Downloaded from on May 17, 2017 - Published by group.bmj.

BJO Online First, published on March 7, 2017 as 10.1136/bjophthalmol-2016-310063
Clinical science

Differences in incidence of diabetic retinopathy
between type 1 and 2 diabetes mellitus:
a nine-year follow-up study
Pedro Romero-Aroca,1,2 Raul Navarro-Gil,1,2 Aida Valls-Mateu,3
Ramon Sagarra-Alamo,4 Antonio Moreno-Ribas,3 Nuria Soler1,2
Ophthalmology Service, ABSTRACT both DM types. DR screening uses a non-mydriatic
University Hospital Sant Joan, Background/aims To determine the incidence of any fundus camera, a cost-effective way of screening
Reus, Spain
Institut de Investigacio diabetic retinopathy (any-DR), sight-threatening diabetic DM populations.4 Screening frequency varies
Sanitaria Pere Virgili (IISPV), retinopathy (STDR) and diabetic macular oedema (DMO) according to DM type.5 Our group rolled out a
Universitat Rovira and Virgili, and their risk factors in type 1 diabetes mellitus (T1DM) screening programme in 2000 that included general
Reus, Spain
over a screening programme. practitioners and endocrinologists,6 and we
Department of Computer
Methods Nine-year follow-up, prospective population- reported an increase in the incidence of DR in a
Engineering and Mathematics,
Universitat Rovira and Virgili, based study of 366 patients with T1DM and 15 030 previously published study.7
Reus, Spain with T2DM. Epidemiological risk factors were as follows: In this study, we determine the incidence of
Health Care Area Reus-Priorat, current age, age at DM diagnosis, sex, type of DM, any-DR, sight-threatening retinopathy (STDR) and
Institut Catala de la Salut duration of DM, arterial hypertension, levels of diabetic macular oedema (DMO) in patients with
(ICS), Institut de Investigació
Sanitària Pere Virgili (IISPV),
glycosylated haemoglobin (HbA1c), triglycerides, T1DM and its differences in patients with T2DM.
Universitat Rovira and Virgili, cholesterol fractions, serum creatinine, estimated
Reus, Spain glomerular filtration rate (eGFR) and urine albumin to MATERIALS AND METHODS
creatinine ratio (UACR). Setting: The reference population in our area is
Correspondence to
Results Sum incidence of any-DR was 47.26% with 247 174. The total number of patients with DM
Dr Pedro Romero-Aroca
Hospital Universtario Sant annual incidence 15.16±2.19% in T1DM, and 26.49% registered with our healthcare area is 17 792 (7.1%).
Joan, Avenida, Doctor, Josep with annual incidence 8.13% in T2DM. Sum incidence Design: A prospective, population-based study,
Laporte 2, Reus 43204, Spain; of STDR was 18.03% with annual incidence 5.77 conducted from 1 January 2007 to 31 December, ±1.21% in T1DM, and 7.59% with annual incidence 2015. A total of 366 patients with T1DM and 15
2.64±0.15% in T2DM. Sum incidence of DMO was 030 with T2DM were screened.
Received 14 December 2016 8.46% with annual incidence 2.68±038% in patients Power of the study: Our epidemiologist estimates
Revised 31 January 2017 with T1DM and 6.36% with annual incidence 2.19 the detection of a ±3% increase in risk and 95%
Accepted 10 February 2017 ±0.18% in T2DM. Cox’s survival analysis showed that accuracy.
current age and age at diagnosis were risk factors at Method: Screening for DR was carried out with
p<0.001, as high HbA1c levels at p<0.001, LDL one 45° field retinography, centred on the fovea. If
cholesterol was significant at p<0.001, eGFR was DR was suspected, a total of nine retinographies of
significant at p<0.001 and UACR at p=0.017. 45° were taken and a complete screening is described
Conclusions The incidence of any-DR and STDR was elsewhere.8 Due to the difficulty in obtaining images
higher in patients with T1DM than those with T2DM. from patients with T1DM under 12 years old, only
Also, the 47.26% sum incidence of any-DR in patients those aged >12 years were included.
with T1DM was higher than in a previous study In this study, DR is classified into (i) no-DR, (ii)
(35.9%), which can be linked to poor metabolic control any-DR—level 20–35 of the ETDRS, (iii) STDR—
of DM. Our results suggest that physicians should be defined as level 43 or worse by the ETDRS. The
encouraged to pay greater attention to treatment term ‘DMO’ includes ‘extrafoveal’ and/or ‘clinically
protocols for T1DM in patients. significant macular oedema (CSMO)’ according to
the ETDRS classification.9
Measures of kidney diabetic disease were deter-
INTRODUCTION mined by (i) serum creatinine; (ii) estimated glom-
It is estimated that more than 200 million people erular filtration rate (eGFR), measured by the
worldwide currently have diabetes and that number chronic kidney disease epidemiology collaboration
is predicted to rise by over 120% by 2025.1 It has equation (CKD-EPI); (iii) urine albumin to creatin-
become a chronic disease with several complica- ine ratio (UACR), classified in normoalbuminuria
tions. Diabetes mellitus (DM) is classified as type 1 defined as UACR <30 mg/g, microalbuminuria as
diabetes (T1DM) or type 2 diabetes (T2DM), ges- UACR 30–299 mg/g and macroalbuminuria as
tational diabetes, monogenic diabetes and second- UACR ≥300 mg/g.
To cite: Romero-Aroca P, ary diabetes.2 There is a current trend towards At the end of the study, all patients with T1DM
Navarro-Gil R, Valls- more children developing T1DM and more than were visited, and a fundus nine-field retinographies
Mateu A, et al. Br J
Ophthalmol Published
half a million children are estimated to be living was carried out by an ophthalmologist to confirm
Online First: [ please include with the disease. the number of patients with DR and if any new
Day Month Year] The most important ocular complication is dia- patients with DR are previously not diagnosed.
doi:10.1136/bjophthalmol- betic retinopathy (DR), a common cause of blind- Inclusion criteria: Patients with T1DM >12
2016-310063 ness in Europe.3 Development of DR is similar in years old, and all patients with T2DM.
Romero-Aroca P, et al. Br J Ophthalmol 2017;0:1–6. doi:10.1136/bjophthalmol-2016-310063 1
Copyright Article author (or their employer) 2017. Produced by BMJ Publishing Group Ltd under licence.

49% 15.6% 2.7 7. qualitative data was made by the χ2 table and the determination of the Fisher test for quantitative data.63% 5.71–15 5.12 for each patient with Ethical adherence: The study was carried with the approval of T2DM.39±12. DMO.60% 56.8 4–15.7 65. diabetic macular oedema.4 8.64±1. in 74 (47.88% 7.77±1.38±1.67% 2.67 12.1136/bjophthalmol-2016-310063 . diabetic retinopathy. Differences <5 years present DR and HbA1c percentages decrease and in were examined using the two-tailed Student’s t-test to compare patients with any-DR with an increase in DM duration. It is interest- values.74% T2DM 390 384 411 424 407 533 489 529 415 Any-DR 7.62± Clinical science Exclusion criteria: Patients with other specific types of dia.1 33. 66 (61.58% 65.72±8.3 8.1±10.11% 65.001).49 8.40% 2.33% 56.42% 5. Table 2 shows by the determination of mean.48 6.32 65.4 7.11±10.61±1. et al.13% 6.84% 2.49% 2.19±1. Each patient with T1DM was screened 4.71 12. SD.9 4.48% 2.9–14 4.25% T1DM STDR 6 7 7 8 7 9 7 8 7 5. and the 95% CI.11 35.77 T1DM mean HbA1c 8. DR.63 ±8.05% T1DM mean age 33.25% 5. By current age.78 8. Excluding RESULTS differences in age. Multivariate analysis was Study of differences between patients with T1DM and carried out using Cox survival regression analysis. Table 1 Descriptive values of the sample Year 2007 2008 2009 2010 2011 2012 2013 2014 2015 Number of patients 117 116 121 124 121 144 129 142 127 with T1DM Number of patients 4910 4873 5191 5243 5264 6193 5494 5983 5026 with T2DM T1DM men 75 76 79 81 78 93 84 93 83 64.77 times over betes.38±1.7 5.58% 2. T1DM.36% 2.97% 15.02 35.00% 2.001.29% in patients with T2DM.76% T1DM DMO 2 3 3 4 3 5 3 4 4 1.3–15.34±6.9–15 4.57±6.5 4.8–15 4.08±10. men are more frequent in both DM types but Demographic variables of sample size less in T1DM.8 4.2–15 Incidence of DR and its severity.22±12.86±8.66±6.83 8.37±6.73% 8.07 65.4 7. and patients with gestational DM.33±1.78% 6.58±10. we used the ana.67% 15. being significant at p<0.32±1.82±1.78% on May 17.05 was considered to indi. ation ( p<0.63±1. T2DM.47 8.81% 3.69 12.5 7.31±1. age at Helsinki.5 4–15.1–14. sight-threatening diabetic retinopathy.81 8.22±10.32% 64. Mean HbA1c values were 8.0 20–30 years. Year 2007 2008 2009 2010 2011 2012 2013 2014 2015 T1DM 16 18 17 18 19 23 20 22 20 Any-DR 13.69% 57.32 5–14.78±8.92 8.16% 54. For qualitative data.44 8.81±8.42 years.33±12.25±1.5% 65.5–15.51% 14.86±10. we describe the incidence of DR and its different types.62±1.84±12. diagnosis 22.56 8.84% 8.24 7.69±11.3–15.41 7. the local ethics committee (approval no.31% 56. 2 Romero-Aroca P.08 34.9–14.37±1.1 34.12 65.04% 14.08 65.47±1.64±10.bmj.29 65.49% 65.41 65.81±8.22.11±0.49% 2.29±6.Published by group.12 8.02% 2.77 T2DM duration 8.5–14. 2017 .79±8. T2DM Table 1 shows differences between both DM types.27±12.02±1.47% 3.23±6.08% 2.08% of DR incidence.1% 65. DR did not appear in patients Statistical methods aged <20 years but was present in 27 (39. Downloaded from http://bjo. a total of 366 patients with T1DM and 15030 with the two-tailed Student t-test.05% 7.11 8. minimum and maximum the HbA1c percentages according to DM duration.28±1.69±8.68% 2.14 T2DM mean age 64. the statistical In the 9-year follow-up (1 January 2007 to 31 December analysis of mean differences between T1DM and T2DM.6 4.19±10.47% 3.46% 64.1 T2DM mean HbA1c 7.5% 15.06% 8.88±11.82 8.5 8.19±10.71 12.9% 8.39 T1DM duration 12.74 12. cate statistical significance.17±10. doi:10.3–15.70%) patients aged Data evaluation and analysis was carried out using SPSS V.03% 55.51 8.59±1.5 3.01 33.22 8. Also.23±6.11% 2.94% 7.37–12 3.3±1.51% 15.0:1–6.35±6.47% 2.2 5.28±6.2 5–15.5 7.32% 2. also. ing to observe that 14.15% T2DM DMO 104 101 112 114 110 150 135 153 122 2.11 years and DM duration 13.77 12.22% 2. the 9 years compared with 3.11%) patients aged >40 years.6% 8.74±8.71% 2.42% Values are presented as number or mean±SD and range.77±8.7% 15.001) and HbA1c levels ( p<0.58 12.05 34.03% 5.bmj.87±12.88% 2.14 4.40±1.94 65.51% T2DM STDR 131 125 132 134 141 170 162 174 139 2.45% 5. using 2015). type 1 diabetes mellitus.72% 55.03 35. type 2 diabetes mellitus.16% in patients with Descriptive statistical analysis of quantitative data was made T1DM and 7.1 5–14. STDR.78 12. was significant for diabetes dur- T2DM were screened (table 1).74%) patients aged 30–40 years and in statistical software package and p<0.12% 2.5% 2.29±1.35% T2DM men 2881 2802 2890 3007 2933 3594 3131 3511 2817 57. 13-01-31/proj6) and in Sample characteristics of patients with T1DM at the end of accordance with revised guidelines of the Declaration of study were as follows: current age 35. Inferential analysis for only 81. which two variables or using one-way analysis of variance if we were might explain why patients with >20 years DM duration have comparing more than two variables. Br J Ophthalmol 2017.03 years.41% 56.7% of patients with DM duration lysis of frequency and percentage in each category.04±9.23 66.

therefore with 15–20 30 (66.57±1.08) No DR 7. age at diagnosis DM duration (years) DR incidence patients (%) HbA1c HDL cholesterol and triglycerides are not significant.bmj.68±038% (1.289.56% a time variable changes the univariate statistical study.1136/bjophthalmol-2016-310063 3 . probably due to the oldest Mean 8.06%–8. Table 2 Values of HbA1c and incidence of diabetic retinopathy (DR).45%) and and we did not find any new patient with DR.bmj.001. Romero-Aroca P.51% more time to develop DR.001 and UACR (>30 mg/g)+eGFR 5–10 36 (41.13%–6.21% (5. 3982 patients with T2DM devel.622.49%).01±1. Downloaded from http://bjo. we confirmed that no patient had been misdiagnosed during the screening follow-up. Any-DR 8. Similar age at diagnosis was signifi- Any-DR on May 17.03% At the end of the Clinical science Study of incidence of DR sum incidence in patients with T2DM was 7.53.78% In Cox’s proportional regression analysis (table 4 and figure Mean 8. with an HR value of 12.26%) developed annual incidence of 2.19% the lipid study. Sum incidence of STDR in patients with T1DM was 18.54±0.02. HbA1c Any-DR 10. Gender remains Mean 7.70) No DR 7. All other variables are significant: current age p<0.41% nificant risk variable at p<0.82% Metabolic DM control measured by HbA1c values was a sig- Any-DR 8. the introduction of different variables with DM duration as 10–15 32 (54.47% p<0.48%–2. presence of arterial hypertension p<0.74±1.012. eGFR p<0.37) No DR 8. therefore.94±1.47% cant at p<0.48±2.49%) with a mean annual incidence of 8.93±1. with an annual incidence of 2. UACR Mean 8. of 2. doi:10.68% not significant in the survival analysis.0:1–6. all patients with T1DM were visited. with an annual incidence of 5. male gender.91±1.85±1.Published by group. Current Any-DR 9. according to duration of diabetes mellitus (DM) Statistical analysis at the end of study In the univariate analysis (table 3).71%–3.18% (2%–2.03±1.45±1.73 m2) p<0.59% with an A total of 173 patients with T1DM (47. creatinine p=0.001.001 with an HR value of 90. LDL cholesterol p=0.46% ±2.19% (13.001.41% patients having a longer duration of diabetes. LDL cholesterol remains a significant variable at p<0.19±0.11±2. >20 60 (81.97%).36% with an annual incidence (7.001.9%) (figure 1A and table 1).61% p<0.82% 1B).22%) and oped any-DR (26. 2017 . diabetes duration <5 15 (14. (B) Survival analysis graphs. Br J Ophthalmol 2017.47% age remains significant at p<0. et al.66) No DR 7.77±1.67%–15.39±0.001.86) No DR 7.78% (<60 mL/min/1.19% p<0.93±1. No other lipid variables (HDL Figure 1 (A) Incidence of diabetic retinopathy (DR) and its severity.15% (2.001 and an HR of Sum incidence of DMO in patients with T1DM was 8.64±0.88%).001. In Mean 7.57±1.13% sum incidence in T2DM was 6. any-DR at 9 years with mean annual incidence of 15.

94 (95% CI 1. OR 2. estimated glomerular filtration rate. 2017 .39 Age at diagnosis No DR 22.78) Diabetes duration No DR 11.4 p=0.3±9.84±0. diabetic retinopathy.001. only surpassed by age at diagnosis and T2DM. et al. conclude that 3.001. Also the due to DMO or ischaemic retina secondary to severe DR.63 DR 17.86) Male No DR 46.93 (95% CI 1.6±8.74±25.42 years in patients with T1DM in the T1DM group with annual incidence of 15.61% DR 28.77±0.6±7.16±10.25±6. analysis of variance. doi:10.15 p=0. Fisher-Snedecor distribution. F=3.65 (95% CI 1.175.08±17.75 LDL No DR 96. F=1.86 to 1.901. OR 3.59 p=0.001. if we subtract patients with DMO of total STDR we can were significative at p<0.62 DR 85.19% in T2DM.0:1–6.99 to 6. is difficult to compare ition. On the contrary.25 to 4. OR 2.91±18.525.383.002.23 p=0.bmj. F=8.329.38±1.58) ANOVA.23 (95% CI 0.16 DR 0.91 (95% CI 0.65±0.43 p<0.001.55% of patients with T1DM have STDR due other causes than DMO.bmj.001. In add- compared with 8.35) Creatinine No DR 0.03) eGFR No DR 106.64 DR 110.38±1.46) DR 9.6 p<0. (95% CI 1.004. The renal function study is interesting.48 p<0.3%) in the group the survival analysis ( p=0.1136/bjophthalmol-2016-310063 . with T2DM). there are few studies that deter.142). Br J Ophthalmol 2017.97% DR 48.79±27.33 p=0. F=2.003. similar values to the Scottish study.76±1. in this association of UACR >30 mg/g and eGFR <60 mL/min/1. OR 1. OR 3. those DM duration-dependent variables.36 to 13.54 p<0.001 5. F=0. in patients with T1DM compared with 7.54 p<0. We think that most similar with our study is the Scottish T1DM.07 to 2.24 p=0. causes a higher incidence of DR in patients with lation.05±19. OR1.63±8.54 p=0.19% compared with Triglycerides No DR 96.044. OR 2.13 DR 21.01 p<0.4 DR 38.93 DR 62. OR 2.73 m2 study. Downloaded from http://bjo. OR Clinical science Table 3 Statistical analysis at the end of 9-year follow-up study.2±8.004.77±0. Therefore.37±2.Published by group. a significant variable at p=0.08±41.001. urine albumin to creatinine ratio. but not significant in (21.78 to 4.001. this percentage is higher than 0.89) HDL No DR 73.017 in the survival analysis.7%) than in those with T2DM (13. F=13.29% in patients mine incidences of DR in T1DM and T1DM in the same popu.16±2.62 to 1. 4 Romero-Aroca P.41 p=0.012. Creatinine had a reports a higher cumulative incidence in patients with T1DM significant value in the univariate analysis. bad metabolic control.301.28 (95% CI 0.23% p<0.82 (95% CI 1.15% in patients with and an HR of 4. which was higher duration of DM (13.88±27.14 (95% CI 0. type 1 diabetes mellitus. OR 2. UACR.671.67% compared with 2.84 to 1. UACR remains without DR at baseline.001 and an HR of 3. cholesterol or triglycerides) were significant in the survival ana. measured by HbA1c (8. OR1. there were more patients with This study should be judged in the context of previous authors’ STDR probably due to retinal ischaemia in T1DM.52 p=0. National Diabetic Retinopathy Screening Programme11 that lysis.55 DR 100. DR. but The incidence of STDR was higher in patients with T1DM at eGFR was a more significant variable than UACR at p<0. studies. OR 1. F=1.86) Arterial hypertension No DR 9.94 (95% CI 1.23% p=0.02.06±9. F=1. F=2.15 p=0.06±1.16% our results with other studies.22 (95% CI 0.85) HbA1c No DR 7.1 years in patients with T2DM). F.38±39.11 STDR might be current age.63 p<0.60 to 2.78 to 4.45% in DISCUSSION patients with T2DM.53) UACR No DR on May 17. based on the 366 patients with T1DM studied Mean values Two-tailed Student’s t-test/ANOVA Univariate study Age No DR 34. eGFR. F=2.11±15.7 10 The difference between patients with T1DM and Higher STDR values in T1DM are probably due to a longer T2DM according to the incidence of any-DR.57 to 5.13±11.35 (95% CI 1. T1DM.059.85 p=0.28 p=0.6 DR 28.82 to 8.18 p=0.

638) (0.367 (0.008 HR 2.567 p<0.402 p<0. with a mean of 2. which can demonstrate the valid- values being linked to DR.011 HR 0.730 p<0. In this study.984 p<0.10 A value of 47. is more significant than creatinine values.489) (1.001 HR 1.656 HR 0.999 (0.36%).22±0.837 p=0. review time lapse after onset of T1DM.38±1.337 HR 0. which reported that these differences.73 m2. urine albumin to creatinine ratio.456 p<0. higher total cholesterol was linked to DMO.001 HR 3. Perhaps.987 (0.222) (0.354 p=0.002) (0.992) (0.665) Arterial hypertension p=0.972 to 0. such as Martín-Merino et al.255 to 2.506) (1.156 to 8.924 p=0.295 to 1.431) (1. which reported values of 97%.658 to 1.813 to 6. higher HbA1c levels in this study might have caused versy.689 to 0.001 HR 0.49%) in patients renal function.20 published in 2006 and The lipid study shows that LDL cholesterol is a risk factor in based on a population in Barbados.001 HR 0.184) (2.001 HR 4.7±1.009 p=0.002) (0. UACR Higher any-DR values in T1DM perhaps can be explained by seems less significant than the eGFR in Cox’s survival regres- two different causes: (i) a longer duration of DM (13. we revised all 366 patients to determine were reported by the fenofibrate study.42% than this study (8.6%. Furthermore.001 HR 1. From our data.361 p=0.305 p=0.349 HR 1.992) (0.450) (1. we found that 47. is that incidence at 9 years is also higher than other published studies.1136/bjophthalmol-2016-310063 5 .002) (0.978 to 1.000 p=0. probably due to a relaxation in metabolic control of Retinopathy.272 HR 0.738 p<0.118 HR 0.096 HR 1.038 p=0.25 hypertension or infection makes a masquerade effect in ±6.990 to 1. doi:10.218 HR 1.715 p<0.044 (2.681 to 0. The incidence of DMO shows similar percentages in both an increase in UACR.22%) in and decreases in advanced stages.739 to 6.42 years in patients with T1DM compared with 8.653 HR 0.998 p=0.689 to 0.46%) than those with T2DM (6. At the end of this study.12 which might be explained by methodology and the lower mean HbA1c reported a 92% prevalence of any-DR.305 to 1. despite final sum incidence was higher in patients CKD-EPI equation.542) (1. UACR.789) (0.997 (0.997 to 1.868) LDL cholesterol p<0.995 to 1.074 Clinical science Table 4 Multivariate analysis. However.003) (0.17 A T1DM compared with 7.993 to 1.801 to 2.201 p<0. This is important data for would encourage further studies to determine the CKD-EPI screening programmes. and similar data At the end of study.662 to 1.001 HR 2.16%).29% in patients with T2DM).017 HR 2. 2017 . the present sample of patients.002) (0.061 p=0.739 to 6.988) Age of DM diagnosis p<0.bmj.387 p<0. The differences patients.991 to 1. study.665 to 0.305 to 9.018) eGFR p<0. et al.022) UACR+eGFR p<0.996 p=0.7% of between CKD-EPI values and DMO.496 to 3.999 p=0.329 (1.71%–3.977 to 5. in which observe a decrease in the incidence of DR in this group of only 120 developed DR at 10 years (35.833 to 2.980 HR 1.144 (1.095) HDL cholesterol p=0.251 to 4.560 p=0.294 to 1.63 sion (table 3).747 HR 1.740 p=0.780 HR 1. estimated glomerular filtration rate.001 HR 1. Downloaded from http://bjo.716 to 6. it is frequent to study:10 on a sample size of 334 patients with T1DM.127 to 1.bmj.970) (0. These data contrast with our previously published of >20 years developed DR.211 (2. with bolic control with low levels of HbA1c.340) (0.772) (0.001 HR 0.632) (1.999 p=0.125) (0.001 HR 0. lower than previous levels in the previous study of 7.19 based on a UK population.32%). Lipid studies often create contro.001 HR 0.483 p=0. cohort study by Man et al18 reported a significant relationship The incidence of any-DR according to DM was 14.065 HR 1.0:1–6. diabetes mellitus.973 to 0.011 HR 5.991) (0.001 HR 1.13 A possible patients with T1DM in recent years.782) (0.38 studies such as the Wisconsin Epidemiological Study of Diabetic ±1.073 HR 1.995 to 1. it is evident that the eGFR with T1D (8.787) (0. The eGFR increases in early-stage DM types of DM.937 (1.001 HR 1. shown in table 2.003) (0. which generally include a revision at equation in patients with T1DM as an important DR risk 5 years if no DR is present in T1DM.987 to 3. but we observed that no one of patients registered as Kidney function can be evaluated by UACR or eGFR.16 Changes in eGFR occur prior to ity of our screening programme.877) UACR+eGFR=if UACR was >30 mg/g and eGFR <60 mL/min/1.000 p=0.050) (2.899) HbA1c p<0.762 (0.975 to 0.993) (0.868) (1. DM.410) (0.189) UACR p=0.001 HR 4.844) (0. using Cox’s proportional regression analysis Significance HR Significance HR Significance HR Significance HR Significance HR Variable (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) Current age p<0.19% (2%–2. and Leske et al.770) (0.697 HR 0. a 23.981 p<0.26% patients devel- Also in this study. thus the Wisconsin Diabetes Registry Study.421) (0.517 to 1. eGFR was determined using the with T2DM.974 to 0.411 to 9.001 HR 0.724 HR 0. both normal fundus developed DR. we must remember Romero-Aroca P.983 p<0.573 (0.166 HR 0.144 to 4.Published by group.005 HR 6. Br J Ophthalmol 2017.330 to 1.899 to 3.697 HR 1.907 to 3.668 to 1.1 years in patients with T2DM) and (ii) bad metabolic UACR.389 p<0.996 p=0.001 HR 0.08% patients with a DM duration oped DR.997 to 1.055 HR 1.982 p=0.011) Creatinine p=0. we patients with <5 years’ duration.151 HR 0.001 HR 0.9% at 9 years.872) (0. perhaps we must change marker. In recent studies.16% in patients with eGFR is recommended by various medical societies. such as the Yau et al14 meta-analysis.15 which reported slow if any developed DR and was not reported previously during progression and development of DR with the use of fenofibrates.651 to 1.004) Triglycerides p=0.285 HR on May 17. eGFR.485) (1. only 81.001 HR 0.001) (0.287 HR 0. Recently.002) (0.684 to 0.999 p=0.736 p<0. Perhaps microalbuminuria secondary to arterial ±8.342 (1. patients with DM duration of over 15 years have a better meta. reflecting the decline in patients with T1DM and 2.997 to 1.001 HR 4.989 to 1.38% (1.871 HR 1.001 HR 3.001 HR 4. with an incidence of 39.819 p<0. Determination of CKD-EPI equation as a reference for control measured by HbA1c (8.26% sum explanation for these differences.295 to 1.38±1.804) Sex p=0.68±0.996 to 1.432 p<0.

et al.39(Suppl 1):S72–80.0/ Including patients with T1DM in a T2DM screening pro- gramme is feasible but it is important to remember that more frequent screening is difficult to achieve. and twenty-year procedures. et Clinical science that a study of the peripheral retina can detect more lesions and properly cited and the use is non-commercial. Vision Loss Expert Group of the Global after the onset of T1DM. Molecular mechanisms and future perspectives. there are few studies on the incidence of DR in T1DM. previous study (35. Sagarra R.100:1366–71. (9) Microvascular complications and foot care. Behar-Cohen F. statistical analysis. Fernandez-Ballart J. et al. 10 Romero-Aroca P. Contributors PRA: contributed to study conception and design. Rivero-Ferrer E.120:2587–95. A new equation to estimate glomerular filtration rate. Study Group.150:604–12. but the increase or decrease of only one patient can Diabetes Care 2015. the CKD-EPI on May 17. Cavallerano JD.2:17. build upon this work non-commercially. and severity. we might expect to treat a lot of complications in the Wisconsin Epidemiologic Study of Diabetic Retinopathy. et al. 13-01-31/proj6]. contributing the Barbados Eye Studies. general practitioners retinopathy.. Cost-effectiveness of interventions to prevent and population. Invest Ophthalmol Vis of the version sent for publication. Barker Diabetes Care 2012. Zhang P. It is important that future studies investigate type 1 diabetes. were only six referenced studies of DR incidence in patients 9 ETDRS. Brussels. Sun JK. CKD-EPI (Chronic Kidney Disease reviewed the translation. contributed to Ann Intern Med 2009. 23 American Diabetes Association.124:250–5. Detection of diabetic macular oedema study no 5. permits others to distribute. Health Creative Commons Attribution Non Commercial (CC BY-NC 4. Changes observed in Strengthens of our study are (i) the screening programme. https://www. as a marker of eGFR for DR develop. interpreted research data and contributed to the interpretation of 19 Martín-Merino E. Diabetes Res Clin Pract 2011. NS: contributed to ophthalmological data collection. Wu SY. 21 Silva PS. AV-M: contributed to study design and 2012. carried out the laboratory diabetic retinopathy. Sabanayagam C. reviewed the statistical analysis. Wong TY. carried out the retinographies. Hennis A. Fortuny J.77 visits 1 International Diabetes Federation. DR over the 9-year follow-up period was 173 (47.bmj. Prevalence and risk of diabetic retinopathy when age at diagnosis is and endocrinologists in our area who have helped us to implement the new 30 or more years.38:437–44.35:556–64. Kohner EM. Wang JJ. Peripheral lesions identified by mydriatic Funding This study was funded by research projects FI12/01535 June 2013 and ultrawide field imaging: distribution and potential impact on diabetic retinopathy FI15/01150 July 2015 (Instituto de Investigaciones Carlos III (IISCIII) of Spain). Diabetes Res developed STDR. Br J Ophthalmol 2014. Diabetes Care 2000.56:4810–16. 2. 11 Looker HC. interpreted the 18 Man RE. J Diabetes Complicat approval of the version sent for publication. Belgium. Risk factors in a sample of people with profile and GFR. Rates of referable eye disease in the Scottish National Diabetic Retinopathy Screening Programme. wrote the discussion and edited the retinopathy. Relationship between data collection.36:631–7. which patients with T1DM of our area are being included. Incidence of retinal complications the study findings. et al.9:e100283. contributing to the final approval of the version sent for Epidemiology Collaboration). which Technol Assess 2015. diagnosed diabetic macular oedema. adapt. RS-A: contributed to study conception and design.28%. et al. III.0:1–6. AM-R: contributed to study design and the Sci 2015. 25 Lee R. et al. the statistical analysis. and our camera 14 Yau JW. If our results are confirmed by other studies in different type 1 diabetes: comparing findings from the Wisconsin Diabetes Registry Study and populations. Klein BE.88:184–8.24 Burden of Disease Study.nice. The number of patients. Microvascular complications and foot care. RN-G: contributed to study conception and design. Epidemiology of diabetic retinopathy. Nyangoma SO. de la Riva-Fernandez S. doi:10. NICE guideline (NG18) Published date: Last updated: November Competing interests None declared.102:527–32. implies that a patient visits only every 2013.19:1–116.26:506–12. 6 Romero P. Valls-Mateu A. Eye Vis (Lond) 2015. The increase in any-DR (47. and especially DMO.22 23 Patients and clini. 24 Scanlon PH. IDF diabetes Atlas. externally peer reviewed. provided the original work is macular edema and related vision loss. et al. Downloaded from http://bjo. Diabetes Care 2016. Cromie DT. Br J Ophthalmol 2017. Diabetes Care DR in the future. wrote the discussion and made a critical review. http://www. Diabetes Care 2010. microalbuminuria and overt nephropathy.94:126–32. Baget-Bernaldiz M. Prevalence and causes of vision loss in high-income countries and in Eastern and Central Europe: 1990–2010. REFERENCES Current T2DM screening. Fenofibrate: a new treatment for diabetic data. Meta-Analysis for Eye Disease (META-EYE) technicians for their work and interest in the diabetes screening. publication. made a critical review and 17 Levey AS. Ten-year incidence population. contributing to the final incidence follow-up of a sample of type 1 diabetic patients. contributed to ophthalmological 16 Romero-Aroca P. Provenance and peer review Not commissioned. et al. to the final approval of the version sent for publication. Klein R.0) license. Palta M. Scottish Diabetes Research Network ment. Rogers SL.20:3258– 2 The Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Jonas JB. Development of a cost-effectiveness model Open Access This is an Open Access article distributed in accordance with the for optimisation of the screening interval in diabetic retinopathy screening. 2016. interpreted the research data. Ferrer J. 6 Romero-Aroca P. Kawasaki R. Mellitus. Leal J. 6th edn. et al. 7 Romero-Aroca P. Nine-year incidence of diabetic retinopathy in retinographies and OCT procedures and interpreted the research data.Published by group. remix. Arch Ophthalmol 2006.38:S58–66. carried out 20 Leske MC. et al. Ophthalmology 2013.1136/bjophthalmol-2016-310063 . over a 9-year period. The number of patients with T1DM who developed control diabetes mellitus: a systematic review..9:746–51. contributing to the final approval of the version sent for in a cohort of newly diagnosed diabetic patients. is small and can bias the Clin Pract 2010.idf. diabetic and license their derivative works on different terms.98:790–5.26%) compared with our Epidemiology Group. Reyes-Torres J. Scottish Diabetic Retinopathy Collaborative. and (iii) the large amount of data. (ii) also. The incorporation of family physicians in the change the results in a 0. contributing to the final approval filtration rate with diabetic retinopathy and macular edema. Sasongko MB. 2017 .26% of the 5 American Diabetes Association. contributing to the final approval of the version sent for publication.21 licenses/by-nc/4.11±0. Stevens LA.9%) is another important consideration. who assessment of diabetic retinopathy by non-mydriatic fundus camera. sample). 13 Klein R. Curr Med Chem manuscript. The association of estimated glomerular research data and helped to write the manuscript. et al. Ophthalmology with T1DM. Methodology for retinal photography and present. cians should aim to make yearly retinography checks 5 years 3 Bourne RR.98:629–38. statistical analysis. et al. Baget-Bernaldiz M. with T1DM and 15 030 patients with T2DM in our screened 4 Li R. assessment of diabetic retinopathy: the EURODIAB IDDM complications study. Arch Ophthalmol 1984. et al. because it would seem to be linked to bad metabolic control of 12 LeCaire TJ. in diabetic retinopathy: eight-year follow up of a Spanish can change the severity of retinopathy. 22 Diabetes (type 1 and type 2) in children and young people: diagnosis and Ethics approval Hospital Universitario Sant Joan de Reus Ethics Committee type-1-diabetes [approval no. et al. Aldington SJ. et al. Roy S. The Wisconsin epidemiologic study of diabetic Acknowledgements The authors thank all paediatricians.23(Suppl 1):S4–19. Br J Ophthalmol A limitation of our study is the small sample of 366 patients 2014. the long follow-up period of our T1DM 1989. at 8 Aldington SJ. 2013.33:1872–94. Assessing progress in retinopathy outcomes in T1DM. Global prevalence and major risk factors of diabetic retinopathy. PLoS ONE 2014. despite the recommendation for patients with Report of the Expert Committee on the Diagnosis and Classification of Diabetes T1DM being annual from 5 years on. Br J Ophthalmol 2016. diabetes mellitus data collection. Moss SE. Lee et al25 carried out a literature review in 2015 but there Diabetologia 1995. et al. collected research 15 Simó R.bmj. Flaxman SR.. See: http://creativecommons. et al.18 years. screening system using the non-mydriatic fundus camera. Meuer S. publication. Schmid CH. with a mean of 4. such as lipid of diabetic retinopathy and macular edema. et al. 2013. FEDER funds.

Published by on May 17. build upon this work non-commercially. Aida Valls-Mateu.bmj.bmj. 3#BIBL Open Access This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4. Downloaded from http://bjo. Topic Articles on similar topics can be found in the following collections Collections Open access (250) Notes To request permissions go to: http://group. remix. Antonio Moreno-Ribas and Nuria Soler Br J Ophthalmol published online March 7.bmj. Sign up in the service box at the top right corner of the online To subscribe to BMJ go to: 3 These include: References This article cites 23 articles.0) license. 2017 Updated information and services can be found at: See: Differences in incidence of diabetic retinopathy between type 1 and 2 diabetes mellitus: a nine-year follow-up study Pedro Romero-Aroca. Ramon To order reprints go to: http://journals.bmj. Raul Navarro-Gil.bmj.0/ Email alerting Receive free email alerts when new articles cite this . and license their derivative works on different terms. 2017 . which permits others to distribute. provided the original work is properly cited and the use is non-commercial. 8 of which you can access for free at: http://bjo.