Diabetes mellitus has become a major public health problem in India. Global estimates projected that India would have the largest number of people with diabetes in the world, rising to about 79 million by 2030 [1]. The National Urban Diabetes Survey found a high prevalence of diabetes in Indian cities [2], and the Indian Council of Medical Research-INdia DIABetes study estimated that 62.4 million Indians had diabetes in 2011 [3].

Asian Indians develop diabetes at a younger age and at lower body weight than Western populations [4], so they live longer with the disease and its complications.

Diabetic retinopathy is a specific microvascular complication of diabetes and a leading cause of blindness in working-age adults. Its prevalence rises with the duration of diabetes [5], and diabetic macular oedema is the main cause of visual loss in type 2 diabetes [6].

Population studies outside India have reported retinopathy

in about 15% to over 40% of people with diabetes [7-10].

In India, population-based studies in Chennai found diabetic retinopathy in 17.6% to 18% of people with diabetes [11,12], and studies in Hyderabad and rural Tamil Nadu reported 22.4% and 10.5% respectively [13,14]. Clinic-based studies from south India found higher rates, and some patients already had retinopathy at the time diabetes was diagnosed [15,16].

Most Indian data come from the southern states, and there are few hospital-based studies from western India. Many patients in India are unaware of the need for regular eye examination and present late, when vision-threatening retinopathy has developed [17].

Knowledge of the local prevalence and risk factors helps in planning screening services. The aim of this study was to estimate the prevalence of diabetic retinopathy and to identify the factors associated with it among patients with type 2 diabetes mellitus attending a tertiary care hospital in Pune, western India.

Study design and setting This hospital-based cross-sectional study was conducted in the Department of Ophthalmology, Smt. Kashibai Navale Medical College and General Hospital, Narhe, Pune, a tertiary care teaching hospital. Patients attending the ophthalmology outpatient department and those referred from the medicine department and diabetes clinic were included. Written informed consent was obtained from every participant. Participants Patients aged 30 years or more with a diagnosis of type 2 diabetes mellitus, based on a documented physician diagnosis or current treatment with oral hypoglycaemic drugs or insulin, were included consecutively. Patients with type 1 diabetes, gestational diabetes, media opacities that prevented a clear view of the fundus, other retinal vascular diseases such as retinal vein occlusion, or previous vitreoretinal surgery, and those who did not consent, were excluded. Sample size The sample size was calculated from an expected prevalence of diabetic retinopathy of 18% in Indian patients with diabetes [12]. With an absolute precision of 7% and a 95% confidence level, the formula n = Z² p (1 - p) / d² gave a minimum of 116 patients, and 120 patients were included. Data collection A structured proforma was used to record age, sex, residence, duration of diabetes, type of treatment, family history of diabetes and smoking. Height, weight and blood pressure were measured, and body mass index was calculated. Hypertension was defined as blood pressure of 140/90 mmHg or more or current use of antihypertensive drugs. Fasting and postprandial blood sugar, glycated haemoglobin (HbA1c), serum total cholesterol, triglycerides, high-density and low-density lipoprotein cholesterol, serum creatinine and the urinary albumin to creatinine ratio were measured. Albuminuria was defined as a urinary albumin to creatinine ratio of 30 mg/g or more. Patients were also asked whether they had ever had a fundus examination. Ocular examination All patients underwent recording of best corrected visual acuity on a Snellen chart, slit lamp examination of the anterior segment, and intraocular pressure measurement by applanation tonometry. After pupillary dilatation with tropicamide 0.8% and phenylephrine 5%, the fundus was examined by slit lamp biomicroscopy with a +90 D lens and by indirect ophthalmoscopy with a +20 D lens. Retinopathy was graded in the worse eye according to the International Clinical Diabetic Retinopathy severity scale as no apparent retinopathy, mild, moderate or severe non-proliferative retinopathy, or proliferative retinopathy [18]. Clinically significant macular oedema was diagnosed by the Early Treatment Diabetic Retinopathy Study criteria [19]. Sight-threatening retinopathy was defined as severe non-proliferative retinopathy, proliferative retinopathy or clinically significant macular oedema. Statistical analysis Data were entered in Microsoft Excel and analysed with Python version 3.13 using the SciPy and statsmodels libraries. Continuous variables were expressed as mean and standard deviation and categorical variables as number and percentage. Prevalence was reported with an exact 95% confidence interval. Patients with and without retinopathy were compared by the chi-square test for categorical variables and the independent t-test for continuous variables, and trends across categories were tested with the Cochran-Armitage test. Variables with p < 0.20 on univariate analysis were entered into a multivariable logistic regression model; fasting blood sugar was not entered because of its close correlation with HbA1c. Results were expressed as adjusted odds ratios (aOR) with 95% confidence intervals. A p-value below 0.05 was considered significant.

A total of 120 patients with type 2 diabetes were examined. Their mean age was 55.2 ± 13.1 years, 48.3% were men and the mean duration of diabetes was 11.3 ± 7.1 years. Of these, 56 (46.7%) were on oral hypoglycaemic drugs, 27 (22.5%) on insulin and 37 (30.8%) on both.

The mean HbA1c was 8.2 ± 1.6%, and 86.7% had hypertension. A total of 52 patients (43.3%) had never had a fundus examination (Table 1).

Table 1. Characteristics of patients with and without diabetic retinopathy

SD: standard deviation; HbA1c: glycated haemoglobin; LDL: low-density lipoprotein. Chi-square test for proportions; independent t-test for means.

Diabetic retinopathy was present in 62 patients, giving a prevalence of 51.7% (95% CI 42.4 to 60.9). Moderate non-proliferative retinopathy was the most common grade (21.7%), followed by mild non-proliferative retinopathy (14.2%), and proliferative retinopathy was seen in 8.3%. Clinically significant macular oedema was present in 12 patients (10.0%), and 30 (25.0%) had sight-threatening retinopathy (Table 2).

Table 2. Distribution of diabetic retinopathy

Graded in the worse eye by the International Clinical Diabetic Retinopathy severity scale [18]. CI: exact confidence interval.

The prevalence of retinopathy was 42.9% in patients with diabetes for less than 5 years, 24.2% at 5 to 9 years, 71.4% at 10 to 14 years and 71.1% at 15 years or more (p for trend < 0.001). It rose steadily with HbA1c, from 30.0% in patients with HbA1c below 7% to 75.7% in those with HbA1c of 9% or more (p for trend < 0.001) (Table 3).

Table 3. Prevalence of retinopathy by duration of diabetes and HbA1c

Cochran-Armitage test for trend.

On univariate analysis, retinopathy was associated with longer duration of diabetes, higher HbA1c and fasting blood sugar, insulin treatment, albuminuria and, unexpectedly, lower total cholesterol, but not with hypertension, body mass index, triglycerides or smoking (Table 1). On multivariable logistic regression, duration of diabetes (aOR 2.41 per 5 years, 95% CI 1.57 to 3.70; p < 0.001), HbA1c (aOR 2.10 per 1%, 95% CI 1.49 to 2.95; p < 0.001) and insulin treatment (aOR 3.56, 95% CI 1.33 to 9.49; p = 0.011) were independently associated with diabetic retinopathy. Albuminuria showed a positive association of borderline significance (aOR 2.33, 95% CI 0.90 to 6.02; p = 0.081), and total cholesterol, age and sex were not significant after adjustment (Table 4).

Table 4. Multivariable logistic regression of risk factors for diabetic retinopathy

Variables with p < 0.20 on univariate analysis were entered into the model. OR: odds ratio; CI: confidence interval. Model fit: Hosmer-Lemeshow p = 0.41.

In this hospital-based study from western India, diabetic retinopathy was present in 51.7% of patients with type 2 diabetes, and 25.0% had sight-threatening retinopathy. Duration of diabetes, poor glycaemic control and insulin treatment were independent risk factors, and 43.3% of the patients had never had their fundus examined. The prevalence in our study is much higher than that reported in Indian population-based studies. The Chennai Urban Rural Epidemiology Study found retinopathy in 17.6% of people with diabetes [11], the Sankara Nethralaya study in 18% [12], and a study from Hyderabad in 22.4% [13], while a rural study from Theni district found 10.5% [14]. Clinic-based studies usually report higher rates because patients with longer and more severe diabetes attend hospitals [15]. Rates in other Asian and Western populations range widely [7-10], reflecting differences in the duration of diabetes, glycaemic control and methods of examination. Our patients had a mean duration of diabetes of 11.3 years and a mean HbA1c of 8.2%, more than half were on insulin, and many were referred from the medicine department, often because of symptoms. These factors, together with careful examination by slit lamp biomicroscopy and indirect ophthalmoscopy, explain the high prevalence, which should not be taken as the prevalence in the general diabetic population. Duration of diabetes was a strong risk factor in our study, with each additional 5 years more than doubling the odds of retinopathy, as in the Wisconsin study, where the prevalence of retinopathy rose steeply with the duration of diabetes [5]. Retinopathy was found in 42.9% of our patients with diabetes of less than 5 years' known duration, which agrees with the finding that many Indian patients already have retinopathy at the time of diagnosis [16], because type 2 diabetes is often present for years before it is detected. This supports the recommendation that every patient with type 2 diabetes should have a dilated fundus examination at diagnosis and every year thereafter. Poor glycaemic control was strongly associated with retinopathy: its prevalence rose from 30.0% in patients with HbA1c below 7% to 75.7% in those with HbA1c of 9% or more, and the odds increased 2.10 times for every 1% rise in HbA1c. The Diabetes Control and Complications Trial showed that intensive control reduced the development and progression of retinopathy in type 1 diabetes [20], and the United Kingdom Prospective Diabetes Study showed that intensive blood glucose control reduced microvascular complications in type 2 diabetes [21]. Higher HbA1c was also a risk factor for the incidence and progression of retinopathy in the UKPDS [22]. Insulin treatment was an independent risk factor in our study, probably because patients who need insulin have longer-standing diabetes with greater loss of beta-cell function and poorer control than those on oral drugs. Hypertension was not associated with retinopathy in our study, probably because 86.7% of our patients were hypertensive, which left little contrast between the groups. Tight blood pressure control nevertheless reduced the progression of retinopathy and the deterioration of vision in the UKPDS [23,24], and blood pressure, lipids and obesity were associated with retinopathy in the Hoorn study [25]. Serum lipids were associated with retinopathy in urban south Indians [26], but in our study total cholesterol was lower in patients with retinopathy and had no independent effect, possibly because of more frequent statin use among patients with complications. In the ACCORD Eye study, intensive glycaemic control and the addition of fenofibrate to statin therapy slowed the progression of retinopathy [27], and fenofibrate reduced the need for laser treatment in the FIELD study [28]. Albuminuria, a marker of diabetic nephropathy, was more common in patients with retinopathy, as in the Sankara Nethralaya study [29], which reflects the shared microvascular basis of the two complications, although the association was of borderline significance after adjustment. Similar risk factors have been reported in rural Indian populations [30]. These findings have important implications. Since 43.3% of our patients had never had a fundus examination and one in four already had sight-threatening disease, eye screening should be integrated into diabetes clinics, and physicians should refer every patient with diabetes for an annual dilated fundus examination. Screening models that combine fundus photography with referral have been shown to work in both urban and rural Indian settings [17]. Good control of blood sugar, together with control of blood pressure and lipids and early detection of nephropathy, should be part of the prevention of blindness from diabetic retinopathy. Strengths and limitations The strengths of this study are the use of a standard grading scale, examination of the fundus by slit lamp biomicroscopy and indirect ophthalmoscopy, and the measurement of a wide range of systemic risk factors. It has limitations. It was a hospital-based study with a modest sample, so the prevalence is higher than in the general population and the confidence intervals are wide. The cross-sectional design does not allow conclusions about cause and effect. Seven-field stereo photography and optical coherence tomography were not used in every patient, so mild retinopathy and macular oedema may have been missed, and information on statin use was not analysed. Patients with dense media opacities were excluded.

Diabetic retinopathy was present in 51.7% of patients with type 2 diabetes attending a tertiary care hospital in Pune, and 25.0% had sight-threatening disease. Longer duration of diabetes, poor glycaemic control and insulin treatment were the main risk factors. Regular dilated fundus examination of all patients with diabetes, starting at diagnosis, together with good control of blood sugar, blood pressure and lipids, is essential to prevent blindness from diabetic retinopathy.

DECLARATIONS

Acknowledgement: The author thanks the faculty, residents and staff of the Departments of Ophthalmology and Medicine, Smt. Kashibai Navale Medical College and General Hospital, Pune, and all patients who took part in the study.

Funding: None.

Conflict of interest: The author declares no conflict of interest.

Author contribution: Dr. Deepak Pathak designed the study, examined the patients, analysed the data, wrote the manuscript and approved the final version.

Data availability: The data supporting the findings of this study are available from the corresponding author on reasonable request.