Abstract
By 2030, sub‐Saharan Africa is forecast to see the steepest rise in the number of people with hypertension of any world region. Hypertensive retinopathy is known to be a common complication of hypertension in developed countries and some studies suggest it is associated with the presence of other hypertension‐related end‐organ damage (EOD) such as stroke and cardiovascular disease. In Tanzania hypertension is relatively more common than in other parts of sub‐Saharan Africa, especially in the older population; however, the prevalence of hypertensive retinopathy and its association with EOD remain unknown. The authors conducted a cross‐sectional study of elderly, community‐dwelling, rural Tanzanians to determine the prevalence of hypertensive retinopathy and its association with hypertension and other forms of EOD. Hypertensive retinopathy was diagnosed based on retinal imaging. In a cohort of 61 patients with gradable images, the authors found the overall prevalence of hypertensive retinopathy to be 64% (n = 39), which was strongly associated with hypertension (X 2 [1] = 4.207, P = .004), with a significant trend towards more severe retinopathy with more severe hypertension (r = .377, P = .003). The authors did not find hypertensive retinopathy to be associated with other forms of EOD. Hypertensive retinopathy is highly prevalent in this population and is associated in most but not all cases with hypertension. These findings do not suggest that it could be used as a screening tool for EOD, but it is important to identify and educate patients with retinopathy about possible complications of the condition.
Keywords: end‐organ damage, hypertension, low‐ and middle‐income countries, retinal imaging, retinopathy, sub‐Saharan Africa, Tanzania
1. INTRODUCTION
1.1. Hypertensive end‐organ damage in SSA
Sub‐Saharan Africa (SSA) will see the largest growth of any world region in the proportion of people living with hypertension.1 In SSA, where diet, lifestyle, patterns of disease, and genetic factors vary markedly from those seen in high‐income settings, the effects of hypertension on the organs of the body are less well understood than in most high‐income settings.2 Part of the reason for this may be that most methods of end‐organ damage (EOD) assessment are either invasive or require analysis of urine or blood samples, raising issues of cultural appropriateness and affordability, particularly in a community setting.
1.2. Hypertensive retinopathy in SSA
Hypertension leads to a series of changes in the retinal microvasculature termed hypertensive retinopathy. There is considerable interest in whether retinal imaging, to identify retinopathy, may be a means of predicting hypertension‐related damage elsewhere in the body.3 Furthermore, retinal imaging is unlikely to be influenced by the white‐coat effect (whereby observations are altered by the procedure of testing), which can artificially distort findings from blood pressure monitoring, and moderate its importance as a predictor of clinical outcomes. There is recent evidence that the white‐coat effect is relatively common in SSA.4 However, studies of hypertensive retinopathy in SSA have yielded highly variable results, making it hard to generalize the findings with regard to prevalence and correlates.5, 6, 7, 8, 9, 10
We assessed the prevalence of hypertensive retinopathy in a community‐dwelling population of older adults living in rural Tanzania known to have high background rates of uncontrolled hypertension.11 We also investigated the association between hypertensive retinopathy and other markers for EOD.
2. METHODS
2.1. Ethics and consent
Ethical approval was obtained from the National Institute of Medical Research, Dar‐es‐Salaam, Tanzania. All participants provided informed consent. Those who were unable to write to sign their name provided a thumbprint. Assent from a close relative was obtained for participants who lacked capacity.
2.2. Setting
This study was conducted as part of a 6‐year follow‐up of a cohort of people aged 70 years and older who lived in the rural Hai district of northern Tanzania. A full discussion of the study population is provided elsewhere and briefly described below.12 A population census was completed at baseline in 2009‐2010 and reported the population of the 52 villages of the Hai district demographic surveillance site as 161 119. At baseline, 12 villages were randomly selected and from these a cohort of 2232 people 70 years and older was recruited. During the 6‐year follow‐up, one village was selected at random as a pilot site for a study of retinal imaging as a means of identifying EOD. From this village, 73 people were recruited to the study. As for the demographic surveillance site as a whole, the village has an economy based largely on agriculture at a subsistence level.
2.3. Timing and recruitment
Data were collected between February 29 and May 13, 2016. Participants at baseline data collection who had died or who refused or were unable to attend the clinic for imaging were excluded. Any participant who failed to attend the clinic after three invitations was considered to have refused participation.
2.4. Assessment
Assessment took place at the participants’ local health clinic. Methods for data collection of basic demographics (age, sex, weight, socioeconomic status, and education), arterial hypertension, and EOD have been previously described.2 EOD markers used included 12‐lead electrocardiography to assess left ventricular hypertrophy, IDEA cognitive screen to assess cognitive function, ankle‐brachial pressure index to assess peripheral arterial disease and arterial stiffening, urine dip stick to assess proteinuria (a marker of renal impairment), and self‐reported stroke history. Further details on these assessment methods have been previously published.2 Three blood pressure measurements were obtained with the patients in a seated position, and the average of the second and third readings was used, as per the WHO STEPwise approach to surveillance protocol. Hypertension was classified in alignment with current UK National Institute for Health and Care Excellence and British Heart Foundation guidelines, which define hypertension as grades I, II, and III (see Table 1 for cutoffs).13 Further lower classifications of <120/80 mm Hg (normotensive), 120 to 129/<80 mm Hg (elevated), and ≥130/80 mm Hg (hypertensive) are also presented to reflect recent updates to US guidelines.14
Table 1.
Demographic, health makers, and hypertension grade by retinopathy grade
| Retinopathy grade | Total | |||
|---|---|---|---|---|
| None | Moderate | Severe | ||
| Demographic characteristics and health status | ||||
| Median age (interquartile range), y | 82.5 (78‐85.3) | 80 (78‐84) | 77 (76.5‐79.3) | 80 (78‐84) |
| Men, No. (%) | 14 (63.6) | 14 (42.4) | 3 (50.0) | 31 (50.8) |
| Smoking, No. (%) | ||||
| Never | 16 (72.7) | 23 (69.7) | 5 (83.3) | 44 (72.1) |
| Former | 6 (27.3) | 8 (24.2) | 1 (16.7) | 15 (24.6) |
| Current | 0 | 2 (6.1) | 0 | 2 (3.3) |
| Alcohol | ||||
| Never | 9 (40.9) | 9 (27.3) | 1 (16.7) | 19 (31.1) |
| Former | 9 (40.9) | 12 (36.4) | 4 (66.7) | 25 (41.0) |
| Current | 4 (18.2) | 12 (36.4) | 1 (16.7) | 17 (27.9) |
| Known diabetes mellitus, No. (%) | 0 | 2 (6.1) | 1 (16.7) | 3 (4.9) |
| Hypertension status 2016 | ||||
| Normotensive (<140/90 mm Hg) | 10 | 6 | 2 | 18 |
| Grade I hypertension (140‐159/90‐99 mm Hg) | 11 | 8 | 1 | 20 |
| Grade II hypertension (160‐179/100‐109 mm Hg) | 0 | 12 | 1 | 13 |
| Grade III hypertension (>180/110 mm Hg) | 1 | 7 | 2 | 10 |
| Total | 22 | 33 | 6 | 61 |
Visual acuity was tested using the best available correction and with a pinhole if visual acuity was ≤ 6/12. Pupils were then dilated with topical tropicamide 1%. Two 45° retinal photographs (one optic disc–centered and one macula‐centered) were taken by personnel from the Kilimanjaro Diabetic Screening Programme using a fundus camera (Topcon Medical Systems, Inc., Oakland, NJ). The images were reviewed and graded by two medical students (RJ and HWIP) who underwent training and were deemed competent in image interpretation by a consultant ophthalmologist (DS) experienced in image grading. Each student graded the images independently and was masked to the other study data, with arbitration grading, in cases of discrepancy by DS (discrepancy occurred in 5% of gradings). Ungradable images based on image quality were recorded and excluded. Retinopathy signs were evaluated in each image noting the following findings: generalized arteriolar narrowing, focal arteriolar narrowing, arteriovenous nicking, opacity (“copper wiring”) of arteriolar wall, hemorrhage (blot, dot, or flame‐shaped), microaneurysm, cotton‐wool spot, hard exudate, and disc swelling. The individual signs from each image were then used to form a final grade as per the three‐point hypertensive grading system (mild, moderate, and malignant) proposed by Wong and Mitchell15 and based on the highest graded image per patient. The moderate and malignant stages were grouped together for analysis, because of the small number within the malignant group. Cases with obvious alternative causes of signs such as hemorrhage or exudate associated with age‐related macular degeneration were noted and excluded.
2.5. Statistical analysis
Statistical analysis was supported by SPSS for Windows version 23 (IBM, Armonk, NY). Age data were not normally distributed and thus were described in terms of the median and interquartile range. All remaining data were categorical and were thus summarized in frequencies and associations investigated through odds ratios with 95% confidence intervals. Fisher exact test was used to assess significant differences. Spearman rho was used to assess correlation. Significance was set at P = 5%.
3. RESULTS
Of the baseline cohort, 143 were still alive and were approached, of whom 73 (51.0%) were able to attend the clinic for retinal imaging and provided consent. Of 73 images, 61 (83.6%) participants had a set of retinal images that could be interpreted and coded.
Demographic and health status variables for the 61 participants are presented in Table 1. Although 11 people were taking antihypertensives, all were hypertensive on assessment; therefore, no one had hypertension that was controlled by medication. Of those with hypertension at grade I or above, 72% (Table 1) had hypertensive retinopathy (X 2 [1] = 4.207, P = .004). There was a general trend towards the presence of retinopathy being more severe with increasing hypertension grade (r = .377, P = .003). Interestingly 44.4% of normotensive participants had hypertensive retinopathy. The association between retinopathy and hypertension at the low cutoff value of 130/80 mm Hg is detailed in Table 2.
Table 2.
Relationship between hypertension status and retinopathy at lower cutoffs14
| Retinopathy grade | Total | |||
|---|---|---|---|---|
| None | Moderate | Severe | ||
| <120/80 mm Hg | 3 | 1 | 2 | 6 |
| 120‐129/<80 mm Hg | 2 | 1 | 0 | 3 |
| ≥130/80 mm Hg | 17 | 31 | 4 | 52 |
Table 3 summarizes the association between hypertensive retinopathy and other markers of EOD. There was no statistically significant association between any of the other markers for EOD and hypertensive retinopathy.
Table 3.
Markers for end‐organ damage and hypertension status with regard to hypertensive retinopathy
| Signs of hypertensive retinopathy (n = 39) | No signs of hypertensive retinopathy (n = 22) | Odds ratio (95% confidence interval) | |
|---|---|---|---|
| Hypertension (≥140/90 mm Hg) | 31 (79.5) | 12 (54.5) | 3.2 (1.0‐10.1) P = .040 |
| Grade II or above hypertension (≥160/100 mm Hg) | 22 (56.4) | 1 (4.5) | 27.2 (3.3‐222.7) P < .001 |
| LVH | 2 (5.1) | 2 (9.1) | 0.5 (0.1‐4.1) P = .615 |
| PAD | 11 (28.2) | 4 (18.2) | 1.8 (0.5‐6.4) P = .383 |
| Proteinuria | 4 (10.5), 1 missing value | 3 (14.3), 1 missing value | 0.7 (0.1‐3.6) P = .691 |
| Cognitive impairment | 3 (7.7) | 3 (14.3), 1 missing value | 0.5 (0.1‐2.9) P = .655 |
| Arterial stiffening | 1 (2.6) | 1 (4.5) | 0.6 (0.03‐9.3) P = 1.000 |
| Previous stroke | 1 (2.6) | 1 (4.5) | 0.6 (0.03‐9.3) P = 1.000 |
Values are expressed as number (percentage). Bold values indicate statistical significance.
LVH, left ventricular hypertrophy; PAD, peripheral arterial disease.
4. DISCUSSION
Our study is the first to look at prevalence of hypertensive retinopathy and its relation to hypertension and hypertensive EOD in Tanzania. Assessments were relatively fast and noninvasive and the imaging equipment was available locally. Although the numbers were small for this pilot study, and our study is not definitive, retinal imaging appeared to be of limited use in identifying individuals with other forms of EOD as measured in this study. However, patients with hypertensive retinopathy are at greater risk of hypertension‐related eye pathology including retinal artery or vein occlusion and ischemic optic neuropathy; thus, identifying these patients with retinal imaging would provide an opportunity for education and treatment.16
4.1. Comparison of prevalence with other data from SSA
The prevalence of hypertensive retinopathy among Tanzanians with hypertension was on the higher end of the spectrum when compared with other studies conducted in SSA. This, in part, is a reflection of high background rates of hypertension and low rates of effective control.11 Other studies have reported prevalence rates of hypertensive retinopathy of 20% to 70%.5, 6, 7, 8, 9, 10 One possible reason for this broad range is differences in imaging techniques and interpretation of images. For example, a study in Nigeria used a Kowa portable fundus camera attached to a smartphone.5 Our study used a stationary fundus camera that produced widefield, high‐quality images. Interobserver variability makes comparison between studies problematic and most studies use just one person to interpret images. We had two, and where necessary three, people assessing the images. Another explanation for the variability in prevalence estimates is the age discrepancy between the populations studied. None of the previous studies focused solely on older people. Nevertheless, some studies of participants in this age range have also found a high prevalence.6, 7, 10 Genetic diversity could also play a role. There is greater genetic diversity within Africa than in the rest of the world combined.17 Consequently, comparing our findings with those from Nigeria, Ghana, Congo, and Afro‐Caribbeans of any origin living in London may be of limited value.5, 6, 7, 8, 9, 10 Environmental factors are also important to consider. In the Hai region of Tanzania, salt consumption rates are some of the highest in SSA, thus hypertension is not only more prevalent but also more severe.18
4.2. Hypertensive retinopathy in participants with normotension
Two fifths of individuals with normotension have hypertensive retinopathy. Only one other study has looked at the prevalence of hypertensive retinopathy in normotensive individuals of African origin.7 They found a prevalence of 13% in Afro‐Caribbean men and 20% in Afro‐Caribbean women aged 40 to 64 years.
A number of factors could explain our findings. First, evidence exists that hypertensive retinopathy precedes hypertension in some cases1 Diabetes mellitus may also have played a role, as it produces retinal vascular changes that can be indistinguishable from those caused by hypertension.19 A limitation of our study is that, as a result of limited resources, blood glucose measurements were not performed. Only three patients reported having diabetes mellitus and, although all had signs of retinopathy, they were also hypertensive, making the assignment of a cause difficult. Nevertheless, diabetes mellitus is not commonly encountered in rural Tanzania.20
Using the lower cutoff for the presence of hypertension of 130/80 mm Hg identified all but nine people as having hypertension, making it hard to draw any conclusions regarding the relationship between retinopathy and hypertension. However, even at this cutoff, four of the nine nonhypertensives had signs of retinopathy. This tends to emphasize that, in this population where hypertension and/or the white‐coat effect are common, reliable methods to identify those most at risk for hypertension‐related adverse events are needed.
4.3. Hypertensive retinopathy and EOD
The prevalence of EOD in our population reflects the lack of adequate diagnosis and treatment of hypertension in this setting.2 Although changes in diet may help control hypertension to some extent, few people in this setting are able to make active dietary choices. Farming is at a subsistence level and crops are grown that give a reliable yield in this climate; maize is a staple for most people and usually made into porridge. The lack of correlation between hypertensive retinopathy and other forms of EOD has been documented in other studies both in Africa and in the United Kingdom.8, 21 Our population may be particularly resistant to EOD, given that they have survived to over 75 years in this resource‐poor setting. Other factors such as the presence of diabetes mellitus, smoking, and hypercholesterolemia are also likely to alter the pattern of EOD. Despite our findings, a significant body of evidence exists for a relationship between hypertensive retinopathy and stroke.8, 21, 22 Other studies have found that hypertensive retinopathy correlates with stroke even in the absence of hypertension.5 The lack of association between retinopathy and stroke in our cohort is likely to reflect the small numbers seen and low rates of stroke awareness in Hai, making self‐report a crude measure of stroke prevalence.23
4.4. Limitations
In addition to the limitations previously acknowledged, we recognize that we were only able to recruit 73 of 143 people approached to be assessed. The main reasons for nonrecruitment were lack of time to attend assessments and immobility. This is likely to indicate that those recruited may not have been representative of the background population, and, therefore, this may have influenced our findings. The limitation of the methods of accurately detecting EOD (most notably for left ventricular hypertrophy and proteinuria), are reviewed in our previous publication on EOD in this cohort.2 Finally, we recognize that our findings may be influenced by the effect of lifelong sun exposure on the retina.
5. CONCLUSIONS
Hypertensive retinopathy was highly prevalent among our cohort, although it was not associated with the presence of other forms of EOD. Our findings do not suggest that hypertensive retinopathy is sensitive or specific enough to be used as a tool for screening for hypertension nor its related EOD in this population. Larger studies to determine its reliability as a risk stratification tool for stroke could highlight further utility of retinal imaging in this population.
CONFLICT OF INTEREST
There are no conflicts of interest to declare.
AUTHOR CONTRIBUTION
Design/conception: Richard Walker, William Gray, Rebecca Jones, Harry Putnam, Charles Cleland, and David Steel. Literature search: Rebecca Jones and Joanna Klaptocz. Data collection: Rebecca Jones, Harry Putnam, and Heiko Philippin. Data analysis: William Gray, Rebecca Jones, and Joanna Klaptocz. Interpretation of results: Richard Walker, William Gray, Joanna Klaptocz, Harry Putnam, and Rebecca Jones. Writing of article and review: all authors.
ACKNOWLEDGMENTS
We wish to acknowledge all of the help we have received from the healthcare workers, translators, village enumerators, officials, caregivers, family members, and participants in the Hai district who helped with assessments, examinations, data collection, and input. We would like to thank Norma Cardill and Victoria Ferguson for assisting in data entry for the 2013 follow‐up, and Gillian Tough for administrative support (all Northumbria Healthcare NHS Foundation Trust).
Jones R, Putnam HWI, Philippin H, et al. Retinal imaging to identify target organ damage in older Africans: A pilot study. J Clin Hypertens. 2018;20:1296–1301. 10.1111/jch.13352
Funding information
This study received no specific funding. RJ and HWIP completed this work in part as fulfillment of the requirements of a Master's in Research degree at Newcastle University, UK.
REFERENCES
- 1. World Health Organization . A global brief on hypertension: silent killer, global public health crisis. Geneva, Switzerland: World Health Organization; 2013. [Google Scholar]
- 2. Putnam HWI, Jones R, Rogathi J, et al. Hypertension in a resource‐limited setting: is it associated with end organ damage in older adults in rural Tanzania? J Clin Hypertens (Greenwich). 2018;20:217‐224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Ong YT, De Silva DA, Cheung CY, et al. Microvascular structure and network in the retina of patients with ischemic stroke. Stroke. 2013;44:2121‐2127. [DOI] [PubMed] [Google Scholar]
- 4. Ivy A, Tam J, Dewhurst MJ, et al. Ambulatory blood pressure monitoring to assess the white‐coat effect in an elderly East African population. J Clin Hypertens (Greenwich). 2015;17:389‐394. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Oluleye ST, Olusanya BA, Adeoye AM. Retinal vascular changes in hypertensive patients in Ibadan, Sub‐Saharan Africa. Int J Gen Med. 2016;9:285‐290. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Ladipo GO. Hypertensive retinopathy in Nigerians. A prospective clinical study of 350 cases. Trop Geogr Med. 1981;33:311‐316. [PubMed] [Google Scholar]
- 7. Sharp PS, Chaturvedi N, Wormald R, McKeigue PM, Marmot MG, Young SM. Hypertensive retinopathy in Afro‐Caribbeans and Europeans. Prevalence and risk factor relationships. Hypertension. 1995;25:1322‐1325. [DOI] [PubMed] [Google Scholar]
- 8. Kabedi NN, Mwanza JC, Lepira FB, Kayembe TK, Kayembe DL. Hypertensive retinopathy and its association with cardiovascular, renal and cerebrovascular morbidity in Congolese patients. Cardiovasc J Afr. 2014;25:228‐232. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Akinkugbe OO. The rarity of hypertensive retinopathy in the African. Am J Med. 1968;45(3):401‐404. [DOI] [PubMed] [Google Scholar]
- 10. Addo J, Smeeth L, Leon DA. Hypertensive target organ damage in Ghanaian civil servants with hypertension. PLoS One. 2009;4:e6672. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Dewhurst MJ, Dewhurst F, Gray WK, Chaote P, Orega GP, Walker RW. The high prevalence of hypertension in rural‐dwelling Tanzanian older adults and the disparity between detection, treatment and control: a rule of sixths? J Hum Hypertens. 2013;27:374‐380. [DOI] [PubMed] [Google Scholar]
- 12. Gray WK, Dewhurst F, Dewhurst MJ, et al. Rates and predictors of three‐year mortality in older people in rural Tanzania. Arch Gerontol Geriatr. 2016;62:36‐42. [DOI] [PubMed] [Google Scholar]
- 13. National Institute for Health and Clinical Excellence (NICE) . Hypertension: Clinical management of primary hypertension in adults (CG127). London, UK: NICE; 2011. [Google Scholar]
- 14. Whelton PK, Carey RM, Aronow WS, et al. 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood pressure in adults: a report of the American College of Cardiology/American Heart Association Task Force on clinical practice guidelines. Hypertension. 2018;71:e13‐e115. [DOI] [PubMed] [Google Scholar]
- 15. Wong TY, Mitchell P. Hypertensive retinopathy. N Engl J Med. 2004;351:2310‐2317. [DOI] [PubMed] [Google Scholar]
- 16. Wong TY, Mitchell P. The eye in hypertension. Lancet. 2007;369:425‐435. [DOI] [PubMed] [Google Scholar]
- 17. Tishkoff SA, Reed FA, Friedlaender FR, et al. The genetic structure and history of Africans and African Americans. Science. 2009;324:1035‐1044. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Oyebode O, Oti S, Chen YF, Lilford RJ. Salt intakes in sub‐Saharan Africa: a systematic review and meta‐regression. Popul Health Metr. 2016;14:1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Wong TY, McIntosh R. Hypertensive retinopathy signs as risk indicators of cardiovascular morbidity and mortality. Br Med Bull. 2005;73–74:57‐70. [DOI] [PubMed] [Google Scholar]
- 20. 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. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Bhargava M, Ikram MK, Wong TY. How does hypertension affect your eyes? J Hum Hypertens. 2012;26:71‐83. [DOI] [PubMed] [Google Scholar]
- 22. Cheung CY, Ikram MK, Chen C, Wong TY. Imaging retina to study dementia and stroke. Prog Retin Eye Res. 2017;57:89‐107. [DOI] [PubMed] [Google Scholar]
- 23. Walker R, Whiting D, Unwin N, et al. Stroke incidence in rural and urban Tanzania: a prospective, community‐based study. Lancet Neurol. 2010;9:786‐792. [DOI] [PubMed] [Google Scholar]