Diabetic retinopathy , visual impairment and ocular status among patients with diabetes mellitus in Yemen : A hospital-based study

Results: Our series comprised 350 patients suff ering from DM. The duration of diabetes was ≥15 years in 101 (29%) patients. Physician was treating 108 DM patients with insulin. The prevalence of DR was 55% (95% CI 49.660.1). The proportions of background diabetic retinopathy (BDR), preproliferative diabetic retinopathy (PPDR), proliferative diabetic retinopathy (PDR) and diabetic macular edema were 20%, 13%, 17% and 22% respectively. The prevalence of blindness among DM patients was 16%. The prevalence of cataract and glaucoma was 34.3% and 8.6%. Duration of DM was the predictor of DR. One-Þ ft h of the patients had sight-threatening DR and needed laser treatment.

Diabetic retinopathy (DR) is a priority disease in the 'VISION 2020' initiative for the global elimination of avoidable blindness.The World Health Organization (WHO) has recommended its member countries to integrate a program approach for DR within their prevention of blindness programs. [1]In industrialized countries, the magnitude of DR is high and it is the leading cause of blindness. [2]But countries with rapidly evolving economies and metros of developing countries also face the challenge of epidemic proportion of DR. [3,4] To our knowledge, no information on eye complications of diabetes in the Yemeni population has been reported in the past.
We treat patients of both the middle and lower middle-income groups in our institute.Eye examination with the help of modern instruments like indirect ophthalmoscope, bio-microscope, Volk lenses, applanation tonometer, autoperimeter, gonioscopes, etc is possible in our institute.Medical care services were provided at the diabetic center.This center has facilities to diagnose and manage Diabetes Mellitus (DM) and its systemic complications.The objective of our study was to estimate the magnitude and selected risk factors of eye complications of DM in our institution.On its basis, we also recommend a public health approach to address DR in the study area.

Materials and Methods
The ethical and research committ ee of our university hospital approved this study.We obtained writt en consents from the patients to participate in this study.The procedures followed were in accordance with the ethical standards of the responsible committ ee on human experimentation (institutional or regional) and with the Helsinki Declaration of 1975, as revised in 2000.This cross-sectional type of study was conducted from January 2004 to December 2004.The patients with DM att ending ophthalmic clinic were our study population.These patients were referred by a physician for their eye examination.
We included '1 to 6' patients with diabetes on Saturdays, '7 to 12' on Mondays, '13 to 20' on Wednesdays.On an average the ophthalmologists used to see 20 DM patients in a day.The patients were sequentially enrolled and those giving consent to participate were examined as per protocol for this study.
One physician and two ophthalmologists who were experienced in treating DM and DR were our Þ eld investigators.DM was deÞ ned as a person having fasting glucose level of ≥ 7 mmol /liter.If a patient was already taking medicine to control hyperglycemia, he/she was labeled as a case of 'previously physician-diagnosed' diabetes. [5]A medical history was obtained to determine the medical treatment being given and the duration of DM.The duration of DM was deÞ ned as the time interval in years between the date of Þ rst time diagnosis of DM and the date of present evaluation.A person was deÞ ned to suff er from hypertension if three repeated measurements at diff erent instances in the day showed a reading of more than 140 mm Hg systolic blood pressure and/or >90 mm Hg diastolic blood pressure. [6]If patient was already taking medicine to control hypertension, he/she was labeled as hypertensive even though blood pressure measurements were within the normal range.
To assess the presence of systemic co-morbidities, the physician performed detailed clinical examination of cardiovascular system, nervous system, renal system.Electrocardiogram was obtained.Laboratory tests were performed for renal function tests and complete blood count and lipid proÞ le.A person with diabetes who had clinically detectable albuminuria (≥300 mg/L) and did not have other renal disease explaining protein loss in urine was considered to suff er from diabetic nephropathy.If macrovascular late complications resulted in atherosclerosis in vessels of heart and patient had symptomatic coronary artery disease, the person was deÞ ned to have cardiac complication of diabetes.If a person suff ered from symmetric distal and predominantly sensory polyneuropathy mainly resulting in stocking glove type of sensory loss, the person was considered to have complication of diabetic neuropathy. [7]hthalmologists examined all these patients.Vision of each eye was noted with the best possible correction.Snellen's projection chart was used for this purpose.The WHOrecommended deÞ nitions of visual disabilities were adopted in our study.Blindness was deÞ ned as vision less than 10/200 in bett er eye aft er correction.Low vision was deÞ ned as vision less than 20/60 in bett er eye aft er correction.Anterior segment of each eye was examined using slit-lamp bio-microscope (Haag Streit).The ocular pressure was measured by an applanation tonometer (Zeiss).For patients suspected to have glaucoma, the Þ eld of vision was tested on Octopus automated perimeter.We asked patients to move their eyes in all eight directions to test the ocular mobility.
The pupils were dilated by instilling one drop of 1.0% tropicamide.If the pupil did not dilate aft er 30 min, we added one drop of 2.5% phenylephine to the previous one.The fundus was examined with +90 D Volk lens and bio-microscope.This enabled us to have a stereoscopic view of the retina and its vasculature.The presence and grading of DR was according to the International clinical DR and macular edema disease severity scale. [8]If exudates, microaneurysms and hemorrhages were present in retina but in an area other than the macula, we graded DR as background diabetic retinopathy (BDR).
Presence of avascular zone and development of neovascular vessels was graded as preproliferative diabetic retinopathy (PPDR).If gliosis was present on the optic disc or along with blood vessels, the stage was graded as proliferative diabetic retinopathy (PDR).If the macula had edema or exudates with or without avascular zone, it was termed as diabetic maculopathy.If retinal detachment due to proliferation of glial tissue exerted traction, we graded the stage as tractional retinal detachment.If blood was present in vitreous and obscured details of the retina, we graded the stage as vitreous hemorrhage.We did not perform validity test and assumed that the diagnosis of two senior ophthalmologists adopting similar classiÞ cation of DR would be similar.
The management of DR as recommended by the ophthalmologist was also evaluated.The WHO has given guidelines for laser treatment of cases with DR.We had used these guidelines for deÞ ning criteria for active intervention (laser treatment or surgery); in the rest of the cases primary prevention was recommended. [2]pre-tested form was used to collect the information for this study.The data was entered in an Excel Microsoft ® spreadsheet.It was checked for consistencies and duplications.For a univariate type of parametric analysis, we used Statistical Package for Social Studies (SPSS 11.5).Frequencies, percentage, their 95% conÞ dence intervals were calculated.To compare the rates among variables, we used STATCALC of EPI6 info soft ware to calculate Odd's ratio, its 95% conÞ dence interval and Mantel-Henzal P value.To review the interaction of diff erent risk factors, we carried out a binominal regression analysis.Presence or absence of DR was the dependent variable.Age, sex, duration of DM, hypertension and type of treatment for DM were the independent variables.A variable that did not have statistical signiÞ cance was removed from the regression model.
All the cases with DR were explained about their ocular status and treatment was given at concession rates.DM patients with blindness or low vision disability were referred for rehabilitative services.

Results
The physician and ophthalmologists examined 350 patients with DM in our study.The mean age when DM was Þ rst diagnosed in these patients was 44 years (Standard Deviation = 12.6 years).The mean age of patients with DR was 54.4 years (Standard Deviation = 11.8 years).One hundred and eighty-four (52%) patients had diabetes for more than 10 years.One hundred and eleven (31.7%) patients were taking injection insulin to treat their diabetes.Physicians noted that cardiac, cerebro-vascular, renal, peripheral vascular, neuropathy and other complications were present in 13, 8, 1, 2, 1 and 17 patients respectively.The mean duration of DM was 9.9 years (Standard Deviation = 7.77 years).
The prevalence of DR in our series was 55% (95% ConÞ dence interval 49.6-60.1).The rate of DR by sex, duration, type of treatment and hypertension was calculated [Table 1].The DR was signiÞ cantly associated with the duration of DM (χ 2 = 33.9,degree of freedom = 3, P < 0.001).
We also compared the magnitude of DR in our patients with the rates reported in other studies in the Middle Eastern countries [Table 2].
The patients with DM were categorized as no DR, BDR, PPDR and PDR.The rates of DR with macular edema and DR with rubeosis iridis were calculated separately.The number and percentage of DR cases and their 95% conÞ dence intervals are given in Table 3. Bilateral serous macular edema was noted in 52 (14.9%) and unilateral serous macular edema was observed in 18 (5.1%)patients.Bilateral and unilateral ischemic macular edema were found in four (1.1%) and three (0.9%) patients respectively.
The presence of co-morbidity in eye with less vision was taken into account to calculate magnitude of co-morbidity [Table 4].In addition to DR, cataract and glaucoma were present in a signiÞ cant number of cases with DM.
The presence or absence of DR was the dependent variable.
In the binominal regression model, we included age, sex, and duration of diabetes, and the mode of treating DM.The adjusted Odds ratio, 95% conÞ dence intervals and P value are given in Table 5. Duration of DM was the predictor of DR in our study.

Discussion
The rate of DR among patients with DM in our study conducted in an eye unit of a hospital in the capital of Yemen was 54.9%.The rate of diabetic macular edema was 22%.The duration of diabetes was positively associated with the presence of DR.
The prevalence of DM was reported to be 4.6% and 9.75% in two diff erent studies conducted in Yemen. [9,10]Risk factors such as obesity, impaired glucose tolerance, hypertension and hyperlipidemia exist in the Yemeni population. [9]Thus the burden of DM is likely to be signiÞ cant and with improved socioeconomic conditions, especially in urban areas of Yemen, it is likely to further rise. [11]With improved health services, mortality due to renal and cardiac complications of DM will reduce and patients with DM would live longer. [12]DR that is associated with the duration of DM is therefore likely to increase. [13]Hence an organized approach to address DR within 'VISION 2020' initiative is now recommended. [14]For proper planning of a public health program, evidence-based  El-Asrar AM, et al. [13] 1998 502 Saudi Arabia
Our study had a few limitations.With a limited sample of patients visiting one institution, the results should be extrapolated with caution.The institution-based case selection could have introduced health-seeking bias.In addition, misclassiÞ cation bias could have been introduced since the sample included a case mix of patients referred by physicians and patients presenting at the eye clinic directly.
The prevalence of DR (54.9%) among patients with DM in our study was high.Eye complications of DM in the Yemeni population are expected to be high because familial clustering and high rate of consanguinity are the genetic risk factors for DM which are reported to exist in the Yemeni population. [15,16]In addition, acquired factors like hypertension, obesity and hyperlipidemia are also reported to be common in this population. [9,10]A study in Yemen in 1997 had shown a rate of 45% micro-vascular complications in diabetic patients. [11]he patients in this study belonged to the manual labor class and this study did not represent the 'well to do' families of Sana city of Yemen.Risk of diabetes and eye complications of diabetes in this group are likely to be lower compared to the urban Yemeni population of our study.Hence comparison of the outcomes of this study with the results of the present study should be done with caution.
The rate of DR in our study was lower than that (64.1%) reported in Jordan. [12]This could be due to longer duration of DM and use of fundus ß uorescein angiography (FFA) in the Jordan study.The mean duration of DM was 12 years in the Jordan study compared to 9.9 years in our study.In a study conducted in the United Arab Emirate (UAE), the prevalence of DR was 19%; much lower compared to our study. [13]Use of fundus photography to document the retinal changes of DR in UAE could be more precise.In Oman, the eye screening of patients with DM was introduced in 2000 while the diabetes control program through primary healthcare existed since 1990.The study showing 14.5% in Oman in 2003 could have been due to free and easy access to the health services to registered diabetics in Oman. [14]A study in Iran covered patients with non-insulin-dependent diabetes patients with a mean duration of diabetes of 6.9 years.This could explain the DR rate of 8.9% reported in the Ishfahan province of Iran. [15]A study in Lebanon that had 112 diabetic patients and retinopathy evaluation was conducted by using direct and indirect ophthalmoscopes reported 17% prevalence of DR. [16] In a study of diabetic patients in a hospital of Cairo, the prevalence of DR in diabetics was 42. 2%. [17]Thus it is evident that even in the Arab population of diff erent Middle Eastern countries there was wide variation of DR.Diff erent lifestyles, care for diabetes, time studies conducted and variation in tribes of Arab population could be the reasons for this variation.
The rates of DR were similar among males and females in our study.In contrast, Al Maskari et al., and Khandekar et al.,  reported higher rates of DR in males. [18,19]Less access of eye care to female patients in early stages of diabetes in Yemen could have resulted in this observation.
Duration of diabetes was positively associated with DR in our study.It was also noted in many other studies. [2,20,21]Early detection of diabetes through screening and regular follow-up and primary prevention is therefore recommended to reduce the risk of severe blinding complications of DR.
The prevalence of DR has been documented to be higher in Type 1 DM compared to Type 2 DM. [22]We had information on patients currently being treated with insulin and/or by other treatment modalities.We found that the rate of DR among patients using insulin was signiÞ cantly higher compared to those using other medications.But long duration of DM and poor glycemic control could have prompted physicians to treat these cases with insulin.Thus association of type of treatment with DR could have been confounded by the duration and poor glycemic control in our study.PDR was found in 17.1% of our patients.In Oman, El Hadad reported PDR in 12.8% patients. [22]High magnitude of PDR in our study could be due to lack of facilities for treating DR and poor compliance of patients for the laser treatment and for ignoring primary prevention of DM.It is interesting to note that the rate of maculopathy was 22% in our study.Al-Adsani reported 10.3% with maculopathy in Saudi Arabia and it was 5.1% in Oman. [3,14]n the absence of a large number of cases with nephropathy, we cannot explain the reason for a high rate of maculopathy in our study.The resources and skills required for treating maculopathy are diff erent compared to the laser treatment for other stages of DR.The national programs aiming to off er care to cases of DR with maculopathy should plan accordingly.
The rate of un-operated cataract in patients with DM in our study was 34.3%.It was 38% in Jordan. [12]In the capital areas in both these studies, lack of eye care services could not be the reason for the backlog of cataract in patients with DM.Perhaps the criteria to operate cataract in DM patients in Jordan and Yemen might be diff erent compared to those adopted for senile cataract cases.
Glaucoma was noted in 8.6% of patients with DM in our study.A study in Oman reported that 8.9% of patients with DM were suff ering from glaucoma. [24]sual disabilities among diabetics are reported to be signiÞ cantly higher compared to the general population. [25]The rates of bilateral blindness and unilateral blindness matched with a study in Jordan. [17]This information will be useful to the national planners of VISION 2020 in Yemen.Program approach adopted to address DR will reduce visual disabilities.The intervention strategies proposed for patients with DR in our study were primary prevention, prophylactic treatment by laser and management of macular complications.But awareness campaigns to improve regular follow-up and primary prevention have to be organized.Eye care services at an aff ordable cost should also be made available to all patients with DM.

Conclusion
The prevalence of DR was high in our study.A public health approach is recommended to address eye complications of DR in the study area.Information on co-morbidities in the eye and visual disabilities among patients with DM should be further veriÞ ed by studies with a larger sample and representing other parts of Yemen.