiabetic subjects, inOwhom this measurement was made, was 1,094 f 41 A. In the nondiabetic control group, only a single determination of glycosylated hemoglobin and basement membrane width was made. At baseline, the mean glycosylated hemoglobin level for the patients in the conventional treatment group was 10.6 f 0.3 percent (Figure 1). It remained essentially unchanged over the average 33.5 f 3.7 months of follow-up and averaged 10.4 f 0.3 percent. The baseline glycosylated hemoglobin level in the experimental treat-
ment group was IO. 1 f 0.4 percent, a value that was not different from that in those patients treated by conventional therapy (Figure 1). Following a mean of 31.4 f 2.5 months of the experimental treatment, the glyCOSylated hemoglobin level fell to an average value of 7.2 f 0.3
percent, p
value
in the experimental
treatment
group
attests to the effectiveness of the experimental program. The severity of the retinopathy at entry in both study groups was the same (Table II). Using the modified ETDRS grading system, there was a preponderance of patients with very minimal or no retinopathy. In addition, although there were no statistical differences between the
Volume
81
EFFECT
OF GLYCEMIC
CONTROL
ON MICROVASCULAR
two groups at entry, the experimental treatment group had a few more patients with more severe retinopathy. This was true irrespective of whether the ETDRS grading system or the macular microaneurysm counts were used (data not shown). The results of the effect of diabetes control on the progression of retinopathy are shown in Table III. In terms of the modified ETDRS grading system, in the conventional treatment group, the condition of 18 patients remained stable, the condition of six progressed, and none had improvement. In the experimental treatment group, the condition of 23 patients remained stable, the condition of three progressed, and four showed improvement. With respect to the macular microaneurysm counts in the conventional treatment group, the condition of 19 remained stable, the condition of five worsened, and none showed improvement. In the experimental treatment group, the condition of 28 remained stable, the condition of one worsened, and one showed improvement. Using either grading systems, there were significant differences between the results from the two treatment groups. The, ETDRS grading showed a significant relationship between diabetes control and the progression of diabetic retinopathy as defined by the categories better, stable, or worse (p <0.05). In terms of macular microaneurysm counts (since only one patient in the experimental group improved), only the categories stable and worse were analyzed and again both treatment groups were significantly different (p <0.05). The width of skeletal muscle capillary basement membrane in the conyentional treatment group at baseline was 1,675 f 750 A, a value identical to that of the experimental treatment group, 1,617 f 91 A (Figure 2). Both baseline values were greater than in the nondiabetic group (p
COMPLICATIONS
TABLE II
342
Figure 2. Skeletal muscle capil/aIy basement membrane width at baseline and follow-up in patients in the experimental treatment group and in the co* ventional treatment group. l l =p < 0.05 versus baseline.
ET AL
Degree of Retinopathy at Study Entry (Modified Early Treatment Diabetic Retinopathy Study Grading) Conventional Treatment Group
(l,l)to (4,X4) (6,<6) (88) to
TABLE Ill
Effect of Diabetes of Retinopathy Modilied Early Treatment Diabetic Retlnopathy Study Experimental” Conventional Treatment Treatment
Better Stable Worse
Experimental Treatment Group
22 2 0 0
(3,3) to (575) to (8,<8) (10,lO)
4 23 3
21 6 3 0
Control
on Progression
Macular Microaneurysm Count Experimental* Conventional Treatment Treatment
0 18 6
* p <0.05, experimental (log-likelihood chi-square).
1 28 1
treatment
versus
0 19 5
conventional
treatment
brane wktth in the conventional treatment group after 33.5 f 3.7 months of follow-up was 1,794 f 100 A, a value that was identical to the baseline value in this treatment group. However, in the experimental treatment group, the average follow-up basement membrane width was 1,417 f 50 A, a value that was significantly less (p (0.005) than the baseline value and significantly different from the follow-up value in the conventional treatment group (p
ExDerimental
-
IN DIABETES-ROSENSTOCK
Conventional
Rx
Rx
2000
2000 1800
“5
1800
.$
1600
g
1600
$
1400
3Q 5 1400
2
1200 1000
i
2E 3E
800
24 1000 E 2 800
z
600
3 Baseline
December
1986
1200
600 Baseline
Followup
The American
Journal
of Medicine
Volume
Followup
81
1015
EFFECT
OF GLYCEMIC
CONTROL
ON MICROVASCULAR
COMPLICATIONS
EXPERIMENTAL F$ RETINOPATHY Better Stable (4) / (23)
IN DIABETES-ROSENSTOCK
CONVENTIONAL s, RETINOPATHY
Worse (3)
Better (0) / Baseline Followup +p
Stable (18)
Worse (8)
0
i f 2000 P 2 1800 22. 1600 g 1400 +g 1200 L+i1000 m”
q
q
VSH
Figure 3. Skeletal muscle capillary basement membrane width at baseline and follow-up in patients in the experEmental treatment group and in the colt ventional treatment group as a function of the status of their diabetic retinopathy.
800
given in Figure 3. This figure shows that the 27 patients in the experimental treatment group whose retinopathy remained stable or improved had a significant reduction of basement membrane width from 1,636 f 99 A at baseline to 1,404 f 54 A at follow-up. Those 18 patients in the conventional treatment group whose retinopathy remained stable (none improved) had identical basement membrane width at baseline and at follow-up (1,627 f 120 A and 1,704 f 93 A, respectively). In both treatment groups, in those patients whose retinopathy worsened, there was a tendency for the skeletal muscle capillary basement membrane width to increase in thickness, although no statistically significant relationship was appreciated. This was probably due to the small sample size.
These data, although exciting, contrast to some extent with other studies performed to date. Specifically, we make reference to the Steno [ 141 and KROC [ 151 studies. In these prospective, randomized, but shorter-duration clinical trials, there was actually a tendency towards progression of diabetic retinopathy in those patients who had the best diabetes control. The progression of retinopathy in these groups was usually the result of the appearance of cotton-wool spots. With longer follow-up, these cotton wool spots tend to regress [ 161. The precise reason why our data differ from these is not entirely clear, but the longer duration of follow-up at six-month rather than shorter intervals and the fact that all of our patients had early diabetic retinopathy are two explanations that seem readily apparent. Our finding that long-term normoglycemia causes a reduction in skeletal muscle capillary basement membrane widths confirms and extends our previous report [ 171. These data show that the reduction in skeletal capillary basement membrane width that occurs after two years of therapy [ 171 continues with prolonged follow-up. Of interest in this study is the association between the improvement (i.e., reduction in thickness) seen in skeletal muscle capillary basement membrane width and clinical microangiopathy. One of the major issues regarding the relevance of measurements of skeletal capillary basement membrane width is the fact that it has been difficult in the past to correlate those measurements with clinical microangiopathy [ 18,191. In this study, those patients in the experimental treatment group whose retinopathy improved or remained stable had a significant reduction in their skeletal muscle basement membrane width. Patients whose diabetic retinopathy progressed, irrespective of treatment, tended to have an increase in the width of the basement membrane in the capillaries of their skeletal muscle.
COMMENTS Our data show that in patients with type I diabetes mellitus, there appears to be a significant effect of improved glycemic control, using an experimental treatment program, on the small blood vessel complications of diabetes. Clearly, patients in our experimental treatment group had long-term improvement of blood glucose control at normal or near-normal levels as evidenced by significant falls in levels of their total glycosylated hemoglobin. In addition, it should be emphasized that in our experimental treatment group, the diabetic retinopathy exhibited by these patients showed significantly less deterioration over time, i.e., the retinopathy was either stable or improved more often in the experimental treatment group than in the conventional treatment group, where the retinopathy tended to worsen. Thus, in our study, there appears to be a significant advantage in terms of the slowing of the progression of early diabetic retinopathy in patients undergoing an experimental treatment program that resulted in near-normal glycemia for periods of up to five years.
1016
December
ET AL
1986
The American
Journal
of Medicine
Volume
81
EFFECT
OF GLYCEMIC
CONTROL
ON MICROVASCULAR
In this regard, it is important to note that others have been unable to correlate skeletal muscle capillary basement membrane width to the presence or absence of diabetic nephropathy [20]. There are several possible explanations. First of all, skeletal muscle capillary basement membrane thickness may be more indicative of retinopathy than of nephropathy. In addition, the studies on nephropathy were cross-sectional in nature and not longitudinal as is ours. Furthermore, they used a very selective and heterogeneous group of patients, those being evaluated for possible pancreas transplantation, with different stages of microvascular complications. Finally, they fixed their muscle biopsy specimens in gluteraldehyde, a fixative that has been shown to produce better photographs, but to be less specific in documenting diabetic microangiopathy [2 11. Several factors relating to the design of this study and interpretation of the results should be discussed. First and most imortant, this was a prospective study, although patients were not randomly assigned to a treatment group. All patients in the study were offered enrollment in the experimental treatment program. It is possible that those patients who were more interested in improving their level of diabetic control and who believed that this could be important to their overall well-being would be most likely to choose the experimental treatment. It is also possible that there were some inherent differences between the two groups that were not apparent in biologic measures. Clearly, these subjects were similar in age, duration of diabetes, arterial blood pressure, and the number who
COMPLICATIONS
IN DIABETES-ROSENSTOCK
ET AL
smoked. At entry into the study, both treatment groups had identical values for the width of skeletal muscle capillary basement membranes and for glycosylated hemoglobin. The results of this study must be viewed in the context of the entire clinical spectrum of insulin-dependent diabetes mellitus. Although our data, for the first time, indicate an advantage of improved diabetic control in terms of the progression of diabetic retinopathy, we cannot at present recommend this type of intensive treatment to all patients with insulin-dependent diabetes. This study certainly provides substantial data that need to be verified by a larger prospective randomized clinical trial such as the Diabetes Control and Complications Trial [22] that is presently underway. ACKNOWLEDGMENT We wish to thank Patricia Challis, R.N., B.S.N., Suzanne Strowig, B.S.N., M.S.N., and Susan Cercone, RD., M.S. for their invaluable help; George Sanborn, M.D., for helping with the grading of the retinal photographs and fluorescein angiograms; Lisa Gagliano, Arthur Ojirika, and Kevin Sullivan for their technical assistance; Debra Elkins and William Anderson for the retinal photographs; Bette Newton for her assistance in the preparation of the manuscript; and the staff and nurses of the General Clinical Research Center for the care of our patients. Finally, our heartfelt thanks to our patients, without whose loyalty and dedication this study could not have been possible.
REFERENCES 1. 2. 3.
4.
5. 6.
7.
8.
9.
Raskin P: Diabetic regulation and its relationship to microangiopathy. Metabolism 1978; 2: 235-252. Cahill GF Jr, Etzwiler DD, Freinkel N: “Control” and diabetes. ?I Engl J Med 1976; 294: 1004-1005. Siperstein MD, Foster DW, Knowles HC Jr, Levine R, Madison LL, Roth J: Control of blood glucose and diabetic vascular disease. N Engl J Med 1977; 296: 1060-1063. Raskin P: Treatment of type I diabetes mellitus with portable insulin infusion devices. Diabetes Care 1982; 1: (suppl 5) 48-52. Skyler JS: Patient self-monitoring of blood glucose. Clin Diabetes 1983; 1: 12-17. American Diabetes Association: Principles of nutrition and dietary recommendations for ‘individuals with diabetes mellitus. Diabetes 1979; 28: 1027-1030. Raskin P, Pietri A, Unger R: Changes in glucagon levels after four to five weeks of glucoregulation by portable insulin infusion pumps. Diabetes 1979; 28: 1033-1035. Raskin P: Treatment of insulin-dependent diabetes mellitus with portable insulin infusion devices. Med Clinic North Am 1982; 66: 1269-1283. Puklin JE, Tamborlane WV, Felig P, Gene1 M, Sherwin RS: Influence of long-term insulin infusion pump treatment of
December
10.
11.
12. 13.
14.
15.
16.
1986
type I diabetes on diabetic retinopathy. Ophthalmology 1982; 89: 735-747. Klein, BE, Davis MD, Segal P, et al: Diabetic retinopathy. Assessment of severity and progression. Ophthalmology 1984; 91: 10-17. Siperstein MD, Unger RH, Madison LL: Studies of muscle capillary basement membranes in normal subjects, diabetic, and prediabetic patients. J Clin Invest 1968; 47: 1973-1999. Palade GE: A study of fixation for electron microscopy. J Exp Med 1952; 95: 285-298. Shannon WA Jr, Rockholt DL, Bates SB: Computer-assisted measurement of the thickness of biological structures. Comput Biol Med 1982; 12: 149-155. Lauritzen T, Frost-Larsen K, Deckert T, et al: Steno Study Group. Effect of 1 year of study of near-normal blood glucose levels on retinopathy in insulin-dependent diabetics. Lancet 1983; I: 200-204. KROC Collaborative Study Group: Blood glucose control and the evolution of diabetic retinopathy and albuminuria. A preliminary multicenter trial. N Engl J Med 1984; 311: 365-372. Lauritzen T, Frost-Larsen K, Larsen HW, et al: Steno Study
The American
Journal
of Medicine
Volume
81
1017
EFFECT
17.
18.
19.
1018
OF GLYCEMIC
CONTROL
ON MICROVASCULAR
COMPLICATIONS
Group. Continuous subcutaneous insulin infusion. Lancet 1983; I: 1445-1446. Raskin P, Pietri A, Unger. R, et al: The effect of diabetic control on skeletal muscle capillary basement membrane width in patients with type I diabetes mellitus. N Engl J Med 1983; 300: 1546-1560. Kilo 6, Volger N, Williamson JR: Muscle capillary basement membrane changes related to aging’and to diabetes mellitus. Diabetes 1972; 21: 881-905. Siperstein MD: Capillary basement membranes and diabetic microanglopathy. Adv-Intern Med 1972; 18: 325-344.
December
1986
The American
Journal
of Medicine
IN DIABETES-ROSENSTOCK
20.
Steffes MW, Sutherland DER, Goertz FC, Rich SS, Mauer SM: Studies of kidney andmuscle biopsy specimens from identical twins discordant for-type I diabetes mellitus. N Engl J Med 1985; 312: 1282-1286. Siperstein MD, Raskin P, Bum .H: Electron microscopic quantification of diabetic retinopathy. Diabetes 7973; 22: 514-527. . Proposed protocol for the clinical trial to assess the relationship between metabolic control and the early vascular complications of insulin-dependent diabetes mellitus. Diabetes ‘1982; 31: 1132-l 133.
21.
22.
Volume
ET AL
61