• Users Online: 252
  • Print this page
  • Email this page


 
 Table of Contents  
ORIGINAL ARTICLE
Year : 2017  |  Volume : 8  |  Issue : 1  |  Page : 1-6

Incidence and factors associated with diabetic retinopathy among diabetic patients at arbaminch general hospital, gamo gofa Zone (longitudinal follow up data analysis)


1 Department of Public Health, Welkite University, Hawassa, Ethiopia
2 Department of Public Health, Mekelle University, Mekelle, Ethiopia

Date of Web Publication9-May-2017

Correspondence Address:
Yilma Chisha
Department of Biostatistics and Health Informatics, Arba Minch College of Health Sciences, Hawassa
Ethiopia
Login to access the Email id

Source of Support: None, Conflict of Interest: None


DOI: 10.4103/jod.jod_6_17

Get Permissions

  Abstract 

Background: Currently, 93 million people are estimated as living with diabetic retinopathy (DR) worldwide. The incidence, retinopathy-free survival time and associated factors of DR in developed countries have been well documented; but in Ethiopia, national data on incidence and associated factors of DR are lacking. Objective: The objective of this study is to determine incidence and factors associated with the development of DR among diabetic patients at Arbaminch General Hospital, Ethiopia. Materials and Methods: Longitudinal follow-up data analysis with record review of 400 diabetic patients was conducted at Arbaminch General Hospital. Among 400 diabetic patients, 270 diabetic patients with baseline information and without a history of hypertension at baseline were included in this study. Whereas, pregnancy induced diabetes and patients with retinopathy at baseline were excluded from this study. Consecutive sampling technique was applied to select study participants. Data of cohorts were extracted from medical record using pre-tested structured extraction checklist. Data cleaning, coding, categorising, merging and analysis carried out by STATA version 12. Descriptive statistics was done and summarised accordingly. Bivariate Cox proportional hazard regression analysis was done to select potential candidates for the full model atP ≤ 0.25 and multivariable Cox proportional hazard regression analysis was made to estimate the independent effect of predictors on the incidence of DR. Model diagnostic tests were performed and final model fitness was checked by Cox and Snell residuals; finally, statistical significance was tested atP < 0.05. Results: Overall incidence of DR at Arbaminch General Hospital among diabetic patients ever enrolled since 1990 E.C. was ~36 cases per 1000 patients per year and an estimated median time to develop was approximately 10 years. The incidence of diabetic retinopathy was high among patients whose disease duration was >12 years. Adjusted analysis showed that the hazard of developing DR was statistically and positively associated with baseline age, baseline systolic blood pressure (SBP) level and fasting blood glucose (FBG) level. Conclusion: In the current study, the incidence of DR was high. Since baseline age, baseline SBP level and high FBG level were statistically and positively related with the development of DR; special care should be given in addition to routine care.

Keywords: Diabetes, diabetic retinopathy, incidence


How to cite this article:
Chisha Y, Terefe W, Assefa H. Incidence and factors associated with diabetic retinopathy among diabetic patients at arbaminch general hospital, gamo gofa Zone (longitudinal follow up data analysis). J Diabetol 2017;8:1-6

How to cite this URL:
Chisha Y, Terefe W, Assefa H. Incidence and factors associated with diabetic retinopathy among diabetic patients at arbaminch general hospital, gamo gofa Zone (longitudinal follow up data analysis). J Diabetol [serial online] 2017 [cited 2017 Jul 21];8:1-6. Available from: http://www.journalofdiabetology.org/text.asp?2017/8/1/1/205982


  Introduction Top


Diabetes mellitus (DM) is a metabolic disorder of carbohydrate, fat and protein and it affects the body's ability to process and use glucose for energy.[1] Main causes for DM are defects in insulin secretion, insulin action, or both.[1],[2] About 5%–10% and 90%–95% of patients with diabetes have Type 1 and Type 2, respectively.[1],[2] People with diabetes are increasing due to population growth, aging, urbanisation and sedentary lifestyle.[3] Since the last decade, DM has emerged as an important clinical and public health problem throughout the world and its prevalence reached an epidemic proportion.[4],[5]

In 2014, 422 million people in the world had diabetes with the prevalence of 8.5% in the adult population.[6] The prevalence of diabetes has been steadily increasing for the past three decades and is growing most rapidly in low- and middle-income countries.[4],[5],[6] The epidemic raised in diabetes poses significant public health and socioeconomic challenges through diabetic complications, of which diabetic retinopathy (DR) or damage to the small blood vessel of retina is the most common and feared diabetic complication that result in blindness.[7]

There are 93 million people are approximated as living with DR worldwide.[8] According to the systematic review report, DR affects 7%–63% of Sub-Saharan diabetic patients; and in Tikur Anbessa Specialized Hospital, Ethiopia, it was about 37.8%.[9] Among diabetics In Jimma University Tertiary Hospital, Ethiopia, the prevalence of DR was 25.4%.[10] Incidence and retinopathy-free survival time in developed countries have been well documented, and risk factors are well-known. In Ethiopia, national data on incidence and associated factors of DR are lacking. This effort constitutes the first attempt to estimate the incidence and associated factors of DR among diabetics at Arbaminch General Hospital.


  Materials and Methods Top


This longitudinal follow-up data analysis with record review was conducted at Arbaminch General Hospital from 27th November 2014, to 12th January 2015 on diabetic patients ever enrolled since 1990 E.C. Arbaminch Hospital was located at Arbaminch town, the capital city of Gamo Gofa zone, which is 505 km from Addis Ababa (capital city of Ethiopia) and 280 km from Hawasa, Center of Southern Nation's Nationalities and People Regional State. Arbaminch has two sub-cities, Secha and Sikela. The hospital is located at Secha which is the administrative centre of Arbaminch town.

Although Arbaminch Hospital is technically a regional hospital, it is acting as a referral hospital and provides preventive, curative and rehabilitative care for 100,000–200,000 people per year, and of them, more than 400 was diabetic patients.

Data collection procedure, sampling and collecting instrument

After taking medical record number of diabetic patients from chronic care follow-up clinic, the patient folder was drawn from card room by health information technician. Among 400 diabetic patients under follow, 270 patients with baseline information and without a history of hypertension at baseline were selected using consecutive sampling technique. Whereas, pregnancy induced diabetes and patients with retinopathy at baseline were excluded from the study.

The record reviews were done by two Bachelor of Science (BSc) nurses and facilitated by principal investigators. Training was given for data extractors before extraction; pre-test was done on 5% of patients outside the study area. Data of cohorts was extracted from medical record by using pre-tested structured checklist which was taken from previous studies. After completing data extraction, it was transferred to STATA analysis software version 12 for cleaning, coding, categorising, merging and to check completeness, consistence and outliers or extreme values.

Study variables

Dependant variables

Diabetic retinopathy

DR was taken as an event when diabetic patients confirmed for the development of DR by the ophthalmologist. The ascertainment of DR status was confirmed by a history of clinical presentation, visual acuity test result, slit lamp microscope examination and direct ophthalmoscope examination findings. Meanwhile, DR was taken as censored when diabetic patients did not develop DR throughout the study period, loose to follow-up, died before the onset of DR, withdraw or drop out, transferred out.

Time

It is time to clinical endpoint. This is the difference of index date or date of being confirmed for diabetes and date of event or censorship.

Independent variables at baseline

Age, sex, place of residence, family history of diabetes, level of systolic blood pressure (SBP), level of diastolic blood pressure (BP), blood glucose level, duration of diabetes, type of diabetes, type of anti-diabetic agents taken are covariates at baseline.

Statistical analysis

After data pre-analysis work was completed, it was analysed by STATA version 12 (STATA company, Stata Corp, USA, Texas) and summarised accordingly. Frequency with percentage to categorical variables and median along with interquartile range (IQR) to skewed continuous variables was used. Bivariate Cox proportional hazard regression was carried out to select potential candidate predictors to the full model with cut-off point P≤ 0.25.[11]

Multivariable Cox proportional hazard regression analysis was done to estimate the independent effect of predictors on the occurrences of DR. Model was built and compared by stepwise backwards elimination procedure and likely hood ratio test, respectively. Interactions and confounders were checked by change in beta coefficient with cut-off point beta change >20%.[11]

Proportionality assumption of Cox proportional hazard regression for the fitted model was checked using global test (significance test) based on Schoenfeld residuals. Similarly, Instability of beta-coefficient (multicollinearity) for variables in the final fitted model was checked using variance inflation factor (VIF) with cut-off point mean VIF >10.[12] Goodness of fit for the final fitted model was checked using Nelson-Aalen cumulative hazard function against Cox-Snell residual and classifying ability was checked by Harrell's concordance statistics. Association between predictors and hazard of DR was summarised using adjusted hazard ratio (AHR), and statistical significances were tested at P < 0.05. Finally, Model equation was written as:



h (t) = h0(t) e [1.94 (baseline age) + −2.57 (duration of diabetes) + 1.40 (baseline SBP level) + 0.002 (baseline FBG level)]. Where,

h (t) = Hazard rate at time t

h0 (t) = The baseline hazard at time 0

X1., Xk = Are potential predictor variables

FBG = Fasting blood glucose.

Ethical consideration

This study was conducted after getting Ethical approval from Mekelle University Ethical Review Board, permission to undertake the study was obtained from Gamo Gofa Zone Health Department and Arbaminch General Hospital.


  Results Top


Sociodemographic characteristics of study subjects

From the total study subjects, 220 (81.5%) were within age category <60 years. With regard to gender, 138 (51.1%) of study subjects were male. Concerning to place of residence, 150 (56.6%) of study participants were from the urban residence. Among 270 study subjects, 231 (85.9%) has no family history of diabetes [Table 1].
Table 1: Descriptive statistics of diabetic patients in Arbaminch Hospital, 2007 E.C. (n=270)

Click here to view


Clinical and biochemical characteristics of study subjects

Of the total study subjects, 200 (74.1%) were Type 2 diabetic. Concerning to duration of diabetes, 213 (78.9%) were with diabetic duration <6 year. Pertaining towards anti-diabetic agents taken, 78 (28.89%) were used insulin alone, 25 (9.26%) were used insulin and oral anti-diabetic agents together, and the rest 167 (61.85%) were used oral anti-diabetic agents alone to manage their diabetes. Looking to the baseline SBP and diastolic BP level, 232 (85.9%) and 240 (88.9%) of study subjects have baseline SBP and diastolic BP level ≤140 mmHg and ≤90 mmHg, respectively. Fasting plasma glucose level was not normally distributed among cohorts and the median (IQR) fasting plasma glucose level was 176 (IQR, 145–253) mg/dl. Similarly, follow-up time was not normally distributed and the median follow-up time with IQR was 2.6 years (IQR, 1.25–5.75) [Table 1].

Survival analysis

Patients were followed for median of 2.6 years. The incidence rate of DR was assessed every 3 years interval. The incidence of DR was high on diabetic patients with diseases duration >12 years and followed by diseases duration 6–9 years. An overall incidence rate of DR among diabetic patients at Arbaminch General Hospital was 35.96 cases per 1000 persons per year [Table 2].
Table 2: Overall incidence and incidence rate of diabetic retinopathy every 3 years interval among diabetic patients in Arbaminch General Hospital, 2007 E.C. (n=270)

Click here to view


The probability of median retinopathy-free survival time (50% of patients will counteract DR, and 50% will be free from DR was approximately 10 years [Figure 1].
Figure 1: Kaplan–Meier survival estimates of median diabetic retinopathy-free survival time among diabetic patients in Arbaminch General Hospital, 2007 E.C. (n = 270)

Click here to view


Bivariate Cox proportional hazard regression analysis

Bivariate Cox proportional hazard regression analysis was done to select potential candidate variables to the multivariable analysis. Based on the pre-set P value criteria cut-off point ≤0.25,[10] except baseline diastolic BP level, other predictors satisfied the criteria and are a potential candidates for the multivariable analysis [Table 3].
Table 3: Bivariate Cox proportional hazard regression analysis of diabetic patients in Arbaminch General Hospital, 2007 E.C. (n=270)

Click here to view


Multivariable Cox regression

In multivariable Cox proportional hazard regression analysis, baseline age, duration of diabetes, baseline SBP level and baseline fasting plasma glucose level have statistically significant association with development of DR.

Factors influencing development of diabetic retinopathy

By holding the effect of baseline SBP level, duration of diabetes and baseline fasting plasma blood glucose level constant, hazard of developing DR was almost seven times higher in patients with baseline age ≥60 years than their counterparts (AHR = 6.9: 95% confidence interval [CI]; 3.25–14.83). Surprisingly, hazard of developing DR was inversely related with duration of diabetes; it is 92% less likely for patients with duration of diabetes 6 years and above than their counterparts (AHR = 0.08: 95% CI; 0.03–0.22).

After stabilising the effect of other covariates in the model, hazard of developing DR was more than four times higher for patients with baseline SBP level >140 mmHg than their counters parts (AHR = 4.1: 95% CI; 1.76–9.44). Hazard of developing DR was increased by 0.2% when fasting plasma glucose level increase by 1 mg/dl (AHR = 1.002: 95% CI; 1.00–1.003) [Table 4].
Table 4: Adjusted Cox proportional hazard regression analysis of diabetic patients in Arbaminch General Hospital, 2007 E.C. (n=270)

Click here to view



  Discussion Top


The current study revealed that an overall incidence of DR among diabetic patients at Arbaminch General Hospital was 35.9 cases per 1000 persons per year. This figure is high compared to prospective cohort study finding in Rochester Minnesota, USA.[13] That is 17.4 cases per 1000 person-years at risk. Similarly, the figure is high from reported incidence of DR in Tanzania (15 per 100,000 per year) and in Sudan (101 per 100,000 per year).[14] Possible reasons for this inconsistency might be a variation on skill and knowledge of physicians to diagnose DR and methods used to detect. Furthermore, it might be due to difference in health care system and quality of care given for diabetic patients. In cross-sectional study conducted in Jimma University Specialized Hospital [15] diabetic patients developed DR within 5–9 years after diagnosis; but in the current study diabetic patients developed DR with the median time of 9.9 years. Possible reasons for this inconsistency might be underestimation of time diagnosed for diabetes in previous study. Moreover, it might be due to delayed presentation of patients to the health facility after onset of diabetes in previous study. As well, might be due to overestimation of time to event in the current study as some patients developed DR after long follow-up time. Despite this, the finding is consistent with study finding in the UK [16] that is about 10 year.

The current study showed that duration of diabetes was negatively associated with hazard of DR and it is not consistent with longitudinal study finding on Type 1 diabetic patients in England [16] and cross-sectional study finding in Jimma University Tertiary Hospital.[15] Possible reasons for this inconsistency might be majority of patients with DR in the current study were Type 2 diabetics, so they are prone to come late to the health facility because Type 2 diabetes is not as severe as Type 1 and patients with Type 2 diabetes develop DR within short duration of diseases compared to Type 1.[17] Besides, further study is needed to come across the difference.

Elevation of SBP level is a risk factor for chronic non-communicable illness; as well, baseline SBP level was significantly and positively associated with the hazard of DR. This finding is in line with longitudinal study finding of THE United Kingdom.[18] Despite this, the finding of community-based cross-sectional study in Melbourne, Australia, revealed that, no statistically significant association between level of SBP and occurrence of DR.[19] Possible reason for this irregularity might be a variation of patients on self-care practice, variation on BP measuring device and skill of health care providers in measuring BP.

As baseline fasting plasma glucose level increases the hazard of DR correspondingly increase. This finding is consistent with study findings in Benghazi, Libya [20] and West Africa.[21]

Diabetes affects both individuals and their families and has an impact on economic and social development of a country. Information on availability, cost and quality of medical care for diabetes does mostly not exist for many low- and middle-income countries including Ethiopia. Complications from diabetes, which can be devastating, could largely be prevented by wider use of several inexpensive medicines, simple tests and monitoring and can be a cost-saving intervention.

This study will provide an in-depth and comprehensive picture of impacts of diabetes and DR and propose clear recommendations for improving prevention and management of diabetic complications. It will help to develop programs and policies for better management of diabetics and cost effective strategies in the Ethiopian context.

Strengths and limitations

Strength of the study

Extracted information's were recorded in the past at the time when a patient came to the health facility, so the collected data were not depended on patient's memory and it minimized recall bias. Since exposure status of diabetic patients was measured before the outcome, it allows determination of temporal sequence between exposure and outcome.

Limitations of the study

Since the study was conducted using pre-recorded data and the data were not collected for research purpose, there might be a lack of accuracy. Because of institutional-based nature of study and non-probability sampling technique, findings cannot be generalized for total population. The current study was not used experimental design, so it does not show causal relationship among predictors and DR. Due to lack of full records on smoking, physical exercise; level of blood cholesterol and body mass index (BMI), the association of these predictors with the incidence of DR was not estimated.


  Conclusion Top


The incidence rate of DR was high and an estimated median time to develop was approximately 10 years. The study also showed that recommended follow-up examinations and tests for BMI and blood cholesterol analysis are neither done nor documented.

The clinic has no standardized intake form and electronic database system for diabetic patients so that the effects of some of the most important variables were not estimated. Baseline age, baseline SBP level, duration of diabetes and fasting plasma glucose level were statistically and significantly associated risk factors with the incidence of DR. Due to limitations of conclusions drawn from relatively small sample size, we strongly recommended other scholars to conduct prospective study with large sample size to estimate the real life determinants of DR.

Recommendations

Based on the findings of the current study, the following recommendations are drown for the concerned bodies to overcome the problem and its unwelcomed effects. Arbaminch general hospital officials and professionals should strive to reduce the incidence of DR through enhancing quality of care and patients self-care practice. They should provide eye evaluation service as recommended by the WHO at least twice annually for all diabetic patients. In addition to the routine care, especial emphasis should be given for patients with baseline age ≥60 years, high baseline SBP (>140 mmHg) and high baseline FBG level (>140 mg/dl). Knowing and controlling the level of BMI, blood cholesterol level and lifestyle related factors were important primary prevention strategies for chronic non-communicable illnesses; these factors should be routinely examined, evaluated and recorded. Finally, prospective study is highly recommended to identify real life determinants of DR.

Acknowledgement

First of all, we would like to thank Mekelle University college of health sciences department of public health for providing an opportunity and fund to conduct this study. Next, we would also like to give our heartfelt for officials and professionals of Gamo Gofa Zone Health Department and Arbaminch General Hospital for permission to use the data and for welcomed face. Finally, we would like to gratitude data collectors, friends and all individuals who helped directly or indirectly for the success of this study.

Funded agency for this study is Mekelle University.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.



 
  References Top

1.
Fauci AS, Kasper DL, Longo DL, Braunwald E, Hauser SL, Jameson JL, editors. Harrison's Principles of Internal Medicine. 17th ed., Vol. 2. United States of America: McGraw-Hill Inc.; 2008.  Back to cited text no. 1
    
2.
Surrena H, editor. Hand Book for Brunner & Suddarth's, Textbook of Medical-Surgical Nursing. 12th ed., Vol. 4. Philadelphia: Wolters Kluwer Health/Lippincott Williams and Wilkins; 2010.  Back to cited text no. 2
    
3.
Wild S, Roglic G, Green A, Sicree R, King H. Global prevalence of diabetes: Estimates for the year 2000 and projections for 2030. Diabetes Care 2004;27:1047-53.  Back to cited text no. 3
    
4.
IDA. Diabetic Facts; 2011. Available from: http://www.worlddiabetesfoundation.org. [Last accessed on 2014 Apr 30].  Back to cited text no. 4
    
5.
IDF. Diabetes in the Developing World; 2011. Available from: http://www.worlddiabetesfoundation.org/media. [Last accessed on 2014 Apr 2014].  Back to cited text no. 5
    
6.
World Health Organization (WHO), Global Report on Diabetes; 2016. Available from: http//www.who.int. [Last accessed on 2016 Jun 27].  Back to cited text no. 6
    
7.
A Vision 2020th Right to Sight, Guidelines for the Comprehensive Management of Diabetic Retinopathy in India. Aravind Eye Care System: New Delhi, India; 2008.  Back to cited text no. 7
    
8.
Yau JW, Rogers SL, Kawasaki R, Lamoureux EL, Kowalski JW, Bek T, Chen S, et al. Global prevalence and major risk factors of diabetic retinopathy. Diabetes Care 2012;35:556-64.  Back to cited text no. 8
    
9.
Hall V, Thomsen RW, Henriksen O, Lohse N. Diabetes in Sub Saharan Africa 1999-2011: Epidemiology and public health implications. A systematic review. BMC Public Health 2011;11:564.  Back to cited text no. 9
    
10.
Abdella SH, Adem M. Awareness of diabetic patients about their illness and associated complications in Ethiopia. Med Sci 2013;2:512-22.  Back to cited text no. 10
    
11.
Bursac Z, Gauss CH, Williams DK, Hosmer DW. Purposeful selection of variables in logistic regression. Biomed Cent 2008;3:1-8.  Back to cited text no. 11
    
12.
Dorman CF, Elith J, Bacher S, Buchman C, Carl G, Care G. Collinearity: A review of methods to deal with it and a simulation study evaluating their performance. Vol. 1. Nordic Society Oikos; 2012. p. EV1-20.  Back to cited text no. 12
    
13.
Dwyer MS, Melton LJ 3rd, Ballard DJ, Palumbo PJ, Trautmann JC, Chu CP. Incidence of diabetic retinopathy and blindness: A population-based study in Rochester, Minnesota. Diabetes Care 1985;8:316-22.  Back to cited text no. 13
    
14.
Mbanya JCN, Motala AA, Sobngwi E, Assahi FK, Enoru S. Diabetes in Sub-Saharan Africa. Lancet 2010;375:2254-66.  Back to cited text no. 14
    
15.
Worku D, Hamza L, Woldemichael K. Patterns of diabetic complications at Jimma University specialized hospital, Southwest Ethiopia. Ethiop J Health Sci 2010;20:33-9.  Back to cited text no. 15
    
16.
Jones CD, Greenwood RH, Misra A, Bachmann MO. Incidence and progression of diabetic retinopathy during 17 years of a population-based screening program in England. Diabetes Care 2012;35:592-6.  Back to cited text no. 16
    
17.
Sleire L. Diabetes: A Neglected Disease in Sub-Saharan Africa, a Comparative Study between Rwanda and Norway. Medical Student Thesis. Faculty of Medicine, University of Oslo; June, 2011.  Back to cited text no. 17
    
18.
Matthews DR, Stratton IM, Aldington SJ, Holman RR, Kohner EM; UK Prospective Diabetes Study Group. Risks of progression of retinopathy and vision loss related to tight blood pressure control in type 2 diabetes mellitus: UKPDS 69. Arch Ophthalmol 2004;122:1631-40.  Back to cited text no. 18
    
19.
Brazionis L, Rowley K Sr., Itsiopoulos C, Harper CA, O'Dea K. Homocysteine and diabetic retinopathy. Diabetes Care 2008;31:50-6.  Back to cited text no. 19
    
20.
Roaeid RB, Kadiki OA. Prevalence of long term complications among type 2 diabetic patient s in Benghazi, Libya. J Diabetol 2011;3:1-8.  Back to cited text no. 20
    
21.
Rotimi C, Daniel H, Zhou J, Obisesan A, Chen G, Chen Y, et al. Prevalence and determinants of diabetic retinopathy and cataracts in West African type 2 diabetes patients. Ethn Dis 2003;13 2 Suppl 2:S110-7.  Back to cited text no. 21
    


    Figures

  [Figure 1]
 
 
    Tables

  [Table 1], [Table 2], [Table 3], [Table 4]



 

Top
 
 
  Search
 
Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
Access Statistics
Email Alert *
Add to My List *
* Registration required (free)

 
  In this article
Abstract
Introduction
Materials and Me...
Results
Discussion
Conclusion
References
Article Figures
Article Tables

 Article Access Statistics
    Viewed326    
    Printed38    
    Emailed0    
    PDF Downloaded75    
    Comments [Add]    

Recommend this journal


[TAG2]
[TAG3]
[TAG4]