- Email: [email protected]
Effect of lisinopril on progression of retinopathy in normotensive people with type 1 diabetes
EARLY REPORT
Early Report
Effect of lisinopril on progression of retinopathy in normotensive people with type 1 diabetes Nish Chaturvedi, Anne-Katrin Sjolie, Judith M Stephenson, Heidemarie Abrahamian, Marc Keipes, Allesandro Castellarin, Zeljka Rogulja-Pepeonik, John H Fuller, and the EUCLID Study Group*
Summary
Introduction
Background Retinopathy commonly occurs in people with type 1 diabetes. Strict glycaemic control can decrease development and progression of retinopathy only partially. Blood pressure is also a risk factor for microvascular complications. Antihypertensive therapy, especially with inhibitors of angiotensin-converting enzyme (ACE), can slow progression of nephropathy, but the effects on retinopathy have not been established. We investigated the effect of lisinopril on retinopathy in type 1 diabetes.
Retinopathy eventually develops in 70–100% of people with type 1 diabetes (insulin-dependent diabetes mellitus) and remains a common cause of visual impairment and eventual blindness.1–3 The only intervention that prevents development and slows progression of retinopathy is tight glycaemic control.4 However, this approach is not wholly effective, and other interventions must be sought. Blood pressure is an important risk factor for the development of retinopathy.5–7 Antihypertensive therapy, especially inhibitors of angiotensin-converting enzyme (ACE), slow the progression of nephropathy,8,9 but whether these agents have a beneficial effect on retinopathy is much less clear. Previous randomised controlled trials of these drugs in people with type 1 diabetes either did not collect standardised data on retinopathy status,9 or did not have sufficient power10,11 to show an effect. We report the retinopathy results from the EUCLID (EURODIAB Controlled Trial of Lisinopril in Insulin-Dependent Diabetes Mellitus) study.12
Methods As part of a 2-year randomised double-blind placebo-controlled trial, we took retinal photographs at baseline and follow-up (24 months) in patients aged 20–59 in 15 European centres. Patients were not hypertensive, and were normoalbuminuric (85%) or microalbuminuric. Retinopathy was classified from photographs on a five-level scale (none to proliferative). Findings The proportion of patients with retinopathy at baseline was 65% (117) in the placebo group and 59% (103) in the lisinopril group (p=0·2). Patients on lisinopril had significantly lower HbA1c at baseline than those on placebo (6·9% vs 7·3 p=0·05). Retinopathy progressed by at least one level in 21 (13·2%) of 159 patients on lisinopril and 39 (23·4%) of 166 patients on placebo (odds ratio 0·50 [95% CI 0·28–0·89], p=0·02). This 50% reduction was the same when adjusted for centre and glycaemic control (0·55 [0·30–1·03], p=0·06). Lisinopril also decreased progression by two or more grades (0·27 [0·07–1·00], p=0·05), and progression to proliferative retinopathy (0·18 [0·04–0·82], p=0·03). Progression was not associated with albuminuric status at baseline. Treatment reduced retinopathy incidence (0·69 [0·30–1·59], p=0·4). Interpretation Lisinopril may decrease retinopathy progression in non-hypertensive patients who have type 1 diabetes with little or no nephropathy. These findings need to be confirmed before changes to clinical practice can be advocated.
Lancet 1998; 351: 28–31
*Listed at end of paper EURODIAB, Department of Epidemiology and Public Health, University College London, London WC1E 6BT, UK (N Chaturvedi MRCP, J M Stephenson MD, J H Fuller FRCP); Aarhus University Hospital, Aarhus, Denmark (A-K Sjolie MD); Hospital Vienna Lainz and L Boltzmann-Institute for Nutrition and Metabolic Diseases, Vienna, Austria (H Abrahamian MD); Maladies du Metabolisme et de la Nutrition, Luxembourg (M Keipes MD); Università de Verona, Verona, Italy (A Castellarin MD); and Vuk Vrhovac Institute, Zagreb, Croatia (Z Rogulja-Pepeonik MD) Correspondence to: Dr Nish Chaturvedi (e-mail: [email protected])
28
Methods The EUCLID study was a double-blind, randomised, paralleldesign clinical trial of lisinopril, an ACE inhibitor, and placebo in 18 European centres, in 530 patients. Men and women with type 1 diabetes13 aged 20–59 years were recruited if resting blood pressure was 75–90 mm Hg diastolic, and 155 mm Hg or less systolic. Most of the patients were normoalbuminuric (albuminexcretion rate <20 µg/min); 15% were microalbuminuric (albumin-excretion rate 20–200 µg/min). Trained photographers took retinal photographs of two 45–50° fields of each eye—macula temporal and disc nasal—at baseline and at 24 months. Slides were sent to London, UK, and then to Aarhus, Denmark, for classification on a five-level scale. AKS assessed all photographs according to the EURODIAB protocol, based on the modified Airlie House classification.14 She had no access to information about patients, except study number. Concordance between the EURODIAB and other methods of grading is high.14 Three centres (Bucharest, Krakow, and Watford) did not have retinal cameras and did not participate in the retinopathy section of the study. Therefore, photographs were taken in 409 patients, 354 of whom had photographs at baseline and follow-up. Patients were re-examined at 1, 3, 6, 12, 18, and 24 months. Blood samples were taken every 6 months to measure glycated haemoglobin (HbA1c). Patients were started on 10 mg lisinopril or matching placebo daily at randomisation. The dose was increased to 20 mg if diastolic blood pressure was more than 75 mm Hg at 3 months. HbA1c was estimated in London by EIA with a monoclonal antibody raised against HbA1c (Dako Ltd, Ely, UK); the normal range for this assay is 2·9–4·8%.15 We estimated that 500 patients would be needed to show a difference between groups in urinary albumin excretion rate, the primary endpoint. Retinopathy was a secondary endpoint, and we estimated the statistical power of the trial to detect a treatment difference in degree of retinopathy. The WESDR showed that retinopathy progressed over 4 years in 48% of
THE LANCET • Vol 351 • January 3, 1998
EARLY REPORT
530 randomised 121 had no baseline photographs No camera (96) Ungradable (7) No photograph (18) 409 had gradable baseline retinal photographs 207 assigned placebo 28 had no final photograph Photograph lost (9) Withdrawn (17) Violated protocol (2)
202 assigned lisinopril 27 had no final photograph Photograph lost (7) Withdrawn (17) Violated protocol (3)
179 had final photographs
175 had final photographs
168 completed trial as per protocol
156 completed trial as per protocol
patients with young-onset diabetes with a diastolic blood pressure of between 79 mm Hg and 85 mm Hg and nonproliferative retinopathy at baseline. We estimated that retinopathy would progress in 24% of the placebo group over 2 years. We used data from the EURODIAB IDDM Complications Study to estimate the likely rate of retinopathy at baseline.13 With the assumption that we would have 250 patients in each group, of whom half would have retinopathy (non-proliferative) at baseline, we would be able to detect a reduction in retinopathy progression from 24% on placebo to 10% on lisinopril (odds ratio 0·35) with 80% power at 5% significance. Patients were assigned to one of five retinopathy levels by the worst affected eye.14 These assignments were based on lesions graded: none, minimal, moderate, severe non-proliferative, and photocoagulation and proliferative combined. The main outcome for retinopathy was progression of retinopathy by at least one grade. Those who could not show progression (photocoagulated and proliferative at baseline), were excluded from these analyses. Patients whose retinopathy progressed by at least two grades or to proliferative retinopathy were also compared. We compared the development of retinopathy in patients with no retinopathy at baseline, and regression of retinopathy for all patients with any retinopathy at baseline. We calculated odds ratios for progression, development, and regression to compare patients in the two treatment groups by the Mantel Haenszel test.16 Logistic regression was used to adjust these odds ratios for confounders, such as centre, HbA1c, baseline albumin excretion rate, duration of diabetes, age, sex, baseline blood pressure, and baseline retinopathy status; only patients with complete data for confounders were included in these analyses. In some cases, because of small numbers, adjustment could not be made for centre. Between-observer and within-observer variations for retinal grading were assessed by asking the original grader (AKS), and an independent grader (S Aldington, Retinopathy Grading Centre, Hammersmith Hospital, London, UK), to regrade 60 sets of baseline and follow-up photographs for 30 (8·5%) patients randomly selected from all patients who had both sets of photographs. We calculated a weighted score to assess the
None Minimal non-protective Moderate non-proliferative Severe non-proliferative Photocoagulated or proliferative
Lisinopril
Retinopathy
Retinopathy
None M/F Age (years) Duration of diabetes (years) % current smokers Body-mass index (kg/m2) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) HbA1c (%) Albumin excretion rate (µg/min)‡ % with microalbuminuria
41/21 31 (7) 9 (6)
Any
All
79/38 120/59 37 (8)* 35 (8) 18 (7) 15 (8)
35 33 34 23·6 25·2* 24·7 (2·3) (2·8) (2·7) 121 (11) 124 (11) 123 (11) 81 (4) 7·1 (1·8) 7·2 (4·7– 12·4) 13
81 (5) 7·4 (2·0) 7·4 (4·7– 15·2) 15
81 (5) 7·3 (1·9) 7·2 (4·7– 14·7) 14
None
Any
All
42/30 69/34 111/64 32 (8) 36 (10)* 34 (9) 10 (8) 17 (8)* 14 (9) 31 20† 25 24·0 25·2* 24·7 (2·6) (2·9) (2·8) 123 (10) 123 (10) 123 (10) 80 (4) 6·7 (2·2) 5·7 (3·4– 11·6) 10
81 (5) 7·0 (1·6) 7·9 (4·5– 17·1) 22†
81 (5) 6·9† (1·9) 6·9 (3·9– 3·3) 17
p values refer to comparisons between those with and without retinopathy within treatment group, and between treatments for the all category. *p<0·01. †p=<0·05. ‡Median (IQR).
Table 2: Baseline characteristics by retinopathy level and treatment group (mean [SD])
Figure 1: Trial profile
Retinopathy status
Placebo
Treatment Placebo (n=179)
Lisinopril (n=175)
62 (35%) 81 (45%) 23 (13%) 0 13 (7%)
72 (41%) 67 (38%) 18 (10%) 2 (1%) 16 (9%)
Table 1: Retinopathy status at baseline by treatment group
THE LANCET • Vol 351 • January 3, 1998
degree of within-observer and between-observer agreement. Odds ratios for progression were calculated for the three previous trials of ACE inhibitors and diabetic retinopathy,10,11,17 and a combined odds ratio was estimated by the method of Woolf.18
Results 409 of 530 patients had gradable baseline photographs. 121 patients did not have photographs. Therefore, gradable photographs were obtained at baseline and follow-up for 354 (81·6%) of 434 eligible patients (figure 1). The only significant differences in baseline characteristics between patients with and without photograhs were for diastolic blood pressure (81 vs 79 mm Hg in patients with and without photographs; p=0·0001), and distribution of sex (47 vs 37% women of patients with and without photographs; p=0·04). These differences were consistent when assessed separately by treatment group. Patients with both sets of photographs were older (34 vs 31 years, p=0·005), had had diabetes for longer (15 vs 12 years, p=0·02), had higher systolic blood pressure (123 vs 120 mm Hg, p=0·03) and better glycaemic control (HbA1c 7·1 vs 7·8%, p=0·08), and fewer were women (35% vs 51%, p=0·02) compared with those who only had baseline photographs. These differences were consistent when assessed by treatment group separately. Distribution of retinopathy at baseline was similar by treatment for patients with only baseline photographs. Retinopathy was present in 117 (65%) patients in the placebo group and 103 (59%) in the lisinopril group (p=0·2). Most cases of retinopathy were minimal nonproliferative (table 1). In both treatment groups, patients with retinopathy were older, had had diabetes for longer, and had higher body-mass indices than those without retinopathy; in the lisinopril group, those with retinopathy had a higher rate of microalbuminuria (table 2). HbA1c was significantly lower at baseline in the lisinopril group than in the placebo group (6·9 vs 7·3%; p=0·05). HbA1c did not change during the trial in either treatment group.12 Retinopathy progressed by at least one level in the worst eye in 39 (23·4%) of 166 patients in the placebo group, and 21 (13·2%) of 159 in the lisinopril group (p=0·02; table 3). Progression by one level was 29
EARLY REPORT Adjustment
Lisinopril
Odds ratio (95% CI)
p
Progression by one level* Unadjusted 39/166 Centre .. .. HbA1c Full† ..
Placebo
21/159 .. .. ..
0·50 (0·28–0·89) 0·50 (0·28–0·92) 0·55 (0·30–1·03) 0·55 (0·29–1·05)
0·02 0·03 0·06 0·07
Progression by two levels* Unadjusted 11/166 .. HbA1c
3/157 ..
0·27 (0·07–1·00) 0·30 (0·08–1·10)
0·05 0·07
Progression to proliferative retinopathy* Unadjusted 11/166 2/159 .. .. HbA1c
0·18 (0·04–0·82) 0·20 (0·04–0·91)
0·03 0·04
Incidence of retinopathy‡ Unadjusted 15/62 .. HbA1c
13/72 ..
0·69 (0·30–0·59) 0·73 (0·31–1·74)
0·4 0·5
Regression of retinopathy§ Unadjusted 28/117 .. HbA1c
33/103 ..
1·48 (0·82–2·68) 1·41 (0·78–2·58)
0·2 0·3
*Number of progressors/number in whom progression was possible. †Adjustment for centre, HbA1c, age, duration of diabetes, sex, baseline albumin excretion rate, baseline retinopathy grade, baseline blood pressure. ‡Number with incident retinopathy/number without retinopathy at baseline. §Number with regression/number with retinopathy at baseline.
Table 3: Progression, development, and regression of retinopa-
seen equally at all baseline levels of retinopathy. In patients on lisinopril compared with placebo the odds ratio adjusted for centre was 0·50 (95% CI 0·28–0·92, p=0·03). On logistic regression only HbA1c was related to progression of retinopathy. Adjustment for HbA1c at baseline made the odds ratio 0·55 (0·30–1·03, p=0·06; table 3). There was no evidence of a significant interaction by HbA1c (p=0·4). However, a stratified analysis by HbA1c with a cut-off of 7%, the median value, gave an odds ratio related to treatment of 0·34 (0·13–0·94; p=0·04) in patients with HbA1c less than 7%, and 0·61 (0·29–1·29; p=0·2) in those with HbA1c of 7% or more. There was no evidence of other interactions. Systolic blood pressure was 3 mm Hg lower in the lisinopril group than in the placebo group at 1 month, and this difference persisted until the end of the trial.12 We adjusted the model with all other potential confounders for systolic blood pressure at 1 month and the odds ratio was 0·57 (0·30–1·10; p=0·09). Progression by two levels, and progression to proliferative retinopathy was also reduced by treatment after adjustment for HbA1c (table 3). There was disagreement on the level of retinopathy for (8·2)
Larsen et al11
Chase et al10 Ravid et al17
This study Meta-analysis
0
1
2 3 Odds ratio
4
Figure 2: Odds ratios (95% CI) for progression of retinopathy for both groups in EUCLID and previous studies
30
six of 60 photographs when regraded by the same observer. These photographs were equally distributed between baseline and final visit and treatment groups. Two sets were assigned to a worse retinopathy level than originally, and four to a better level. The within-observer agreement () was 0·85 (0·67–1·00), and the betweenobserver was 0·55 (0·39–0·71). Individual odds ratios for the previous three studies all indicate a beneficial, but non-significant effect of ACE inhibitors on retinopathy (figure 2). Combination of these odds ratios produces an overall odds ratio of 0·46 (0·19–1·10; 2 for heterogeneity 0·57, p>0·5). A metaanalysis of all studies, including our study, gave 2 for heterogeneity between studies of 0·59 (p>0·5), with a combined odds ratio of 0·49 (0·30–0·79).
Discussion We show that lisinopril had beneficial effects on progression of retinopathy. Non-significant benefits have been shown previously with other ACE inhibitors in people with type 110,11,19 and type 2 (non-insulin-dependent diabetes mellitus).17 The analyses for retinopathy were included in the EUCLID protocol, but the study was not primarily designed to assess the effects of ACE inhibitors on retinopathy. However, retinal photographs taken by standard methods are an accepted way of assessing changes in retinopathy status,20 and the EURODIAB protocol is well validated.14 Not all patients in our trial had gradable photographs, but we did not find notable differences in baseline characteristics between those who did and did not have photographs. We could analyse only data on patients who completed the trial and, therefore, had better-than-average compliance. Withdrawal rates were similar in the two treatment groups and, therefore, a differential compliance by treatment is unlikely to explain our findings. The within-observer and between-observer differences in the grading of the retinal photographs were not systematically biased by treatment group. Confounders such as glycaemic control can account for only part of the effect of lisinopril on retinopathy progression, and other explanations are needed. Serum concentrations of ACE inhibitor are positively related to the presence and degree of retinopathy,21 and captopril, an ACE inhibitor, limits the abnormal leak of fluorescein from retinal vessels seen in diabetic retinopathy.11,12 This limitation may be due to a direct effect on systemic blood pressure, but studies show that ACE is produced locally by vascular endothelial cells,23,24 which may have direct detrimental effects on retinal flow and vascular structure, independent of changes in systemic blood pressure.25 A direct effect of ACE inhibitors on the eye could be why we found that the retinal effects of lisinopril cannot be accounted for by renal effects. The slowing of progression of retinopathy to proliferative disease, once established, has been by improvement of glycaemic control,4 but control is difficult to achieve26,27 and a proportion of patients will still develop or have advanced retinopathy. Good glycaemic control halves the progression of retinopathy over a 2-year period (progression of one of five levels in EUCLID vs three of 25 in the Diabetes Control and Complications Trial). In EUCLID, however, we found that patients with better glycaemic control had the most benefit from ACEinhibitors, combination may be the best therapeutic
THE LANCET • Vol 351 • January 3, 1998
EARLY REPORT
approach. We show that the beneficial effects of lisinopril seem not to depend on the presence of renal disease; even those without microalbuminuria seemed to benefit. Restriction of ACE inhibitors to patients at risk of retinopathy would exclude many who may benefit. ACE inhibitors may have an important role in limiting the progression of retinopathy in patients with type 1 diabetes, but these findings, and the effects of this treatment on the development and regression of retinopathy, must be studied further in large, randomised, controlled trials before changes in clinical practice are advocated. EUCLID investigators B Karamanos, A Kofinis, C Petrou, Hippokration Hospital, Athens, Greece; C Ionescu-Tirgovisite, C Iosif, Clinic of Diabetes Nutrition and Metabolic Diseases, Bucharest, Romania; G Tamas, G Bibok, 1st Department of Medicine, Semmelweis University, Budapest, Hungary; Z Kerenyi, P Kis-Gombos, J Toth, Szent Imré Teaching Hospital, Budapest, Hungary; J B Ferriss, G Grealy, Cork University Hospital, Wilton, Cork, Ireland; R Rottiers, H Priem, University Hospital, Gent, Belgium; V Koivisto, J Tuominen, E Kostamo, Helsinki University Hospital, Helsinki, Finland; B Idzior-Walus, B Solnica, D Galicka-Latalie, University School of Medicine, Krakow, Poland; G Michel, M Keipes, A Giuliani, A Herode, Hospitalier de Luxembourg, Luxembourg; F Santeusanio, A Bueti, S Bistoni, Dr Cagini, Instituto di Medicina Interna E Scienze Endocrine E Metaboliche, Perugia, Italy; R Navalesi, G Penno, M Nannipieri, L Rizzo, R Miccoli, Instituto di Clinica Medica II, Pisa, Italy; G Ghirlanda, P Cotroneo, A Manto, A Minella, C Saponara, Universita Cattolica del Sacro Cuore, Rome, Italy; J Ward, M Plater, S Ibrahim, S Ibbotson, C Mody, Royal Hallamshire Hospital, Sheffield, UK; N Papazoglou, C Manes, K Soulis, M Voukias, Agios Pavlos General Hospital, Thessaloniki, Greece; M Muggeo, V Cacciatori, M L Gemma, A Dellera, A Castellarin, Ospedale Civile, Verona, Italy; K Irsigler, H Abrahamian, C Gurdet, C Willinger, Hospital Vienna Lainz and L Boltsmanninstitute for Nutrition and Metabolic Diseases, Vienna, Austria; A Nelstrop, C Feben, Watford General Hospital, Watford, UK; S Walford, V McLelland, S Hughes, New Cross Hospital, Wolverhampton, UK; Z Metelko, G Roglic, Dr Z Rogulja Pepeonik, Vuk Vrhovak Institute for Diabetes, Zagreb, Croatia. Steering Committee Members—B Ferriss, chairman, Cork; J H Fuller, London; B Karamanos, Athens; Z Kerenyi, Budapest; G Michel, Luxembourg; M Muggeo, Verona; J Stephenson, London; N Chaturvedi, London; G C Viberti, London; A K Sjolie, Aarhus. Coordinating Centre—J H Fuller, J Stephenson, N Chaturvedi, J Holloway, M Milne, D Webb. Retinopathy Grading Centre—A K Sjolie, Aarhus, Denmark. Central Laboratory—G John, D J Newman, The Royal London Hospital and St Bartholomew’s and The Royal London Hospital School of Medicine and Dentistry, London, UK.
Contributors Nish Chaturvedi jointly supervised the project, carried out analysis, and drafted the paper. Anne-Katrin Sjolie advised on the design of the retinopathy section of the study, graded all retinopathy, and assisted in writing the paper. Judith Stephenson devised the original idea for the project, designed the study protocol, jointly supervised the project, advised on statistical analysis, and helped to draft the paper. Heidemarie Abrahamian and Marc Keipes recruited and assessed patients at their centres and participated in the drafting of the paper. Allesandro Castellarin and Zeljka Rogulja-Pepeonik took the retinal photographs and commented on the paper. John Fuller was the team leader, assisted in the design of the protocol, jointly supervised the project, advised on statistical analyses, and assisted with the drafting of the paper.
Acknowledgments This study was supported by a grant from Zeneca Pharmaceuticals, who also provided the lisinopril and placebo tablets. We thank the EUCLID investigators, Rebecca Hardy, statistician, MRC National Survey of Health and Development, University College, London, who performed the meta-analysis, Steve Aldington and staff at the Retinopathy Grading Centre, Hammersmith Hospital, London, UK, and the people who took part in the study for their collaboration.
3
4
5
6
7
8
9
10
11
12
13
14
15 16 17
18 19
20
21
22
23
24
25
References 1
2
Dwyer MS, Melton III LJ, Ballard DJ, Palumbi PJ, Trautman JC, Chu CP. Incidence of diabetic retinopathy and blindness: a populationbased study in Rochester, Minnesota. Diabetes Care 1985; 8: 316–22. Klein R, Klein BEK, Moss SE, Davis MD, DeMets DL. The Wisconsin Epidemiologic Study of diabetic retinopathy II prevalence and risk of diabetic retinopathy when age at diagnosis is less than 30 years. Arch Ophthalmol 1984; 102: 520–60.
THE LANCET • Vol 351 • January 3, 1998
26
27
Sjolie AK, Stephenson J, Aldington S, et al. Retinopathy and vision loss in insulin-dependent diabetes in Europe. Ophthalmology 1997; 104: 252–60. The Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 1993; 329: 977–86. Klein R, Klein BEK, Moss SE, Davis MD, DeMets DL. Is blood pressure a predictor of the incidence or progression of diabetic retinopathy? Arch Intern Med 1989; 149: 2427–32. Janka HU, Warram JH, Rand LI, Krolewski AS. Risk factors for progression of background retinopathy in long-standing IDDM. Diabetes 1989; 38: 460–64. Teuscher A, Schnell H, Wilson PWF. Incidence of diabetic retinopathy and relationship to baseline plasma glucose and blood pressure. Diabetes Care 1988; 11: 246–51. Lewis EJ, Hunsicker LG, Bain RP, Rohde RD, and the Collaborative Study Group. The effect of angiotensin converting enzyme inhibition on diabetic nephropathy. N Engl J Med 1993; 329: 1456–62. The Microalbumin Captopril Study Group. Captopril reduces the risk of nephropathy in IDDM patients with microalbuminuria. Diabetologia 1996; 39: 587–93. Chase HP, Garg SK, Harris S, Hoops S, Jackson WE, Holmes DL. Angiotensin-converting enzyme inhibitor treatment for young normotensive diabetic subjects: a two-year trial. Ann Ophthalmol 1993; 25: 284–89. Larsen M, Hommel E, Parving HH, Lund-Andersen H. Protective effect of captopril on the blood-retina barrier in normotensive insulindependent diabetic patients with nephropathy and background retinopathy. Graefes Arch Clin Exp Ophthalmol 1990; 228: 505–09. The EUCLID study group. Randomised placebo-controlled trial of lisinopril in normotensive patients with insulin-dependent diabetes and normoalbuminuria or microalbuminuria. Lancet 1997; 349: 1787–92. The EURODIAB IDDM Complications Study Group. Microvascular and acute complications in insulin dependent diabetes mellitus. Diabetologia 1994; 37: 278-85. Aldington SJ, Kohner EM, Meuer S, Klein R, Sjolie A-K, The EURODIAB IDDM Complications Study Group. Methodolgy for retinal photography and assessment of diabetic retinopathy: the EURODIAB IDDM Complications Study. Diabetologia 1995; 38: 437–44. John GW, Gray MR, Bates DL, Beacham JL. Enzyme immunoassay: a new technique for estimating HbA1c. Clin Chem 1993; 39: 663–66. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959; 22: 719–48. Ravid M, Savin H, Jutrin I, Bental T, Katz B, Lishner M. Long-term stabilizing effect of angiotensin-converting enzyme inhibition on plasma creatinine and on proteinuria in normotensive type II diabetic patients. Ann Intern Med 1993; 118: 577–81. Woolf B. On estimating the relation between blood group and disease. Am Hum Genet 1995; 19: 251–53. Jackson WE, Holmes DL, Garg SK, Harris S, Chase HP. Angiotensinconverting enzyme inhibitor therapy and diabetic retinopathy. Ann Ophthalmol 1992; 24: 99–103. The DCCT Research Group. The Diabetes Control and Complications Trials (DCCT): design and methodologic consideration for the feasibility phase. Diabetes 1986; 35: 530–44. Feman SS, Mericle RA, Reed GW, May JM, Workman RJ. Serum angiotensin converting enzyme in diabetic patients. Am J Med Sci 1993; 305: 280–84. Parving HH, Larsen M, Hommel E, Lund Andersen H. Effect of antihypertensive treatment on blood-retinal barrier permeability to fluorescein in hypertensive type 1 (insulin-dependent) diabetic patients with background retinopathy. Diabetologia 1989; 32: 440–44. Danser AH, Derkx FH, Admiraal PJ, Deinum J, de Jong PT, Schalekamp MA. Angiotensin levels in the eye. Invest Ophthalmol Vis Sci 1994; 35: 1008–18. Caldwell PRB, Seegal BC, Hsu KC, Das M, Soffer R. Angiotensin converting enzyme, vascular endothelial localization. Science 1976; 191: 1050–51. Engler CB, Parving H-H, Mathiesen ER, Larsen M, Lund-Andersen H. Blood-retina barrier permeability in diabetes during ACE inhibition. Acta Ophthalmol 1991; 69: 581–85. Herman WH, Dasbach EJ, Songer TJ, Thompson DE, Crofford OB. Assessing the impact of intensive insulin therapy on the health care system. Diabetes Rev 1994; 2: 384–88. The DCCT Research Group. Epidemiology of severe hypoglycaemia in the Diabetes Control and Complications Trial. Am J Med 1991; 90: 450–59.
31
Early Report
Effect of lisinopril on progression of retinopathy in normotensive people with type 1 diabetes Nish Chaturvedi, Anne-Katrin Sjolie, Judith M Stephenson, Heidemarie Abrahamian, Marc Keipes, Allesandro Castellarin, Zeljka Rogulja-Pepeonik, John H Fuller, and the EUCLID Study Group*
Summary
Introduction
Background Retinopathy commonly occurs in people with type 1 diabetes. Strict glycaemic control can decrease development and progression of retinopathy only partially. Blood pressure is also a risk factor for microvascular complications. Antihypertensive therapy, especially with inhibitors of angiotensin-converting enzyme (ACE), can slow progression of nephropathy, but the effects on retinopathy have not been established. We investigated the effect of lisinopril on retinopathy in type 1 diabetes.
Retinopathy eventually develops in 70–100% of people with type 1 diabetes (insulin-dependent diabetes mellitus) and remains a common cause of visual impairment and eventual blindness.1–3 The only intervention that prevents development and slows progression of retinopathy is tight glycaemic control.4 However, this approach is not wholly effective, and other interventions must be sought. Blood pressure is an important risk factor for the development of retinopathy.5–7 Antihypertensive therapy, especially inhibitors of angiotensin-converting enzyme (ACE), slow the progression of nephropathy,8,9 but whether these agents have a beneficial effect on retinopathy is much less clear. Previous randomised controlled trials of these drugs in people with type 1 diabetes either did not collect standardised data on retinopathy status,9 or did not have sufficient power10,11 to show an effect. We report the retinopathy results from the EUCLID (EURODIAB Controlled Trial of Lisinopril in Insulin-Dependent Diabetes Mellitus) study.12
Methods As part of a 2-year randomised double-blind placebo-controlled trial, we took retinal photographs at baseline and follow-up (24 months) in patients aged 20–59 in 15 European centres. Patients were not hypertensive, and were normoalbuminuric (85%) or microalbuminuric. Retinopathy was classified from photographs on a five-level scale (none to proliferative). Findings The proportion of patients with retinopathy at baseline was 65% (117) in the placebo group and 59% (103) in the lisinopril group (p=0·2). Patients on lisinopril had significantly lower HbA1c at baseline than those on placebo (6·9% vs 7·3 p=0·05). Retinopathy progressed by at least one level in 21 (13·2%) of 159 patients on lisinopril and 39 (23·4%) of 166 patients on placebo (odds ratio 0·50 [95% CI 0·28–0·89], p=0·02). This 50% reduction was the same when adjusted for centre and glycaemic control (0·55 [0·30–1·03], p=0·06). Lisinopril also decreased progression by two or more grades (0·27 [0·07–1·00], p=0·05), and progression to proliferative retinopathy (0·18 [0·04–0·82], p=0·03). Progression was not associated with albuminuric status at baseline. Treatment reduced retinopathy incidence (0·69 [0·30–1·59], p=0·4). Interpretation Lisinopril may decrease retinopathy progression in non-hypertensive patients who have type 1 diabetes with little or no nephropathy. These findings need to be confirmed before changes to clinical practice can be advocated.
Lancet 1998; 351: 28–31
*Listed at end of paper EURODIAB, Department of Epidemiology and Public Health, University College London, London WC1E 6BT, UK (N Chaturvedi MRCP, J M Stephenson MD, J H Fuller FRCP); Aarhus University Hospital, Aarhus, Denmark (A-K Sjolie MD); Hospital Vienna Lainz and L Boltzmann-Institute for Nutrition and Metabolic Diseases, Vienna, Austria (H Abrahamian MD); Maladies du Metabolisme et de la Nutrition, Luxembourg (M Keipes MD); Università de Verona, Verona, Italy (A Castellarin MD); and Vuk Vrhovac Institute, Zagreb, Croatia (Z Rogulja-Pepeonik MD) Correspondence to: Dr Nish Chaturvedi (e-mail: [email protected])
28
Methods The EUCLID study was a double-blind, randomised, paralleldesign clinical trial of lisinopril, an ACE inhibitor, and placebo in 18 European centres, in 530 patients. Men and women with type 1 diabetes13 aged 20–59 years were recruited if resting blood pressure was 75–90 mm Hg diastolic, and 155 mm Hg or less systolic. Most of the patients were normoalbuminuric (albuminexcretion rate <20 µg/min); 15% were microalbuminuric (albumin-excretion rate 20–200 µg/min). Trained photographers took retinal photographs of two 45–50° fields of each eye—macula temporal and disc nasal—at baseline and at 24 months. Slides were sent to London, UK, and then to Aarhus, Denmark, for classification on a five-level scale. AKS assessed all photographs according to the EURODIAB protocol, based on the modified Airlie House classification.14 She had no access to information about patients, except study number. Concordance between the EURODIAB and other methods of grading is high.14 Three centres (Bucharest, Krakow, and Watford) did not have retinal cameras and did not participate in the retinopathy section of the study. Therefore, photographs were taken in 409 patients, 354 of whom had photographs at baseline and follow-up. Patients were re-examined at 1, 3, 6, 12, 18, and 24 months. Blood samples were taken every 6 months to measure glycated haemoglobin (HbA1c). Patients were started on 10 mg lisinopril or matching placebo daily at randomisation. The dose was increased to 20 mg if diastolic blood pressure was more than 75 mm Hg at 3 months. HbA1c was estimated in London by EIA with a monoclonal antibody raised against HbA1c (Dako Ltd, Ely, UK); the normal range for this assay is 2·9–4·8%.15 We estimated that 500 patients would be needed to show a difference between groups in urinary albumin excretion rate, the primary endpoint. Retinopathy was a secondary endpoint, and we estimated the statistical power of the trial to detect a treatment difference in degree of retinopathy. The WESDR showed that retinopathy progressed over 4 years in 48% of
THE LANCET • Vol 351 • January 3, 1998
EARLY REPORT
530 randomised 121 had no baseline photographs No camera (96) Ungradable (7) No photograph (18) 409 had gradable baseline retinal photographs 207 assigned placebo 28 had no final photograph Photograph lost (9) Withdrawn (17) Violated protocol (2)
202 assigned lisinopril 27 had no final photograph Photograph lost (7) Withdrawn (17) Violated protocol (3)
179 had final photographs
175 had final photographs
168 completed trial as per protocol
156 completed trial as per protocol
patients with young-onset diabetes with a diastolic blood pressure of between 79 mm Hg and 85 mm Hg and nonproliferative retinopathy at baseline. We estimated that retinopathy would progress in 24% of the placebo group over 2 years. We used data from the EURODIAB IDDM Complications Study to estimate the likely rate of retinopathy at baseline.13 With the assumption that we would have 250 patients in each group, of whom half would have retinopathy (non-proliferative) at baseline, we would be able to detect a reduction in retinopathy progression from 24% on placebo to 10% on lisinopril (odds ratio 0·35) with 80% power at 5% significance. Patients were assigned to one of five retinopathy levels by the worst affected eye.14 These assignments were based on lesions graded: none, minimal, moderate, severe non-proliferative, and photocoagulation and proliferative combined. The main outcome for retinopathy was progression of retinopathy by at least one grade. Those who could not show progression (photocoagulated and proliferative at baseline), were excluded from these analyses. Patients whose retinopathy progressed by at least two grades or to proliferative retinopathy were also compared. We compared the development of retinopathy in patients with no retinopathy at baseline, and regression of retinopathy for all patients with any retinopathy at baseline. We calculated odds ratios for progression, development, and regression to compare patients in the two treatment groups by the Mantel Haenszel test.16 Logistic regression was used to adjust these odds ratios for confounders, such as centre, HbA1c, baseline albumin excretion rate, duration of diabetes, age, sex, baseline blood pressure, and baseline retinopathy status; only patients with complete data for confounders were included in these analyses. In some cases, because of small numbers, adjustment could not be made for centre. Between-observer and within-observer variations for retinal grading were assessed by asking the original grader (AKS), and an independent grader (S Aldington, Retinopathy Grading Centre, Hammersmith Hospital, London, UK), to regrade 60 sets of baseline and follow-up photographs for 30 (8·5%) patients randomly selected from all patients who had both sets of photographs. We calculated a weighted score to assess the
None Minimal non-protective Moderate non-proliferative Severe non-proliferative Photocoagulated or proliferative
Lisinopril
Retinopathy
Retinopathy
None M/F Age (years) Duration of diabetes (years) % current smokers Body-mass index (kg/m2) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) HbA1c (%) Albumin excretion rate (µg/min)‡ % with microalbuminuria
41/21 31 (7) 9 (6)
Any
All
79/38 120/59 37 (8)* 35 (8) 18 (7) 15 (8)
35 33 34 23·6 25·2* 24·7 (2·3) (2·8) (2·7) 121 (11) 124 (11) 123 (11) 81 (4) 7·1 (1·8) 7·2 (4·7– 12·4) 13
81 (5) 7·4 (2·0) 7·4 (4·7– 15·2) 15
81 (5) 7·3 (1·9) 7·2 (4·7– 14·7) 14
None
Any
All
42/30 69/34 111/64 32 (8) 36 (10)* 34 (9) 10 (8) 17 (8)* 14 (9) 31 20† 25 24·0 25·2* 24·7 (2·6) (2·9) (2·8) 123 (10) 123 (10) 123 (10) 80 (4) 6·7 (2·2) 5·7 (3·4– 11·6) 10
81 (5) 7·0 (1·6) 7·9 (4·5– 17·1) 22†
81 (5) 6·9† (1·9) 6·9 (3·9– 3·3) 17
p values refer to comparisons between those with and without retinopathy within treatment group, and between treatments for the all category. *p<0·01. †p=<0·05. ‡Median (IQR).
Table 2: Baseline characteristics by retinopathy level and treatment group (mean [SD])
Figure 1: Trial profile
Retinopathy status
Placebo
Treatment Placebo (n=179)
Lisinopril (n=175)
62 (35%) 81 (45%) 23 (13%) 0 13 (7%)
72 (41%) 67 (38%) 18 (10%) 2 (1%) 16 (9%)
Table 1: Retinopathy status at baseline by treatment group
THE LANCET • Vol 351 • January 3, 1998
degree of within-observer and between-observer agreement. Odds ratios for progression were calculated for the three previous trials of ACE inhibitors and diabetic retinopathy,10,11,17 and a combined odds ratio was estimated by the method of Woolf.18
Results 409 of 530 patients had gradable baseline photographs. 121 patients did not have photographs. Therefore, gradable photographs were obtained at baseline and follow-up for 354 (81·6%) of 434 eligible patients (figure 1). The only significant differences in baseline characteristics between patients with and without photograhs were for diastolic blood pressure (81 vs 79 mm Hg in patients with and without photographs; p=0·0001), and distribution of sex (47 vs 37% women of patients with and without photographs; p=0·04). These differences were consistent when assessed separately by treatment group. Patients with both sets of photographs were older (34 vs 31 years, p=0·005), had had diabetes for longer (15 vs 12 years, p=0·02), had higher systolic blood pressure (123 vs 120 mm Hg, p=0·03) and better glycaemic control (HbA1c 7·1 vs 7·8%, p=0·08), and fewer were women (35% vs 51%, p=0·02) compared with those who only had baseline photographs. These differences were consistent when assessed by treatment group separately. Distribution of retinopathy at baseline was similar by treatment for patients with only baseline photographs. Retinopathy was present in 117 (65%) patients in the placebo group and 103 (59%) in the lisinopril group (p=0·2). Most cases of retinopathy were minimal nonproliferative (table 1). In both treatment groups, patients with retinopathy were older, had had diabetes for longer, and had higher body-mass indices than those without retinopathy; in the lisinopril group, those with retinopathy had a higher rate of microalbuminuria (table 2). HbA1c was significantly lower at baseline in the lisinopril group than in the placebo group (6·9 vs 7·3%; p=0·05). HbA1c did not change during the trial in either treatment group.12 Retinopathy progressed by at least one level in the worst eye in 39 (23·4%) of 166 patients in the placebo group, and 21 (13·2%) of 159 in the lisinopril group (p=0·02; table 3). Progression by one level was 29
EARLY REPORT Adjustment
Lisinopril
Odds ratio (95% CI)
p
Progression by one level* Unadjusted 39/166 Centre .. .. HbA1c Full† ..
Placebo
21/159 .. .. ..
0·50 (0·28–0·89) 0·50 (0·28–0·92) 0·55 (0·30–1·03) 0·55 (0·29–1·05)
0·02 0·03 0·06 0·07
Progression by two levels* Unadjusted 11/166 .. HbA1c
3/157 ..
0·27 (0·07–1·00) 0·30 (0·08–1·10)
0·05 0·07
Progression to proliferative retinopathy* Unadjusted 11/166 2/159 .. .. HbA1c
0·18 (0·04–0·82) 0·20 (0·04–0·91)
0·03 0·04
Incidence of retinopathy‡ Unadjusted 15/62 .. HbA1c
13/72 ..
0·69 (0·30–0·59) 0·73 (0·31–1·74)
0·4 0·5
Regression of retinopathy§ Unadjusted 28/117 .. HbA1c
33/103 ..
1·48 (0·82–2·68) 1·41 (0·78–2·58)
0·2 0·3
*Number of progressors/number in whom progression was possible. †Adjustment for centre, HbA1c, age, duration of diabetes, sex, baseline albumin excretion rate, baseline retinopathy grade, baseline blood pressure. ‡Number with incident retinopathy/number without retinopathy at baseline. §Number with regression/number with retinopathy at baseline.
Table 3: Progression, development, and regression of retinopa-
seen equally at all baseline levels of retinopathy. In patients on lisinopril compared with placebo the odds ratio adjusted for centre was 0·50 (95% CI 0·28–0·92, p=0·03). On logistic regression only HbA1c was related to progression of retinopathy. Adjustment for HbA1c at baseline made the odds ratio 0·55 (0·30–1·03, p=0·06; table 3). There was no evidence of a significant interaction by HbA1c (p=0·4). However, a stratified analysis by HbA1c with a cut-off of 7%, the median value, gave an odds ratio related to treatment of 0·34 (0·13–0·94; p=0·04) in patients with HbA1c less than 7%, and 0·61 (0·29–1·29; p=0·2) in those with HbA1c of 7% or more. There was no evidence of other interactions. Systolic blood pressure was 3 mm Hg lower in the lisinopril group than in the placebo group at 1 month, and this difference persisted until the end of the trial.12 We adjusted the model with all other potential confounders for systolic blood pressure at 1 month and the odds ratio was 0·57 (0·30–1·10; p=0·09). Progression by two levels, and progression to proliferative retinopathy was also reduced by treatment after adjustment for HbA1c (table 3). There was disagreement on the level of retinopathy for (8·2)
Larsen et al11
Chase et al10 Ravid et al17
This study Meta-analysis
0
1
2 3 Odds ratio
4
Figure 2: Odds ratios (95% CI) for progression of retinopathy for both groups in EUCLID and previous studies
30
six of 60 photographs when regraded by the same observer. These photographs were equally distributed between baseline and final visit and treatment groups. Two sets were assigned to a worse retinopathy level than originally, and four to a better level. The within-observer agreement () was 0·85 (0·67–1·00), and the betweenobserver was 0·55 (0·39–0·71). Individual odds ratios for the previous three studies all indicate a beneficial, but non-significant effect of ACE inhibitors on retinopathy (figure 2). Combination of these odds ratios produces an overall odds ratio of 0·46 (0·19–1·10; 2 for heterogeneity 0·57, p>0·5). A metaanalysis of all studies, including our study, gave 2 for heterogeneity between studies of 0·59 (p>0·5), with a combined odds ratio of 0·49 (0·30–0·79).
Discussion We show that lisinopril had beneficial effects on progression of retinopathy. Non-significant benefits have been shown previously with other ACE inhibitors in people with type 110,11,19 and type 2 (non-insulin-dependent diabetes mellitus).17 The analyses for retinopathy were included in the EUCLID protocol, but the study was not primarily designed to assess the effects of ACE inhibitors on retinopathy. However, retinal photographs taken by standard methods are an accepted way of assessing changes in retinopathy status,20 and the EURODIAB protocol is well validated.14 Not all patients in our trial had gradable photographs, but we did not find notable differences in baseline characteristics between those who did and did not have photographs. We could analyse only data on patients who completed the trial and, therefore, had better-than-average compliance. Withdrawal rates were similar in the two treatment groups and, therefore, a differential compliance by treatment is unlikely to explain our findings. The within-observer and between-observer differences in the grading of the retinal photographs were not systematically biased by treatment group. Confounders such as glycaemic control can account for only part of the effect of lisinopril on retinopathy progression, and other explanations are needed. Serum concentrations of ACE inhibitor are positively related to the presence and degree of retinopathy,21 and captopril, an ACE inhibitor, limits the abnormal leak of fluorescein from retinal vessels seen in diabetic retinopathy.11,12 This limitation may be due to a direct effect on systemic blood pressure, but studies show that ACE is produced locally by vascular endothelial cells,23,24 which may have direct detrimental effects on retinal flow and vascular structure, independent of changes in systemic blood pressure.25 A direct effect of ACE inhibitors on the eye could be why we found that the retinal effects of lisinopril cannot be accounted for by renal effects. The slowing of progression of retinopathy to proliferative disease, once established, has been by improvement of glycaemic control,4 but control is difficult to achieve26,27 and a proportion of patients will still develop or have advanced retinopathy. Good glycaemic control halves the progression of retinopathy over a 2-year period (progression of one of five levels in EUCLID vs three of 25 in the Diabetes Control and Complications Trial). In EUCLID, however, we found that patients with better glycaemic control had the most benefit from ACEinhibitors, combination may be the best therapeutic
THE LANCET • Vol 351 • January 3, 1998
EARLY REPORT
approach. We show that the beneficial effects of lisinopril seem not to depend on the presence of renal disease; even those without microalbuminuria seemed to benefit. Restriction of ACE inhibitors to patients at risk of retinopathy would exclude many who may benefit. ACE inhibitors may have an important role in limiting the progression of retinopathy in patients with type 1 diabetes, but these findings, and the effects of this treatment on the development and regression of retinopathy, must be studied further in large, randomised, controlled trials before changes in clinical practice are advocated. EUCLID investigators B Karamanos, A Kofinis, C Petrou, Hippokration Hospital, Athens, Greece; C Ionescu-Tirgovisite, C Iosif, Clinic of Diabetes Nutrition and Metabolic Diseases, Bucharest, Romania; G Tamas, G Bibok, 1st Department of Medicine, Semmelweis University, Budapest, Hungary; Z Kerenyi, P Kis-Gombos, J Toth, Szent Imré Teaching Hospital, Budapest, Hungary; J B Ferriss, G Grealy, Cork University Hospital, Wilton, Cork, Ireland; R Rottiers, H Priem, University Hospital, Gent, Belgium; V Koivisto, J Tuominen, E Kostamo, Helsinki University Hospital, Helsinki, Finland; B Idzior-Walus, B Solnica, D Galicka-Latalie, University School of Medicine, Krakow, Poland; G Michel, M Keipes, A Giuliani, A Herode, Hospitalier de Luxembourg, Luxembourg; F Santeusanio, A Bueti, S Bistoni, Dr Cagini, Instituto di Medicina Interna E Scienze Endocrine E Metaboliche, Perugia, Italy; R Navalesi, G Penno, M Nannipieri, L Rizzo, R Miccoli, Instituto di Clinica Medica II, Pisa, Italy; G Ghirlanda, P Cotroneo, A Manto, A Minella, C Saponara, Universita Cattolica del Sacro Cuore, Rome, Italy; J Ward, M Plater, S Ibrahim, S Ibbotson, C Mody, Royal Hallamshire Hospital, Sheffield, UK; N Papazoglou, C Manes, K Soulis, M Voukias, Agios Pavlos General Hospital, Thessaloniki, Greece; M Muggeo, V Cacciatori, M L Gemma, A Dellera, A Castellarin, Ospedale Civile, Verona, Italy; K Irsigler, H Abrahamian, C Gurdet, C Willinger, Hospital Vienna Lainz and L Boltsmanninstitute for Nutrition and Metabolic Diseases, Vienna, Austria; A Nelstrop, C Feben, Watford General Hospital, Watford, UK; S Walford, V McLelland, S Hughes, New Cross Hospital, Wolverhampton, UK; Z Metelko, G Roglic, Dr Z Rogulja Pepeonik, Vuk Vrhovak Institute for Diabetes, Zagreb, Croatia. Steering Committee Members—B Ferriss, chairman, Cork; J H Fuller, London; B Karamanos, Athens; Z Kerenyi, Budapest; G Michel, Luxembourg; M Muggeo, Verona; J Stephenson, London; N Chaturvedi, London; G C Viberti, London; A K Sjolie, Aarhus. Coordinating Centre—J H Fuller, J Stephenson, N Chaturvedi, J Holloway, M Milne, D Webb. Retinopathy Grading Centre—A K Sjolie, Aarhus, Denmark. Central Laboratory—G John, D J Newman, The Royal London Hospital and St Bartholomew’s and The Royal London Hospital School of Medicine and Dentistry, London, UK.
Contributors Nish Chaturvedi jointly supervised the project, carried out analysis, and drafted the paper. Anne-Katrin Sjolie advised on the design of the retinopathy section of the study, graded all retinopathy, and assisted in writing the paper. Judith Stephenson devised the original idea for the project, designed the study protocol, jointly supervised the project, advised on statistical analysis, and helped to draft the paper. Heidemarie Abrahamian and Marc Keipes recruited and assessed patients at their centres and participated in the drafting of the paper. Allesandro Castellarin and Zeljka Rogulja-Pepeonik took the retinal photographs and commented on the paper. John Fuller was the team leader, assisted in the design of the protocol, jointly supervised the project, advised on statistical analyses, and assisted with the drafting of the paper.
Acknowledgments This study was supported by a grant from Zeneca Pharmaceuticals, who also provided the lisinopril and placebo tablets. We thank the EUCLID investigators, Rebecca Hardy, statistician, MRC National Survey of Health and Development, University College, London, who performed the meta-analysis, Steve Aldington and staff at the Retinopathy Grading Centre, Hammersmith Hospital, London, UK, and the people who took part in the study for their collaboration.
3
4
5
6
7
8
9
10
11
12
13
14
15 16 17
18 19
20
21
22
23
24
25
References 1
2
Dwyer MS, Melton III LJ, Ballard DJ, Palumbi PJ, Trautman JC, Chu CP. Incidence of diabetic retinopathy and blindness: a populationbased study in Rochester, Minnesota. Diabetes Care 1985; 8: 316–22. Klein R, Klein BEK, Moss SE, Davis MD, DeMets DL. The Wisconsin Epidemiologic Study of diabetic retinopathy II prevalence and risk of diabetic retinopathy when age at diagnosis is less than 30 years. Arch Ophthalmol 1984; 102: 520–60.
THE LANCET • Vol 351 • January 3, 1998
26
27
Sjolie AK, Stephenson J, Aldington S, et al. Retinopathy and vision loss in insulin-dependent diabetes in Europe. Ophthalmology 1997; 104: 252–60. The Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 1993; 329: 977–86. Klein R, Klein BEK, Moss SE, Davis MD, DeMets DL. Is blood pressure a predictor of the incidence or progression of diabetic retinopathy? Arch Intern Med 1989; 149: 2427–32. Janka HU, Warram JH, Rand LI, Krolewski AS. Risk factors for progression of background retinopathy in long-standing IDDM. Diabetes 1989; 38: 460–64. Teuscher A, Schnell H, Wilson PWF. Incidence of diabetic retinopathy and relationship to baseline plasma glucose and blood pressure. Diabetes Care 1988; 11: 246–51. Lewis EJ, Hunsicker LG, Bain RP, Rohde RD, and the Collaborative Study Group. The effect of angiotensin converting enzyme inhibition on diabetic nephropathy. N Engl J Med 1993; 329: 1456–62. The Microalbumin Captopril Study Group. Captopril reduces the risk of nephropathy in IDDM patients with microalbuminuria. Diabetologia 1996; 39: 587–93. Chase HP, Garg SK, Harris S, Hoops S, Jackson WE, Holmes DL. Angiotensin-converting enzyme inhibitor treatment for young normotensive diabetic subjects: a two-year trial. Ann Ophthalmol 1993; 25: 284–89. Larsen M, Hommel E, Parving HH, Lund-Andersen H. Protective effect of captopril on the blood-retina barrier in normotensive insulindependent diabetic patients with nephropathy and background retinopathy. Graefes Arch Clin Exp Ophthalmol 1990; 228: 505–09. The EUCLID study group. Randomised placebo-controlled trial of lisinopril in normotensive patients with insulin-dependent diabetes and normoalbuminuria or microalbuminuria. Lancet 1997; 349: 1787–92. The EURODIAB IDDM Complications Study Group. Microvascular and acute complications in insulin dependent diabetes mellitus. Diabetologia 1994; 37: 278-85. Aldington SJ, Kohner EM, Meuer S, Klein R, Sjolie A-K, The EURODIAB IDDM Complications Study Group. Methodolgy for retinal photography and assessment of diabetic retinopathy: the EURODIAB IDDM Complications Study. Diabetologia 1995; 38: 437–44. John GW, Gray MR, Bates DL, Beacham JL. Enzyme immunoassay: a new technique for estimating HbA1c. Clin Chem 1993; 39: 663–66. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959; 22: 719–48. Ravid M, Savin H, Jutrin I, Bental T, Katz B, Lishner M. Long-term stabilizing effect of angiotensin-converting enzyme inhibition on plasma creatinine and on proteinuria in normotensive type II diabetic patients. Ann Intern Med 1993; 118: 577–81. Woolf B. On estimating the relation between blood group and disease. Am Hum Genet 1995; 19: 251–53. Jackson WE, Holmes DL, Garg SK, Harris S, Chase HP. Angiotensinconverting enzyme inhibitor therapy and diabetic retinopathy. Ann Ophthalmol 1992; 24: 99–103. The DCCT Research Group. The Diabetes Control and Complications Trials (DCCT): design and methodologic consideration for the feasibility phase. Diabetes 1986; 35: 530–44. Feman SS, Mericle RA, Reed GW, May JM, Workman RJ. Serum angiotensin converting enzyme in diabetic patients. Am J Med Sci 1993; 305: 280–84. Parving HH, Larsen M, Hommel E, Lund Andersen H. Effect of antihypertensive treatment on blood-retinal barrier permeability to fluorescein in hypertensive type 1 (insulin-dependent) diabetic patients with background retinopathy. Diabetologia 1989; 32: 440–44. Danser AH, Derkx FH, Admiraal PJ, Deinum J, de Jong PT, Schalekamp MA. Angiotensin levels in the eye. Invest Ophthalmol Vis Sci 1994; 35: 1008–18. Caldwell PRB, Seegal BC, Hsu KC, Das M, Soffer R. Angiotensin converting enzyme, vascular endothelial localization. Science 1976; 191: 1050–51. Engler CB, Parving H-H, Mathiesen ER, Larsen M, Lund-Andersen H. Blood-retina barrier permeability in diabetes during ACE inhibition. Acta Ophthalmol 1991; 69: 581–85. Herman WH, Dasbach EJ, Songer TJ, Thompson DE, Crofford OB. Assessing the impact of intensive insulin therapy on the health care system. Diabetes Rev 1994; 2: 384–88. The DCCT Research Group. Epidemiology of severe hypoglycaemia in the Diabetes Control and Complications Trial. Am J Med 1991; 90: 450–59.
31