Epidemiological Issues in Diabetic Retinopathy

Abstract

There is currently an epidemic of diabetes in the world, principally type 2 diabetes that is linked to changing lifestyle, obesity, and increasing age of the population. Latest estimates from the International Diabetes Federation (IDF) forecasts a rise from 366 million people worldwide to 552 million by 2030. Type 1 diabetes is more common in the Northern hemisphere with the highest rates in Finland and there is evidence of a rise in some central European countries, particularly in the younger children under 5 years of age. Modifiable risk factors for progression of diabetic retinopathy (DR) are blood glucose, blood pressure, serum lipids, and smoking. Nonmodifiable risk factors are duration, age, genetic predisposition, and ethnicity. Other risk factors are pregnancy, microaneurysm count in an eye, microaneurysm formation rate, and the presence of any DR in the second eye. DR, macular edema (ME), and proliferative DR (PDR) develop with increased duration of diabetes and the rates are dependent on the above risk factors. In one study of type 1 diabetes, the median individual risk for the development of early retinal changes was 9.1 years of diabetes duration. Another study reported the 25 year incidence of proliferative retinopathy among population-based cohort of type 1 patients with diabetes was 42.9%. In recent years, people with diabetes have lower rates of progression than historically to PDR and severe visual loss, which may reflect better control of glucose, blood pressure, and serum lipids, and earlier diagnosis.

INTRODUCTION

In this issue on epidemiological issues relating to diabetic retinopathy (DR), we include the following sections:

1. Incidence and prevalence of diabetes

2. Prevalence of DR and sight threatening DR (STDR)

3. Risk factors for the development of DR and STDR

4. Progression of DR.

COMPARISON OF STUDIES IN THIS REVIEW

In view of the wide availability, improved quality and simplicity, and relative reducing cost of retinal fundus imaging systems, there is now no realistic alternative to retinal camera-based digital imaging systems to record the retinal appearance. Whilst direct ophthalmoscopy or slit-lamp biomicroscopy remain the methods of choice for some, both approaches have significant disadvantages. Primary amongst these is the lack of a permanent record of the retinal appearance: Rendering impossible any form of multiple opinion or grading for quality assurance purposes.

It is difficult to compare many of the studies in this review that record the incidence and prevalence of DR, STDR or vision threatening DR (VTDR) because of the difference in examination techniques and the different definitions, particularly of STDR and VTDR. However, it is generally true to say that assessment methods based on a clinical examination will generally underestimate mild levels of DR as compared to photographic techniques, and that with more fields photographed, a higher percentage of DR will be recorded.[1]

A range of digital imaging protocols are used throughout the world, ranging from simple single-field per eye 45° systems through to complex seven-fields per eye 30° protocols. Recent technological advantages have also resulted in the availability of ultra-wide field (200°) systems using scanning laser ophthalmoscopy techniques for imaging. In general, camera systems which use a nominal 45° horizontal field are currently the method of choice in established systematic DR screening programs, whilst systems using or supporting narrower (20° or 30°) or wider (60°) fields are more commonly found in specialist hospital eye referral centers.

If a multi-field imaging protocol is used, it is usually necessary to induce pharmacological mydriasis, as the application of the first white light electronic flash invariably causes bilateral pupillary constriction. With single-field imaging it is possible to use a non-mydriatic approach in a darkened room in some groups of patients (particularly the young who tend to naturally dilate reasonably easily), but not all patients will naturally dilate sufficiently to allow adequate imaging.

Similarly, a wide range of image grading (assessment) methods and protocols are used, making direct comparison of results very difficult. The majority of reports on retinopathy prevalence and incidence fail to provide sufficient clarity on exactly how the grading was carried out, under what conditions and frequently by whom.

INCIDENCE AND PREVALENCE OF DIABETES

Amos et al.,[2] estimated in 1997 that 124 million people worldwide had diabetes, with the majority (97%) type 2 and that by 2010 the total number with diabetes was projected to reach 221 million with the greatest potential increases in Asia and Africa, where diabetes rates could rise two or three times. In 2000, Sorensen[3] reported that the World Health Organization (WHO) has recognized that there was a “global epidemic of obesity” and the prevalence of type 2 diabetes was rising in parallel. The global and societal implications of the diabetes epidemic were described by Zimmet et al.,[4] in 2001 with changes in human behavior and lifestyle; such as sedentary lifestyle, rich nutrition, and obesity; resulting in a dramatic increase in incidence of type 2 diabetes worldwide. In 2004, Wild et al.,[5] emphasized the importance of increasing age of the population, suggesting that the most important demographic change to diabetes prevalence across the world appeared to be the increase in the proportion of people <65 years of age.

The International Diabetes Federation (IDF) published their Diabetes Atlas[6] in 2006 and demonstrated that diabetes affects 246 million people worldwide and the latest version[7] in 2011 estimates the number with diabetes to be 366 million people worldwide with a predicted rise to 552 million by 2030. It also estimates that 80% of people with diabetes live in low and middle income countries and the greatest number of people with diabetes is type 2 between 40 and 59 years of age. A trend that has also been concerning is the appearance of type 2 diabetes emerging in obese children.[8] Mohan and Pradeepa[9] reported that, in 1972, the prevalence of diabetes in urban areas of India was 2.1% and this has rapidly climbed to 12-16% representing a 600-800% increase in prevalence rates over a 30 year period.

With respect to type 1 diabetes, the 2011 IDF Diabetes Atlas[7] estimates that 78,000 children develop type 1 diabetes every year. In 2000, Karvonen et al.,[10] reported a global variation in the incidence in different populations with the overall age-adjusted incidence of type 1 diabetes varied from 0.1/100,000 per year in China and Venezuela to 36.8/100,000 per year in Sardinia and 36.5/100,000 per year in Finland. This represents a <350-fold variation in the incidence among the 100 populations. The EURODIAB ACE Study Group[11] reported that the annual rate of incidence in Europe of type 1 diabetes was increasing by 3.4% (95% CI 2.5-4.4%) and in some central European countries it was more rapid than this and they also reported a higher rate of increase in children aged under 5 years. The reported rates[12] in sub-Saharan Africa are generally low, although the accuracy of the figures makes it difficult to give an exact figure, more accurate figures being available for Pakistan of 1/100,000 per year and in Egypt 8/100,000 per year. The variation in incidence between different populations and ethnic groups can be explained by genetic[13] and environmental factors. Feltbower et al.,[14] reported an increased incidence of type 1 diabetes in south Asians in Bradford, UK. Variations in incidence of type 1 diabetes in different seasons have been reported[15,16] and some perinatal risk factors have been identified[17] as increasing the risk of type 1 diabetes such as older maternal age and blood group incompatibility.

PREVALENCE OF DIABETIC RETINOPATHY (DR) AND SIGHT THREATENING OR VISION THREATENING DIABETIC RETINOPATHY (STDR OR VTDR)

In 1984, Klein et al.,[18,19,20,21] reported results from the Wisconsin Epidemiological Study of Diabetic Retinopathy (WESDR study), a population-based study in southern Wisconsin to determine the prevalence and severity of DR and associated risk variables in the following groups of patients:

Group 1: 996 patients with insulin dependent diabetes mellitus (IDDM) diagnosed > 30 years

Group 2: 674 patients with IDDM diagnosed < 30 years

Group 3: 696 patients with non (N) IDDM < 30 years.

The groups showed the following prevalence:

Group 1: DR was detected in 70%, proliferative retinopathy (after 20 years of diabetes) was present in about 50%, and macular edema (ME) (after 15 years of diabetes) was present in about 18%. The prevalence of DR ranged from 17-97.5% in persons with diabetes for less than 5 years and those with diabetes for 15 or more years. Proliferative retinopathy ranged from 1.2-67% in persons with diabetes for less than 10 and 35 or more years, respectively. Prevalence rates of ME varied from 0% in those who had diabetes less than 5 years to 29% in those whose duration of diabetes was 20 or more years

Group 2: DR was detected in 62%, proliferative retinopathy (after 20 years of diabetes) was present in about 25%, and ME (after 15 years of diabetes) was present in about 20%

Group 3: DR was detected in 36%, proliferative retinopathy (after 20 years of diabetes) was present in about 5%, and ME (after 15 years of diabetes) was present in about 12%.

The mean duration of known diabetes was 14.6 years in Group 1, 11.0 years in Group 2, and 6.9 years in Group 3.

Kohner et al.,[22] reported baseline retinopathy levels in 2,964 patients with newly diagnosed type 2 diabetes enrolled in the United Kingdom Prospective Diabetes Study (UKPDS). Retinopathy, defined as microaneurysms or worse lesions in at least one eye, was present in 39% of men and 35% of women. Cotton wool spots or intraretinal microvascular abnormalities were present in 8% of men and 4% of women.

In 2002, Younis et al.,[23] reported baseline results from population screening in Liverpool of 831 people with type 1 diabetes and 7,231 people with type 2 diabetes. The results showed a baseline for type 1 of any DR 45.7%, proliferative DR (PDR) 3.7% and STDR 16.4%. Baseline for type 2 group of any DR 25.3%, PDR 0.5%, and STDR 6.0%.

Kernell et al.,[24] reported the youngest child in the literature (11.8 years) with pre-PDR. Donaghue et al.,[25] reported the youngest child reported in the literature with background DR-7.9 years (duration 5.6 years, HbA1c 8.9%).

Three studies[26,27,28] have demonstrated that, if one screens for type 2 diabetes in different populations, the prevalence of DR in screen positive patients (7.6, 6.8, and 9%) is much lower than the prevalence in the know population of people with diabetes.

Yau et al.,[29] provides worldwide estimates using data from 22,896 individuals with diabetes in 35 studies that used photographic methods of detection. The overall prevalence was 34.6% (95% confidence interval (CI) 34.5-34.8) for any DR, 6.96% (6.87-7.04) for PDR, 6.81% (6.74-6.89) for diabetic ME, and 10.2% (10.1-10.3) for VTDR. The paper mentions limitations of different time points, methodologies, and population characteristics.

Al Ghamdi et al.,[30] studied 3,300 eligible people over 50 years of age in Taif in Saudi Arabia, examining 3,052 (93%). The prevalence of diabetes was 29.7%. Most people (88%) confirmed to have diabetes in these examinations had a previous diagnosis of diabetes. More than half (56%) the people with known diabetes had poor control (random blood glucose < 200 mg/dl). The prevalence of any DR was 34.6%, but what was noticeable was the high level of STDR of 17.5%, which was mostly due to high levels of referable maculopathy (15.9%) and may be related to the high number with poor glycemic control.

Burgess et al.,[31] reported a systematic literature review of studies of DR and maculopathy in Africa. Sixty-two studies from 21 countries were included. In population based studies the reported prevalence range in patients with diabetes was 30-31.6% for any DR, 0.9-1.3% for PDR, and 1.2-4.5% for any maculopathy. The review also found the reported prevalence in diabetes clinic based surveys to be 7.0-62.4% for any DR, 0-6.9% for PDR, and 1.2-31.1% for any maculopathy. A study[32] from the same group of researchers in a diabetes center in Malawi found that in 249 patients with type 2 diabetes the retinopathy prevalence were 32.5% (26.7-38.3%) for any DR, 4.8% (2.2-7.5%) for PDR, and 19.7% (14.7-24.6%) for STDR. In 32 patients with type 1 diabetes, the retinopathy prevalence was 28.1% (12.5-43.7%) for any DR, 12.5% (1.0-24.0%) for PDR, and 18.8% (5.2-32.2%) for STDR.

RISK FACTORS FOR THE PROGRESSION OF DIABETIC RETINOPATHY (DR) TO SIGHT THREATENING OR VISION THREATENING DR (STDR OR VTDR)

Risk factors for the development of DR are generally divided into modifiable and nonmodifiable.

Modifiable risk factors are

Blood glucose

In type 1 diabetes, the Diabetes Control and Complications Trial[33,34,35] (DCCT) included 1,441 people with IDDM, 726 with no DR at baseline (the primary-prevention cohort), and 715 with mild retinopathy (the secondary-intervention cohort); with mean follow-up of 6.5 years. For the primary-prevention cohort, intensive therapy reduced the mean risk for the development of DR by 76% (CI 62-85%), compared with conventional therapy. For the secondary-intervention cohort, intensive therapy slowed the progression of DR by 54% (CI 39-66%) and reduced the development of PDR or severe NPDR by 47% (CI 14-67%).

In type 2 diabetes, the UKPDS[22,36,37,38,39] recruited 5,102, and of these, 3,867 with NIDDM and the effect of intensive blood glucose control with sulfonylureas or insulin was compared with conventional treatment. Compared with the conventional group, there was a 25% risk reduction (7-40, P = 0.0099) in the intensive group in microvascular endpoints, including the need for retinal photocoagulation.

Numerous other studies have found this link with blood glucose control[40,41,42,43,44,45,46] both in type 1 and type 2 diabetes.

Blood pressure

Control of systemic hypertension has been shown to reduce the risk of new onset DR and slow the progression of existing DR.[38,39,45,47,48,49,50,51,52] However, a study[53] reported in 2008 from a tertiary eye clinic in Melbourne that adherence to clinical guidelines for glycemic and blood pressure (BP) control in patients with DR was low, even in a well-developed healthcare system with free ophthalmic care access.

Serum lipids

There is evidence that elevated serum lipids are associated with macular exudates and moderate visual loss and partial regression of hard exudates may be possible by reducing elevated lipid levels.[46,54,55,56,57,58] Funatsu et al.,[59] has suggested that serum lipoprotein A level is an independent risk factor for the progression of NPDR in type 2 diabetes patients. The subgroup analyses from the FIELD study[60] and the ACCORD[61] study have led to further recommendations[62] of control of elevated lipid levels in the modern medical management of DR.

Smoking

There is some evidence that smoking may be a risk factor in progression of DR in type 1 diabetes as described by Muhlhauser and Muhlhauser et al.,[63,64] and Karamanos et al.[65] However, in type 2 disease the evidence is controversial and it may protect[39] against the progression of retinopathy in some patients despite the fact that it is an independent risk factor for myocardial infarction and death from cardiovascular disease in patients with diabetes.

Nonmodifiable risk factors are

Duration

A major nonmodifiable determinant of development and progression of DR is duration[66,67,68] of diabetes.

Age

The Wisconsin Epidemiological Study[18,19] found, in those whose age of diagnosis was less than 30 years and who had diabetes of 10 years duration or less, that the severity of retinopathy was related to older age at examination; whereas when the age at diagnosis was 30 or more years, the severity of retinopathy was related to younger age at diagnosis. In the UKPDS[39] , where all recruited newly diagnosed patients with type 2 diabetes were aged 65 or less, in those who already had retinopathy, progression was associated with older age.

Genetic predisposition

Early studies of identical twins with diabetes mellitus suggest familial clustering of DR and Hietala et al.,[69] has demonstrated a familial clustering of proliferative retinopathy in patients with type 1 diabetes in Finland. Association with human leukocyte antigens (HLAs) and candidate genes have so far only shown weak associations.[70,71,72]

Ethnicity

Emanuele et al.,[73] reported ethnic differences in higher levels of DR between Hispanics (36%), African Americans (29%), and non-Hispanic whites (22%); and Simmons[74] demonstrated that that moderate or more severe retinopathy is more common in Polynesians than Europeans. The prevalence of moderate or more severe retinopathy was 4.0% in Europeans, 12.9% in Maori, and 15.8% in Pacific people (P = 0.003). In neither of these two studies could the differences be accounted for by an imbalance in traditional risk factors such as age, duration of diagnosed diabetes, HbA1c, and blood pressure.

Raymond et al.,[75] found that patients in the UK of south Asian ethnicity compared to white European ethnicity had significantly higher prevalence of DR and of maculopathy. However, they also had significantly elevated systolic and diastolic blood pressures, HbA1c and total cholesterol, and lower attained age and younger age at diagnosis; which made it difficult to separate the ethnic differences from the other known risk factors.

Risk factors difficult to place in the modifiable or nonmodifiable category

Pregnancy

Two studies[76,77] found pregnancy to be independently associated with progression of DR.

Biomarkers

Microaneurysm count has been shown[78] to be associated with progression of DR.

Scanlon et al.,[79] reported the risk of progression is significantly higher for those with background DR in both eyes than those with background retinopathy in only one or in neither eye.

Nunes et al.,[80] described a high microaneurysm formation rate on color fundus photographs appears to be a good biomarker for DR progression to clinically significant ME (CSME) in type 2 diabetic patients with NPDR.

INCIDENCE AND PROGRESSION OF DIABETIC RETINOPATHY

In 1981, Palmberg et al.,[81] described a study of the natural history of DR in 461 people with juvenile onset IDDM. At IDDM diagnosis no DR was found, at 7 years duration DR was present in 50% and after 17-50 years in 90%. PDR was first seen after around 13 years duration and in 26% of the patients at 26-50 years.

In 1984 and 1989, Klein et al.,[18,19,66,67] reported on the natural history of DR in the Wisconsin Epidemiological Study (WESDR study). These studies demonstrated a clear progression in DR to PDR and/or ME with duration of the condition for the following groups of patients:

Group 1a: Patients with IDDM diagnosed >30 years with no DR at first visit

Group 1b: IDDM diagnosed >30 yrs. with DR, but no PDR or ME at first visit

Group 2a: Those with IDDM diagnosed ≥30 years with no DR at first visit

Group 2b: Those with IDDM diagnosed ≥30 years with DR, but no PDR or ME at first visit

Group 3a: Those with NIDDM ≥30 years with no DR at first visit

Group 3b: Those with NIDDM diagnosed ≥30 years with DR, but no PDR or ME at first visit.

The groups showed the following progression rates:

Group 1a: 59% developed DR after 4 yrs

Group 1b: After 4 years, worsening of DR occurred in 41%, improvement in 7%

Group 2a: 47% developed DR after 4 years

Group 2b: Worsening of DR in 34%

Group 3a: 34% developed DR after 4 years

Group 3b: Worsening of DR in 25%.

For progression to PDR and/or ME the rates were:

Group 1b: 11% developed PDR after 4 years, 14% after 13 years of diabetes, and 8.2% ME after 4 years

Group 2b: 7% developed PDR after 4 years, 8.4% ME after 4 years

Group 3b: 2% developed PDR after 4 years, 2.9% ME after 4 years.

Burger et al.,[82] reported results from the Berlin DR study where 231 subjects with Type 1 diabetes mellitus aged 18 ± 4 years were examined between one and six times both by ophthalmoscopy and fluorescein angiography. In subjects less than 15 years of age and diabetes for less than 5 years, retinal lesions were rare. The median time to development of early retinopathy was 9.1 years from diagnosis of diabetes.

Grauslund et al.,[83] reported the 25 year incidence of proliferative retinopathy among a population-based cohort of 727 Danish type 1 diabetic patients was 42.9%.

A systematic review published by Williams et al.,[84] in 2004 on the epidemiology of DR and ME concluded that studies of sufficient size to stratify for age and duration of eye disease show an increase in DR in older age groups with long-standing disease.

In 2006, an article by Leske et al.,[85] on long-term incidence among persons of African origin suggested a possible lower risk of severe/PDR than in whites, while CSME incidence seems comparable or higher.

In 2008 and 2009, Klein et al.,[48,49] reported on the 25-year cumulative progression and regression of DR and cumulative incidence of ME and CSME in type 1 patients in the Wisconsin Epidemiologic Study of Diabetic Retinopathy. The 25-year cumulative rate of progression of DR was 83%, progression to PDR was 42%, and improvement of DR was 18% and the 25-year cumulative incidence was 29% for ME and 17% for CSME.

These studies reported a reduction in incidence and prevalence of PDR and ME in the more recently diagnosed cohorts which was considered to be due to better glycemic control and BP control in the later cohorts.

In 2009, Wong et al.,[86] conducted a systematic review of rates of progression of DR in people with both type 1 and type 2 diabetes during different time periods. The article concluded that since 1985, diabetic patients have lower rates of progression to PDR and severe visual loss. These findings may reflect an increased awareness of retinopathy risk factors; earlier identification and initiation of care for patients with retinopathy; and improved medical management of glucose, blood pressure, and serum lipids.