Pattern and Presentation of Vitreo-Retinal Diseases: Lessons from a Tertiary Eye Care Centre in Nepal

Background: We examined patients presenting in a tertiary eye hospital in Nepal, focussing on information for screening and management programs for vitreo-retinal disease (VR). Methods: We reviewed all patients presenting for the first time to the VR-clinic over one year. We quantified patient demography, symptoms and duration, associated systemic diseases, ophthalmological examinations, diagnostic investigations and final diagnoses. Results: Of the 1905 cases 1148 were males (60.3%). The 25th-percentile of ages was 29 and 38 years for males and females respectively, indicating females presented later (p<0.0001). Hypertension was the commonest systemic disease (40.8%), followed by diabetes (32.5%). Macular degeneration (AMD) and diabetic retinopathy (DR) affected 447 eyes (11.8%), and 416 eyes (10.9%) respectively. Male and female AMD and DR patients did not differ in age or disease duration, which for DR was not correlated with severity. Asymmetry of disease severity between AMD and DR eyes was largest in patients with one normal eye. Presenting acuity was highly asymmetric between eyes (p<0.0001) with people more often reporting when their dominant eyes had acuity of 6/18 or worse. Conclusions: When left to self-report patients tended to not notice visual impairment in their non-dominant eye until their disease was quite advanced, putting them at risk of serious bilateral disease. Screening of blood pressure and glucose levels combined with fundus photography could prevent many from progressing to life-changing visual impairment and blindness. Later reporting by females began at childbearing age, therefore education and ocular screening could be usefully coupled to mother and child health programs.


Background
Vitreo-retinal (VR) diseases are common causes of visual impairment and blindness.
Population-based studies have reported the overall prevalence of VR disorders to be 8.56% (range 10.4% to 21.02%) among the population aged 40 years and above [1,2].
The 1981 Nepal Blindness Survey reported VR disorders as the third leading cause of bilateral blindness, second only to cataract and its complications [3]. A more recent Nepalese population-based study reported VR disorders to be the second commonest cause of bilateral blindness, second only to cataract, and the most common cause among pseudophakics [4].
A population-based study in Bhutan reported that 22.1% of visual impairment and blindness was due to VR pathologies amongst the population aged ≥ 50 years [5]. A similar survey in Bangladesh reported VR diseases as the second leading cause of bilateral blindness accounting for 13.3% [6], and in India 17.1% among persons aged ≥ 30 years [7]. By contrast a Nigerian study reported a prevalence of 8.1%, with AMD, DR, retinal vein occlusion (RVO) and retinal detachment (RD) as the most common retinal diseases [8]. An Ethiopian study reported RD as the commonest cause of both bilateral (59.4%) and unilateral (41.2%) blindness [9]. The Tehran study reported VR prevalence of 8.56% with acquired retinopathies and peripheral lesions as the most common retinal diseases [2].
In developed countries AMD affects nearly 10% of those over 65 years of age, and 25% over 75 years [10], including Australia [11]. In the USA, more than 8 million people have intermediate AMD and nearly 2 million have advanced AMD [12]. In the UK, the prevalence of late AMD was 2.4% among the population aged 50 years and above, 4.8% for ≥ 65 years and 12.2% for ≥ 80 years [13].
DR often affects adults of working age [14]. The American National Health and Nutritional Examination Survey 2005-2008 reported that 28.5% of diabetic patients had some degree 4 of DR, and 4.4% had vision-threatening DR [15]. In 2012 global prevalence was 34.6% for any DR, 6.96% for proliferative DR, 6.81% for diabetic macular edema, and 10.2% for vision-threatening DR [16]. India and China are confronting a growing epidemic of diabetes and DR [17][18][19].
The current study focuses upon the pattern and laterality of VR diseases presenting to a tertiary eye care centre in Nepal

Study population
This study covers all cases presenting to the VR clinic for the first time over one year.
Patients who presented for repeat or follow-up visits were not included here. The study ran from 01.01.2010 until 31.12.2010. A key objective for publishing this study now is to provide a baseline reference for planned follow-up studies that will use similar analyses. A further benefit of recording this hospital-based study is that there are population-based studies of disease prevalence from both Nepal [4,20] and Bhutan [5] [21] to estimate the population means and the standard errors of the means (SE) of those percentiles. We employed 10,000 bootstrap cycles to insure the estimated means and SE converged to within 2 decimal places on 5 independent crossvalidations. We then applied t-tests employing the estimated means and SE. The 1905 6 subjects insured that the bootstrap estimates were statistically conservative.

Results
We will first give an overview of the general presentation of the 1,905 cases. We next present the ocular disease data as: 1) the non-retinal diseases involving posterior segment other than retina, and 2) the main analysis of the retinal diseases. Some less relevant data is presented in 4 Supplementary tables.

General presentation
During the study year 1,905 new cases presented to the VR clinic: 1,148 males (60.3%) and 757 females (39.7%). Their ages ranged from 0.17 years to 116 years, with mean of 49.14, median of 54 and mode of 70 years (83 cases = 4.4%). Fig. 1A gives a breakdown of the presenting ages in 20-year cohorts. Fig. 1B shows that for the whole cohort the 25th percentile is 29 years for males and 38 years for females. A bootstrap analysis (Methods) revealed that the 10th, 25th, 50th and 75th percentiles were significantly older for females than males by 3.95, 9.28, 7.25, and 3.08 years: median 5.6 years. The 10th percentile difference was significant at p<0.003, and the others at p = 0.0001. Notice that the 25th percentiles for the males is generally lower in the 20-year cohorts of Fig. 1A  could not affirm the time period due to missing data (Table S1).
Hypertension was the commonest systemic disease associated in 250 cases (40.8%), followed by diabetes in 199 cases (32.5%) and combined diabetes and hypertension in 124 cases (20.2%) (Fig. 2). The other systemic diseases found in 40 cases (6.5%) are summarised in Table S2. The duration of systemic disease association was found to be less than 5 years in 42.8% of cases, 5 to 10 years in 33.2% and more than 10 years in 24.0%.
[ Figure Table 2.   In part because of the higher rate of smoking by males in Nepal [22], we decided to examine wet AMD eyes relative to fellow eyes. Two patients had familial drusen bilaterally, and for 14 this analysis they were classed as having dry AMD. Four patients had a non-visible fundus (mainly due to cataract) and were eliminated from the analysis. The ages of males whose worst eye had wet AMD was not different to females: 66.0 ± 15.3 vs. 65.8 ± 13.7 years.
There were 40 males whose worst eye had wet AMD, vs. 15 females, which was marginally significant (p=0.063, t-test, correcting for the relative abundance on males in the study population). Interestingly, patients whose worst eye had dry AMD were older than those with a worst eye that was wet AMD, 69.5 ± 14.8 vs. 65.9 ± 10.5 years (p=0.045). Overall there was a suggestion of males developing wet AMD earlier, and in relatively greater numbers relative to females. This might have been an effect of smoking, but needs further investigation. We also examined laterality by scoring normal fundus to wet AMD on a scale from 1 to 3 (normal, dry, wet) and then examining the absolute value of the difference in scores between eyes. Six patients had wet AMD OU. For the 186 patients whose best eye had dry AMD, only 30 had a worst eye with wet AMD. Patients whose best eye was normal were even more heterogeneous. Of those 36 patients 17 had a fellow eye that was dry, and 19 had a fellow eye that was wet. Thus the pattern of progression seemed to be relative heterogeneity early, progressing to ever more bilateral disease, rather than simple bilaterally at each stage. Like the AMD patients the DR group contained more males (123) than females (91), but this was not significant. Their ages also did not differ at 58.7 ± 9.91 years for males, and 58.3 ± 11.21 years for females. Their durations of diabetes (DM) did not differ at 11.5 ± 6.79 and 10.4 ± 6.96 years respectively. Nine patients had a non-visible fundus in one eye and were removed from further analysis. We examined the laterality of DR in the remaining 205 patients. We scored the five DR diagnostic categories from normal fundus to PDR as 1 to 5.
We then binned subjects according to the diagnosis in their least affected eye. To quantify the degree of laterality we took the absolute value of the difference of these severity steps in each pair of eyes. A boxplot of the results is shown in Fig. 4A. Basically, as with AMD, eyes tended to become more similar as severity of the best eye increased. As shown in Fig. 4B severity of DR was not correlated with duration of disease.

Discussion
The visual handicap experienced by individuals suffering from unilateral eye disease, like macular hole, is strongly influenced by ocular dominance [23]. Ocular-dominance and handedness are associated, with about 65% of right-handers, and 43% of left-handers, being right eye-dominant [24]. Some of our data (Fig. 3) suggested eye dominance played a role in patient reporting to the hospital: with the predominantly right-handed subjects only reporting once their right eye had BCVA worse than 6/18. AMD and DR were very homogenous with respect to age and sex, however when the best eye was normal the range of severity in the fellow eye was surprisingly broad (e.g. Fig. 4). Taken together the results mean that if that eye was a dominant eye then patients tended not to notice their sight threatening disease in their fellow eye, an excellent argument for more screening. Given the results on presenting BCVA (Fig. 3) we subsequently analysed other data by eye (Tables  3, 4, S4) in order to elucidate any eye-wise biases, however few were found.
The preponderance of males in the total group was 60.3%, and for AMD and DR patients combined was 56.5%, which was marginally more than females (p=0.052). This agrees with the findings of some hospital-based studies on VR diseases [9,25], but differs from a Nepalese population-based study, in which only 45.5% were males [20], and one in India with only 45% males [1]. Our results could be due to women being reserved due to social norms and therefore do not come forward for medical check-ups. Here the 25th percentile of ages reporting was nearly 10 years higher for females (p<0.0001, t-test 9.27 years; 95% CL 6.05 and 12.5 years). The mean and median age of 49 and 54 years for males and females tally well with other studies of VR diseases [1,9]. Overall ages ranged from 2 months to 116 years. Considering the life expectancy at birth for Nepal is 68.3 years for men and 71.5 years for women [26], the oldest patient in the study, a man aged 116 years, was unexpected.
Only 41 cases (2.2%) presented for routine check-up. Chronic retinal diseases like AMD, DR, RD, RVO, RP, etc., ranked high in the diagnostic list but only 2.2% of cases were found to present for routine check-up indicating that there is a need to emphasize patient education and counselling about the importance of regular check-ups and follow ups.
In 53 cases the interval between the onset of symptoms and presentation to the hospital could not be confirmed due to the lack of recorded data. 565 cases ( [27]. Remote areas and associated difficulty in accessing medical attention, mean people initially rely on traditional methods of healing. Systemic disease association was not found in 1,292 of the 1,905 cases. As shown in Fig. 2 (and Table S2) among the 613 remaining cases the commonest associated diseases were hypertension (40.8%), diabetes (32.5%), and both (20.2%). This agrees with the Tehran eye study which reported hypertension (21.14%), followed by diabetes (15.99%) [2]. By contrast, a Nigerian study found diabetes to be more common (14.6%) than hypertension (13.2%) [25]. Other systemic diseases were found in 40 of our cases (6.5%, Table S2). The duration of systemic disease association was less than 5 years in 210 cases (42.8%), 5 to 10 years in 163 cases (33.2%) and more than 10 years in 118 cases (24.0%).
The commonest VR disease was AMD affecting 11.8%, followed by DR at 10.9%; matching another Nepalese study reported that AMD was the commonest VR disease at 28.3%, followed by DR at 17.9% [20]. They do not match a Nigerian hospital study, which reported DR (24.9%) as the commonest VR disease, followed by hypertensive retinopathy (13.3%) and AMD (24.9%) [25]. In our case 208 of 324 (64.2%) confirmed diabetic cases had some form of DR, which did not tally with population-based studies reporting only 10.5% [1]. In our study 117 left eyes and 113 right eyes of 323 patients had CSME but a population-based study in Nepal found it in only 2 of 305 diabetic cases [20]. This disparity could be explained by the early onset of macular edema causing a high percentage of patients of CSME to present to the TIO.
RD was the third commonest VR disease in our study affecting 112 right eyes (5.9%) and 126 left eyes (6.6%), while a the Nepalese population-based study reported population prevalence of only 0.10% [20]. A hospital-based study in Ethiopia reported RD as the second commonest VR disease at 24.5% [9]. Of 65 cases of FTMH in our study 33 cases (50.8%) affected right eyes, while others have reported only 48% involving right eyes in one study [23]. In our study 94.8% of the FTMH and 84.6% of the LMH were unilateral, which are similarly reported in other studies as macular holes (full thickness or lamellar) and are basically the consequence of factors affecting the macula locally. When such unilateral macular diseases affect the dominant eyes of the individuals they cause greater functional visual impairment [23].
In eastern Asia pathological myopia is a major issue found in 80-90% of school-leavers, and 10-20% of those completing secondary school [28], and contributes to RD numbers. The prevalence of myopia varies from 0.8% to 53.4%, depending on geographical area, age, occupation and ethnicity [29][30][31]. In Nepal, Sherpa children had a prevalence of 2.9% as compared to 21.7% for Tibetan children [32]. A myriad of myopic complications like atrophic retinal holes and RD, choroidal neo-vascular membranes, degeneration, cataract and glaucoma cause visual loss warranting attention [33]. Thapa et al. [20] reported the population prevalence of macular hole as 0.20% in Nepal. Asymptomatic macular holes occur at a prevalence of 6.26% among high myopes with more than -20D [34].

Limitations
Nineteen patients were excluded due to incomplete data. AMD was not classified as per the AREDS classification system. Only the final diagnoses, and not ocular comorbidities, have been considered for analysis which might have altered the reported disease patterns.

Conclusion
This study indicates that low cost screening and management programs for retinal disease could be of immense value in developing countries. We found that without screening 20 programs patients tend to not notice developing visual impairment in their non-dominant eye, often until it is too late. In Nepal, and perhaps in similar countries, females report later for care than men. That was true from child baring age, so education and screening could be usefully coupled to child health programs. The high prevalence of hypertension and diabetes amongst retinal disease patients suggest that a simple screening of blood pressure and glucose levels combined with fundus photography could prevent many from progressing to life-changing visual impairment and blindness.

Declarations
Ethical approval and consent to participate The study has been approved and the need for consent was waived by the Tilganga Institute

Availability of data and materials
The data have not been placed in any online data storage. The datasets generated and analysed during the study are available upon request from the first author.  Presenting BCVA: The presenting BCVAs provided significant asymmetries between eyes (p<0.0001) in the two groups with acuities of 6/60 or better (chi-square = 158 for "6/18 or better", and 283 for "6/18 to 6/60") between DR-severity steps in the two eyes when the severity levels (normal to PDR) are scored as 1 to 5. Generally the DR became more bilaterally symmetric. A small amount of uniformly distributed noise (0 to 0.03) was added to the data to make the outliers more distinct (red+). (B) Diabetes duration as a function of the best eye. Duration did not appear to strongly determine severity.

Supplementary Files
This is a list of supplementary files associated with the primary manuscript. Click to download.