Prevalence and Severity of Diabetic Retinopathy among Diabetic Patients Presenting to a Tertiary Eye Hospital in Nepal.

PURPOSE: The purpose of the study is to evaluate the prevalence and severity of diabetic retinopathy (DR) among all diabetic patients presenting to a tertiary eye care center in Nepal over a 3-year period.
MATERIALS AND METHODS: This was a retrospective review of all clinical records from the initial presentations of diabetic patients at the Tilganga Institute of Ophthalmology (TIO) from 2012 to 2014.
RESULTS: In total, 8855 patients were identified who presented to TIO with a prior diagnosis of diabetic mellitus (DM) during the study period. DR was found in 1714 patients (19.4%) at the time of initial presentation, 1305 (14.7%) of which had nonproliferative DR (PDR), while 617 (6.9%) demonstrated diabetic macular edema (DME) and 409 (4.6%) demonstrated PDR. Of the 1714 patients with DR, 825 (48.1%) required treatment at initial presentation for DME and/or PDR. Male gender, hypertension, age at presentation, and interval from diagnosis of DM to initial eye consultation were significantly associated with the presence of DR (P < 0.05).
CONCLUSIONS: The prevalence of DR among patients with DM was 19.4%, which is lower than previously published estimates. However, among patients diagnosed with DR, over half presented with a vision-threatening complication warranting some measure of initial treatment. Copyright:

Introduction

Diabetic retinopathy (DR) is the leading cause of visual impairment among working-age individuals in the Western world, according to numerous large population-based studies.[12] Detailed understanding of the prevalence of diabetes mellitus (DM) in a general population, as well as the association of DR with other detectable conditions (e.g., hypertension) allows for improved awareness, improved patterns of screening and treatment, and subsequently lower incidence and severity of vision loss.

This question is of critical importance in South Asia and specifically in Nepal. According to the World Health Organization, India has the largest diabetic population in the world, projected to reach 79.4 million by 2030.[3] However, previous population-based screening studies from Nepal have shown a higher percentage prevalence of DR than India,[4] a fact of significant public health concern for the country.

Initial large cross-sectional surveys examining the prevalence of DR were conducted predominately in Western nations; however, studies examining other, genetically diverse populations have been conducted around the world, including a number of smaller studies in Nepal [Table 1].[567891011] These studies have largely been hospital-based, cross-sectional examinations of various subsegments of the diabetic patients presenting to an institution, carried out prospectively over time periods less than a year, with a prevalence of DR ranging widely from 20.31% to 78%.[56789] An additional Nepalese study examining diabetic patients who recruited through community screening found a lower prevalence of DR of 19.3%,[12] and a subsequent community-based study of known diabetics reported a slightly lower DR prevalence of approximately 12%.[13]

Table 1

Results of prior diabetic retinopathy studies from South Asia

Study	n	DR (percentage of total)	NPDR (percentage of total)	PDR (percentage of total)
Shrestha RK, 2011*	128	26 (20.31)	17 (13.28)	9 (7.03)
Thapa R, et al., 2012*	210	164 (78)	129 (61.4)	35 (16.7)
Shrestha S, et al., 2007*	316	67 (21)	N/A	N/A
Shrestha MK, et al., 2007*	371	166 (44.7)	142 (38.2)	24 (6.4)
Rizyal A, 2004*	120	57 (47.3)	56 (46.5)	1 (0.08)
Ramavat PR, et al., 2013§	168	57 (33.9)	43 (25.5)	14 (8.3)
Wahab S, et al., 2008‡	130	20 (15)	N/A	N/A

*Denotes studies conducted with Nepalese populations; §Hospital-based, cross-sectional study from western India, evaluating only type 2 DM patients; ‡Hospital-based, cross-sectional study from Pakistan, evaluating newly-diagnosed type 2 DM patients. DM: Diabetes mellitus, DR: Diabetic retinopathy, NPDR: Nonproliferative diabetic retinopathy, PDR: Proliferative diabetic retinopathy, N/A: Not available

This broad variability may simply reflect the diversity of subpopulations examined, the diversity of the Nepalese population presenting to hospitals in Kathmandu, or may be a result of smaller patient numbers examined. Similar studies from other countries in South Asia also present widely divergent prevalence figures, albeit from widely divergent populations and with variable study criteria.[10111415] Larger studies are needed to clarify these numbers.

Along with examinations of DR prevalence alone, significant attention should be given to the severity of DR at presentation and the need for treatment. Early studies from populations in Southeast Asia and Oceania have shown that up to 10% of diabetic patients have vision-threatening DR at initial presentation, with up to 5% requiring treatment at this initial visit.[1617] Prior studies from Nepal have shown a prevalence of non-proliferative DR (NPDR) ranging from 13.28% to 61.4% and of PDR ranging from 0.08% to 16.7%.[5689] Again, a broad range of prevalence has been seen that warrants further clarification.

Within this context, we conducted a cross-sectional survey of diabetic patients presenting for ocular examination at the Tilganga Institute of Ophthalmology (TIO) from the years 2012–2014, focusing on the prevalence of DR, DR severity, and associated factors.

Materials and Methods

The study was a retrospective, cross-sectional review of all clinical records from all patients presenting to the TIO with a prior diagnosis of DM, from the years 2012 to 2014. Due to the unique position of TIO within the community as both a tertiary eye care center and a charitable organization offering care to all presenting patients, these patients represent a broad cohort referred from both outside eye clinics and primary care, as well as self-referred patients, from around the Kathmandu Valley. As the largest population center in Nepal, a country of widely divergent ethnicities, these patients also represent the most heterogeneous cross-section of the Nepalese population. Demographic data, including age and gender, as well as health data, such as diabetic subclass (i.e., Type 1 or Type 2) and interval from diagnosis to presentation at TIO, were collected. DR was determined to be present through the presence of any retinal vascular abnormalities consistent with this entity – “dot-blot” hemorrhages, microaneurysms, cotton-wool spots, intraretinal microvascular abnormalities, and/or venous beading. In addition, data regarding its staging and severity, the presence or absence of neovascularization, and initial treatment modality (if any required) were noted. To eliminate the extensive variability in attempting to retrospectively substage PDR and NPDR (e.g., mild, moderate, severe), patients were grouped as either “no DR,” “NPDR” (retinal manifestations without neovascularization), and “PDR” (any neovascularization). In addition, the presence or absence of diabetic macular edema (DME) was noted. DR severity for an individual patient was determined according to the staging of the eye with the most severe disease manifestations.

All patient data were appropriately deidentified. This study was approved by the Institutional Review Committee at TIO and adhered to the tenets of the Declaration of Helsinki.

Inclusion/exclusion criteria

Patient records were included in the review if they (a) presented to TIO during the study period from 2012 to 2014, (b) carried a diagnosis of either Type 1 or Type 2 DM at the time of initial presentation, (c) were adult patients >18 years of age at the time of initial presentation, and (d) did not meet any of the exclusion criteria. Patients were excluded if a fundus examination had not been performed at the initial visit (often precluded due to coexisting ocular comorbidities, such as mature cataract), if there was a history of potentially confounding medical conditions (e.g., known hypertensive retinopathy, infection with human immunodeficiency virus, history of ocular radiation, hemoglobinopathy), or if the DM was not classified as Type 1 or Type 2, such as gestational diabetes or steroid-induced diabetes.

Data analysis

For categorical variables, a Chi-square test was used; for continuous variables, Student's t-test or ANOVA test was used for mean difference, where applicable. Statistical analysis was performed using IBM SPSS Statistics (version 20, IBM, Armonk, NY, USA). Statistical significance was set at P < 0.05.

Results

Altogether, 8855 patients were identified who presented with DM during the study period [Table 2]. Of these 8855 cases, 2578 presented in 2012, 2853 in 2013, and the final 3424 in 2014. The overall gender ratio of the patients was 1.2:1 (M:F), with a similar gender distribution in all years studied. Male gender was significantly associated with the presence of DR at initial presentation (P < 0.001, Chi-square test). The mean patient age at time of presentation was 56.9 years (standard deviation ± 11.9 years), and this varied little for each individual year of the study period. The study population was skewed almost entirely toward patients with type 2 DM (8586 patients, 97%), with only 3% (269 patients) reporting a diagnosis of type 1 DM. This was similar across all years of the study period.

Table 2

Demographics, diabetic retinopathy, and severity of participants at the time of presentation (n=8855)

Variables	Categories	Total, n (%)	2012, n (%)	2013, n (%)	2014, n (%)
Gender	Male	4878 (55.1)	1399 (54.3)	1555 (54.5)	1924 (56.2)
	Female	3977 (44.9)	1179 (45.7)	1298 (45.5)	1500 (43.8)
Age	Mean±SD	56.9±11.9	57.0±11.9	56.8±11.9	56.9±12.0
DM types	1	269 (3)	82 (3.2)	87 (3)	100 (2.9)
	2	8586 (97)	2496 (96.8)	2766 (97)	3324 (97.1)
DR	Yes	1714 (19.4)	502 (19.5)	620 (21.7)	592 (17.3)
	No	7141 (80.6)	2076 (80.5)	2233 (78.3)	2832 (82.7)
DR types (worse eye)	NPDR	889 (51.9)	269 (53.6)	335 (54)	285 (48.1)
	NPDR and DME	416 (24.3)	119 (23.7)	141 (22.7)	156 (26.4)
	PDR	208 (12.1)	46 (9.2)	87 (14)	75 (12.7)
	PDR and DME	201 (11.7)	68 (13.5)	57 (9.2)	76 (12.8)
Initial treatment plan	Observation/diet control	7939 (89.7)	2321 (90)	2507 (87.9)	3111 (90.9)
	Laser of any kind	44 (0.5)	3 (0.1)	40 (1.4)	1 (0)
	Intravitreal injection (s)	17 (0.2)	0 (0)	17 (0.6)	0 (0)
	Incisional surgery	9 (0.1)	0 (0)	9 (0.3)	0 (0)
	Multimodal	845 (9.5)	254 (9.9)	279 (9.8)	312 (9.1)

SD: Standard deviation, DM: Diabetes mellitus, DR: Diabetic retinopathy, NPDR: Nonproliferative diabetic retinopathy, DME: Diabetic macular edema, PDR: Proliferative diabetic retinopathy

The overall prevalence of DR (any stage) among all patients with DM at the time of initial presentation was 19.4% (95% confidence interval [CI] 17.5–21.2). This prevalence was fairly consistent across the 3-year study period, with the prevalence rates of 19.5% (95% CI 16.0–22.9) in 2012, 21.7% (95% CI 18.5–25.0) in 2013, and 17.3% (95% CI 14.2–20.3) in 2014.

Of all cases presenting with DR, over half presented with NPDR (1305 patients, 76.1%), for an overall prevalence of 14.7% among all patients with DM. Of patients with NPDR, 416 (24.3% of patients with DR) were noted to have concomitant DME. The remaining 409 patients (23.9% of all cases with DR) presented with PDR, for an overall prevalence of 4.6% among patients presenting with DM, with 201 of these patients (11.7% of patients with DR) having concomitant DME. Across all years of the study, 915 patients (53.4% of patients with DR) were referred for treatment (laser, intravitreal pharmacologics, and/or surgery) at the time of initial presentation.

Forty-eight percent of the overall cohort had an additional systemic comorbidity, and hypertension was the primary comorbidity in 99.5% of all cases with DM [Table 3]. The remaining 0.5% demonstrated a diverse array of comorbidities, including asthma, chronic obstructive pulmonary disease, heart disease, tuberculosis infection, and thyroid dysfunction. Not surprisingly, hypertension was significantly associated with the presence of DR at the initial presentation (P < 0.001, Chi-square test).

Table 3

Factors associated with diabetes mellitus and concurrent diabetic retinopathy

Variables	Categories	Total, n (%)	2012, n (%)	2013, n (%)	2014, n (%)	P
Associated comorbidities (%)	Yes	4258 (48.1)	1398 (54.2)	1224 (42.9)	1636 (47.8)	<0.001
	No	4597 (51.9)	1180 (45.8)	1629 (57.1)	1788 (52.2)
Associated comorbidities	HTN	4222 (99.2)	1391 (99.5)	1199 (98)	1632 (99.8)	<0.001
	HTN with other diseases	13 (0.3)	1 (0.1)	10 (0.8)	2 (0.1)
	Others	23 (0.5)	6 (0.4)	15 (1.2)	2 (0.1)
Duration of onset of DM diagnosis to initial eye consultation (years)	Mean±SD	7.4 (6.6)	7.7 (6.5)	7.1 (6.5)	7.4 (6.8)	<0.05
	Range (years)	0.003-48	0.003-40	0.003-45	0.003-48

HTN: Hypertension, DM: Diabetes mellitus, SD: Standard deviation

The mean duration from initial DM diagnosis to initial eye consultation was also significantly associated with the presence of DR (P < 0.05, one-way ANOVA), with a mean duration of 7.4 years (range: <1 month–48 years) reported across the study period. Less than half the diabetic cohort, 43.2%, presented for an ophthalmic examination within 5 years of their initial diagnosis. An additional 24.3% of patients only sought consultation within the first decade after their diagnosis. No clear trend toward earlier presentation across the study window was noted, with similar patterns of presentation across all 3 years.

Discussion

DM and its ophthalmic complications represent a massive, and growing, public health concern, and one that is especially concerning in low-to-middle-income countries where resources are limited and access to care is difficult.[18] Nepal, like other countries in South Asia, appears to have a nascent epidemic of DM and DR, but published data have been widely varied.[5689101112131415] Previous studies were based on relatively small sample sizes Table 1, so we first sought to examine a larger cross-section of diabetic patients. With a primary cohort of 8855 patients, this stands as the largest review in Nepal to date, and as such, it likely serves as the most reliable source of the in-country DR prevalence for all DM patients presenting for care thus far. A few conclusions may be cautiously drawn from these data then, in spite of the retrospective nature of this study:

First, the DR incidence among patients with DM in Nepal likely accords with the lower end of previously published estimates. The overall prevalence of DR across all years for DM patients presenting to TIO was 19.4%, and this correlates most closely to the two prior studies from the Nepali literature by Shrestha and Shrestha etal.[57] Acknowledging certain limitations within the study design, as well as the selection bias introduced by examining only the population presenting for ophthalmic examination, this is likely the most accurate estimate, to date, for DR prevalence in Nepal's diabetic population. In addition, its near-exact correlation with the community-based study performed by Paudyal et al. lends further credence to this determination.[12]

Second, although likely confirming lower estimates of DR prevalence, the public health concern herein is in no way diminished, as over half of the patients found to have DR upon initial evaluation presented with a vision-threatening complication warranting some measure of initial treatment (i.e., DME and/or PDR). This correlates with the choice for initial therapy among patients found to have DR, with 53.4% being referred immediately for some mode of intervention (laser, intravitreal pharmacologics, and/or surgery), with the vast majority of these cases receiving multimodal therapy.

Third, no dramatic shifts in DR prevalence, severity, or risk factors were noted, over the 3 years of the study period, referral characteristics, demographic data, and DR severity remained relatively stable. This lends credence to our estimates of DR prevalence but does not reveal any changes in public health awareness as, perhaps, a broader population-based study or longer study period might. Future studies may seek to examine these possibilities in greater detail and over longer periods of time. The only exception to this stability was a slight uptick in the number of patient presentations during each year of the study period Table 2, which may reflect the influence of a number of factors: steadily increasing patient volume at TIO, increasing community awareness of diabetic care, or improving networks of patient referrals, for instance.

Correlating patient and provider awareness of the risks of DR may be an additional area for future study. In previous studies from Nepal, awareness of the deleterious effects of DR and the need among diabetic patients for screening has been low.[612] Various studies cite the need for improved patient education and coordination with primary care providers[67812] and increased awareness of DR has been correlated with improved uptake of health services.[19] More accurate data reflecting the prevalence and severity of DR in the Nepali population can assist in these efforts, and our study represents a significant step in that direction.

Certain study limitations do add a measure of caution to our conclusions. As a retrospective study, our study sample is inevitably subject to bias, as it was based in a tertiary care eye center, our cohort may be skewed toward more severe disease, artificially increasing the prevalence of more severe cases of vision-altering DR. Next, our sample was drawn largely from the Kathmandu Valley (with the majority from the urban environs of Kathmandu) and may not reflect the prevalence of DR in the more rural communities of this country. Increased male incidence among patients with DR upon presentation may reflect a true risk factor, or it may reveal gender inequities in access to care, as has been noted previously in some Nepali populations.[20] However, the consistency of DR prevalence and severity across all 3 years of the study period (in spite of variable patient numbers), as well as their approximation of the one prior community-based cross-sectional study,[12] help affirm the observations of this study.

Finally, as the study examined a population of patients being referred to a tertiary eye care center, the results may have been affected by some selection bias in favor of increased disease prevalence (as compared to the actual population prevalence) and increased disease severity (as compared to the actual diabetic population). A 2016 review of approximately 650 patients with DR, selected from the general population using cluster sampling methods, found a much higher proportion of NPDR and lower prevalence of more advanced complications such as DME and PDR (of note, this study also found a DR prevalence of 12% among patients with DM, which approximates the prevalence reported herein).[13] In contrast, a 2017 review of 50 DR patients from an alternate location in the Kathmandu valley demonstrated a PDR prevalence of over 30% among patients with DR, over twice what was detected in our study, and almost five times that found in the 2016 review.[21]

This review detected a more average prevalence between these two extremes, and two details (beyond the number of patients assessed) seem to strengthen the claims of our data in reflecting a more general population: first, although providing tertiary-level ophthalmic care, the TIO receives patients through multiple referral channels, including self-referral and primary care, and therefore, examines a more cross-sectional population than the standard tertiary care center. Second, the correlation of our data with the lower prevalence estimates provided by both published community-based studies (rather than an inflated, higher estimate) supports the concordance of this large study with prior, smaller community surveys.

Conclusions

In summary, we present here the largest cross-sectional survey of the prevalence and severity of DR from a diabetic population in Nepal. We observed a prevalence of DR in this population that was lower than previous estimates, but consistent with a prior community-based survey from Nepal; however, among patients with DR, nearly half presented with a vision-threatening complication requiring initial treatment. These data affirm the growing public health threat of diabetes and its complications and the need for ongoing education and equipping of the ophthalmic community.

Financial support and sponsorship

This study was sponsored through the research department at the Tilganga Institute of Ophthalmology.

Conflicts of interest

There are no conflicts of interest.