© 2010 Wichtig Editore - ISSN 1120-6721 Eur J Ophthalmol (2011; :4) 484-48921 DOI: 10.5301/EJO.2010.5977 484 INTRODUCTION Sickle cell hemoglobinopathies are a group of inherited diseases characterized by an abnormality in the β-chain of the hemoglobin molecule. Sickle cell disease (SCD) is found all over the world except in the Far East and in the Arctic countries. It is most common in Africans and people of African descent but also occurs in the non-black popula- tion of Saudi Arabia and India. The chief manifestations are Ocular manifestations of sickle cell disease at the Korle-bu Hospital, Accra, Ghana Alfred Osafo-Kwaako1, Kahaki Kimani1, Dunera Ilako1, Stephen Akafo2, Ivy Ekem2, Onike Rodrigues2, Christabel Enweronu-Laryea2, Martin M. Nentwich 3 1Department of Ophthalmology, University of Nairobi, Nairobi - Kenya 2University of Ghana, Medical School Accra, Accra - Ghana 3Department of Ophthalmology, Ludwig-Maximilians-University Munich, Munich - Germany Purpose. To determine the magnitude and pattern of ocular manifestations in sickle cell disease at Korle-bu Hospital, Accra, Ghana. Methods. Hospital-based cross-sectional study including all patients with sickle cell disease reporting for routine follow-up at the Sickle Cell Clinic at Korle-bu Hospital, Accra, Ghana. Results. A total of 201 patients with sickle cell disease (67 male and 134 female) were enrolled, com- prising 114 subjects with genotype HbSS, aged 6-58 years, mean 19.26 (SD 11.70), and 87 with genotype HbSC, aged 6-65 years, mean 31.4 (SD 16.76). Visual impairment was found in 5.6% of eyes examined. Causes were cataract, proliferative sickle retinopathy (PSR), optic atrophy, phthisis bulbi, and central retinal artery occlusion. Common anterior segment signs of sickle cell disease, which were more common in HbSC patients, were tortuous corkscrew conjunctival vessels, iris atrophy, and cata- ract. Eyes with iris atrophy or depigmentation were 1.8 times more at risk of PSR than eyes without. Overall, PSR was found in 12.9% of subjects examined (3.5% of HbSS, 25.3% of HbSC; 15.9% of males and 11.2% of females). The prevalence of proliferative sickle retinopathy increased with age and increased systemic severity of sickle cell disease; sex did not have an influence. Conclusions. There is a high prevalence of ocular morbidity in sickle cell disease patients at Korle-bu Hospital. Prevalence increased with age, systemic severity of sickle cell disease, and HbSC genotype. Key Words. Epidemiology, Ocular manifestations, Public health, Sickle cell disease, Sickle cell retin- opathy Accepted: October 5, 2010 ORIGINAL ARTICLE chronic hemolytic anemia and vaso-occlusive crises that cause severe pain as well as long-term and widespread organ damage (1, 2). The eyes offer a unique opportunity for direct observation of the vaso-occlusive process in SCD (2, 3). The conjunc- tiva in sickle cell patients shows tortuous “corkscrew” con- junctival vessels, transient saccular dilatation of vessels, and multiple, short, comma-shaped capillary segments seemingly isolated from the vascular network (2). http://crossmark.crossref.org/dialog/?doi=10.5301%2FEJO.2010.5977&domain=pdf&date_stamp=2010-04-11 © 2010 Wichtig Editore - ISSN 1120-6721 485 Osafo-Kwaako et al Ghana, from March 4 to April 3, 2007. The SCC is a unit in the Department of Medicine dedicated to the management of SCD patients. Sickle cell patients reporting for routine follow-up for SCD at the SCC older than 6 years of age were eligible for this study. Following written informed con- sent, 201 consecutive patients were recruited. Clinical and demographic data were obtained by a structured question- naire administered by the first author and using information from the hospital records. Dependent variables included age, sex, and genotype. The estimated number of admis- sions due to sickle cell crisis was used as an indicator for severity of systemic disease. Vision with the patients’ own correction or pinhole-corrected visual acuity was assessed with a Snellen chart and E chart. Ocular alignment, ex- traocular motility, and pupillary light reflex were assessed for possible neurologic deficits as well as pupillary reac- tions using a pen torch. Anterior segment slit-lamp biomi- croscopy and, in cooperative patients, applanation tonom- etry was carried out. Anterior segment examination was performed to assess the presence or absence of signs due to SCD in the conjunctiva, cornea, anterior chamber, pupil, iris, lens, and anterior vitreous. Afterwards, dilated exami- nation of the ocular fundus was done using indirect binoc- ular ophthalmoscopy. Fundus findings were documented with standard retinal drawings. Statistical analysis was carried out using the Statistical Package for Social Sciences (SPSS) for Microsoft Win- dows, using Fisher exact test and chi-square test. RESULTS A total of 201 subjects were included in the study. The age, sex, and genotype distribution of subjects is shown in Table I. All patients had normal vision with their own correction or with pinhole in the better eye. In patients with HbSS geno- type, 2/228 (0.9%) eyes had abnormal visual acuity, com- pared to 8/174 (4.6%) HbSC eyes. The difference in the number of visually impaired eyes in the 2 genotypes was statistically significant (p=0.02, cross tab, Fisher exact test), with more visually impaired eyes in the HbSC genotype (Tab. II). The reasons for vision loss in patients with visual impai- rment were cataract, vitreous hemorrhage, optic atrophy, phthisis bulbi, central retinal artery occlusion, and sea-fan neovascularization. Some eyes had more than one reason for vision loss. Intraocular pressure was below 21 mmHg The SS genotype form exists in 0.4% of the black popula- tion and causes severe systemic but mild ocular disease, while the less common SC and Sβ-thal genotypes cause mild systemic but severe ocular disease (4). The prevalence of SC and Sβ-thal varies between different black popula- tions. West Africa has the highest prevalence of HbSC in Africa (2). An estimated 1%-2% of Ghana’s approximately 25 million people have SCD, with a relatively high preva- lence of the hemoglobin C gene (1% of the Ghanaian pop- ulation) (2). However, the prevalence and pattern of ocular morbidity due to sickle cell disease in Ghana is unknown. In order to plan ocular screening programs, data on the influence of age, sex, and genotype on ocular morbidity in Ghanaian SCD patients are needed. Such information is currently unavailable, though workers in neighboring Togo found an increased prevalence of proliferative sickle retin- opathy (PSR) with increasing age (5). Proliferative sickle retinopathy has also been found to be more common in males, to increase with age in both genotypes, and to have a tendency of being bilateral (6). It would be useful to know the risk of PSR in Ghanaian SCD patients with abnormal anterior segment signs, since anterior segment examination is technically easier to per- form than funduscopy and therefore easier to teach to personnel who might be recruited for an ocular screening program in a resource-poor setting. Patients with PSR in Ghana usually present late to the oph- thalmologist with vitreous hemorrhage or retinal detachment. This, coupled with the unavailability of a specialized vitreo- retinal surgery unit in Ghana, makes the early detection of PSR very important. Our study aims at determining the mag- nitude and pattern of ocular manifestations in SCD patients at Korle-bu Hospital in Accra, Ghana. It is thus important in several respects. It provides data needed for the develop- ment of an appropriate protocol for screening and manage- ment of PSR, encouraging a multidisciplinary approach in the management of SCD. Involving ophthalmologists in the care of SCD patients has become increasingly important as the expansion of appropriate health care services worldwide enables most sickle cell patients to live into adulthood, al- lowing their ocular changes to become manifest. MATERIALS AND METHODS This hospital-based cross-sectional study was carried out at the Sickle Cell Clinic (SCC), Korle-bu Hospital, Accra, © 2010 Wichtig Editore - ISSN 1120-6721 486 Sickle-cell disease in Ghana: ocular manifestations in all but 6 subjects (3 HbSS and 3HbSC). The clinical fin- dings sorted by genotype are shown in Table III. Tortuous, corkscrew conjunctival vessels, iris atrophy, and cataract were significantly more common in the HbSC genotype (p=0.018, 0.002, and 0.003 respectively). Jaundice was more common in the HbSS genotype (p=0.005). The ave- rage age of patients with cataract was 39.9 years (SD 15.4) (range 32- 65). The black sunburst sign and sea-fan neo- vascularization were significantly more common in genot- ype HbSC (p<0.001). Sea-fan neovascularization was seen TABLE I - AGE, SEX, AND GENOTYPE DISTRIBUTION OF SUBJECTS Age range, years Sex Genotype Male Female HbSS HbSC No. (%) No. (%) No. (%) No. (%) 6-15 16-25 26-35 36-45 46-55 56-65 34 13 14 2 3 1 (16.9) (6.3) (7.0) (1.0) (1.5) (0.5) 34 40 21 17 16 6 (16.9) (19.9) (10.4) (8.5) (8.1) (3.5) 51 33 21 5 3 1 (25.4) (16.4) (10.4) (2.5) (1.5) (0.5) 17 20 14 16 16 6 (8.5) (10.0) (7.0) (8.0) (8.0) (3.0) Total 67 (33.2) 134 (67.3) 114 (56.7) 87 (44.5) TABLE II - �VISION with the PATIENTS’ own correc- tion OR WITH PINHOLE (N=402 EYES) Visual acuity SS n (%) SC n (%) Normal vision 225 (99.1) 165 (95.3) Visual impairment 0 5 (2.9) Severe visual impairment 0 1 (0.6) Blind eyes 2 (0.9) 2 (1.2) Excluded eyes 1 1 Total 228 174 TABLE III - CLINICAL SIGNS VERSUS GENOTYPE (N=201 SUBJECTS: 114 SS, 87 SC) Anterior segment signs SS (n=114) n (%) SC (n=87) n (%) Total (n=201) n (%) p value Conjunctiva   Tortuous corkscrew conjunctival vessels 53 (46.5) 55 (63.2) 108 (53.7) 0.018*   Saccular dilatations of vessels 45 (39.5) 42 (48.3) 87 (43.3) 0.212   Seemingly isolated cap segments 42 (36.8) 39 (44.8) 81 (40.3) 0.253   Injection 0 2 (2.3) 2 (1.0) —   Jaundice 10 (8.8) 0 10 (5.0) 0.005* Iris and lens signs   Iris depigmentation 40 (35.1) 39 (44.8) 79 (39.3) 0.161   Iris atrophy 4 (3.5) 14 (16.1) 18 (9.0) 0.002*   Rubeosis 0 0 0 —   Cataract 3 (2.6) 12 (13.8) 15 (7.5) 0.003* Nonproliferative fundus signs   Increased tortuosity of major retinal vessels 7 (6.1) 1 (1.1) 8 (4.0) 0.072   Pale peripheral retina 24 (21.1) 27 (31.0) 51 (25.4) 0.112   Salmon patch hemorrhage 4 (3.5) 4 (4.6) 8 (4.0) 0.702   Black sunburst sign 9 (7.9) 24 (27.6) 33 (16.4) <0.001*   Angioid streaks 6 (5.3) 6 (6.9) 12 (6.0) 0.636   Central retinal vein occlusion 0 1 (1.1) 1 (0.5) —   Central retinal artery occlusion 0 0 0 — Proliferative fundus signs   Sea-fan neovascularization 4 (3.5) 22 (25.3) 26 (12.9) <0.001*   Vitreous hemorrhage 2 (1.8) 3 (3.4) 5 (2.5) 0.449   Retinal detachment 1 (0.9) 1 (1.1) 2 (1.0) 0.851 *Significant. © 2010 Wichtig Editore - ISSN 1120-6721 487 Osafo-Kwaako et al in 26 subjects (12.9%), 22 HbSC and 4 HbSS (p<0.001); 15.9% were male and 11.2% female (p=0.392). It increa- sed with age, occurring in 1 (aged 10 years, male, HbSS) out of 68 children, compared to 25/133 adults (p=0.001) with 2/7 in the age group 56-65 years affected. Vitreous hemorrhage occurred in 5 subjects with SN 2/233 HbSS and 3/169 HbSC, 1/67 male and 4/134 female (p=0.480), with prevalence increasing with advanced age. Retinal de- tachment was found only in 2 female patients. Category- based prevalence of PSR in different categories of sickle cell patients is shown in Table IV. Other findings were peripheral retinal pallor and salmon patch hemorrhages. Although 50% of children had pale retinas, which may have been due to anemia, only 1.6% had PSR. The black sunburst sign, found in 16.4% of sub- jects, was significantly more common in HbSC (27.6%) than HbSS (7.9%) subjects (<0.001) and increased with age. The prevalence of PSR increased with increasing total number of admissions to the hospital, for both genotypes, as shown in Figure 1A. The prevalence of PSR first in- creases, and then decreases, with increasing admissions per year of life in HbSC patients, but increases with in- creasing admissions per year of life in HbSS patients, as shown in Figure 1B. These results suggest that the number of admissions to the hospital might not be of predictive Fig. 1 - The relationship between the percentage of proliferative sickle retinopathy and (A) the total number of admissions to hos- pital and (B) the number of hospital admissions per year of life. An increasing frequency of admission to hospital in HbSS patients cor- relates with a higher rate of proliferative sickle retinopathy. PSR = proliferative sickle retinopathy. TABLE IV - CATEGORY-BASED PREVALENCE OF PSR IN DIFFERENT CATEGORIES OF SICKLE CELL PATIENTS Subgroup selected for analysis Percentage PSR Total subjects (n=201) 12.9 SS subjects only (n=114) 3.5 SC subjects only (n=87) 25.3 Children only (n=68) 1.5 Adults only (n=133) 18.8 SS adults only (n=63) 4.8 SC adults only (n=70) 31.4 SS adults only, with admission (n=54) 5.6 SS adults only, with no admission (n=9) 0 SS adults only, with admissions per year >0.3 (n=16) 12.5 SS adults only, with admissions per year >0.6 (n=8) 25.0 SS adults only, with admissions per year ≤0.6 (n=55) 3.6 PSR = proliferative sickle retinopathy. A B SC SS SC SS © 2010 Wichtig Editore - ISSN 1120-6721 488 Sickle-cell disease in Ghana: ocular manifestations lence of PSR in Kenya was 1%, Saudi Arabia 1.6%, Nigeria 5.6%, Togo 9.4%, Jamaica 24%, and United States 18% (5, 11-14). Our 12.9% prevalence of PSR (sea-fan neovascular- ization with or without vitreous hemorrhage and retinal detach- ment) is comparable to Nigeria and Togo, while life expectancy of HbSS and HbSC populations in these countries is similar. In Kenya and Saudi Arabia, the HbSS population dominates, while Jamaican and American SCD patients survive longer. PSR was significantly more common in HbSC (25.3%) than HbSS (3.5%) subjects (p<0.001), with a male:female ratio of 15.9%:11.2% (p=0.392), and increased with age (p=0.001). However, Fox et al found a statistically significant higher prev- alence of PSR in males (6). Only 1.5% of children <15 years had SN compared to 28.6% of patients 56-65 years of age. Our SN prevalence in the other age groups was similar to the findings of Balo et al in Togo (5). This was not surprising, giv- en that the 2 countries share similar populations, genotypes, and life expectancy. With regard to vitreous hemorrhage, out of the 5 patients concerned (2.5%), 2 (1.8%) were HbSS and 3 (3.4%) HbSC subjects. For HbSS subjects, our prevalence is higher than the one in Kenyan subjects (0%), but less than the 3.8% found in Nigerian HbSS subjects (3, 10). Our 3.4% prevalence in HbSC patients is much lower than the 10%, 8%, and 18% found by authors in Senegal, Togo, and Curacao (5, 7, 15). Retinal detachment was seen in 2 (1%) subjects, more than the 0% found in Kenya and Nigeria but lower than the 8% prevalence in Curacao, which might be explained by the higher PSR prevalence there (3, 4). Statistically, our findings on the relationship between iris at- rophy/depigmentation and the development of PSR are in- conclusive, but we noted that the eye with the severest atro- phy also had the most florid SN. This may indicate that the increased risk of having PSR might be related to the severity of iris atrophy rather than just its presence, as Acheson et al (16) found that iris atrophy was closely associated with PSR in the same eye. For the patients with HbSC, admission history seems not to be predictive of PSR based on the distribution pattern in Figure 1B. However, for patients with HbSS, there was a positive correlation between PSR and the systemic se- verity of sickle cell based on admission to hospital history. Our study shows that the risk of development of PSR increased with the total number of admissions to hospital (an index of systemic severity of SCD) for both genotypes. This number is in turn related to age; older patients are likely to have more ad- missions, but their increased prevalence of PSR may be more related to age than to the number of admissions. However, the power of this correlation between systemic severity and PSR value in HbSC patients, while an increasing frequency of admission to the hospital in HbSS patients correlates with a higher rate of PSR. Eyes with iris atrophy or depigmenta- tion were 1.8 times more at risk of PSR than eyes without (odds ratio 1.8). DISCUSSION In this study, the number of subjects enrolled declined with in- creasing age, with 68 subjects (33.8%) in the age group 6-15 years and only 7 subjects (3.5%) in the age group 56-65 years. This may reflect the lower life expectancy of sickle cell patients when compared with the general population in Ghana (male 59.5 years, female 60 years). However, other factors such as reduced mobility of older patients might also contribute to these findings. Our study supports other authors’ reports that bilateral visual loss from SCD is rare (7). We found visual impairment in only 2.5% of eyes at presentation, with all patients having normal corrected vision in the better eye. While unilateral loss of vi- sion from SCD is not of serious public health significance, it is significant for the individual patient who lives in fear of los- ing vision in the better eye. Visual loss was more frequent but less severe among HbSC patients (proliferative and nonprolif- erative retinal changes) than HbSS patients (anterior segment ischemia and central retinal artery occlusion/central retinal vein thrombosis). In our study, only 63% of patients showed conjunctival signs, which is a low rate compared to data from Kenya (87%) and Nigeria (77%-81%) (2, 8). Cataract and iris atrophy or depig- mentation were 3 times more common among HbSC than HbSS subjects. In sickle cell patients, cataracts occurred at a younger age (mean 39.9 years, range 32-65) than in the gen- eral population. We found a well-regulated intraocular pressure in almost all patients, without any glaucomatous optic disc changes. In accordance with other studies, increased tortuos- ity of major conjunctival vessels was the most common sign (9, 10). It was significantly more prevalent in HbSS subjects (p=0.018) and in females, the youngest being 12 years old. Our 4.0% prevalence of salmon patch hemorrhage is less than the 6.4% seen in Nigerian patients (3). Similar to the findings of this Nigerian study, the presence of the black sunburst sign was dependent on the genotype (HbSC>HbSS) and increasing age of the patients (3). The prevalence of PSR varies in different countries depending on the life expectancy of SCD patients and genotype. Preva- © 2010 Wichtig Editore - ISSN 1120-6721 489 Osafo-Kwaako et al REFERENCES Serjeant GR. Sickle Cell Disease, 2nd ed. Oxford: Oxford 1. University Press; 1992. Konotey-Ahulu FID. The Sickle Cell Disease Patient. Lon-2. don: Macmillan Education Ltd.; 1991/1992. Obikili AG, Oji EO, Onwukeme KE. Ocular findings in ho-3. mozygous sickle cell disease in Jos, Nigeria. Afr J Med Med Sci 1990; 19: 245-50. van Meurs JC. Ocular findings in sickle cell patients on Cu-4. racao. Int Ophthalmol 1991; 15: 53-9. Balo KP, Segbena K, Mensah A, Mihluedo H, Bechetoille A. 5. Hemoglobinopathies and retinopathies in Lome UHC. J Fr Ophtalmol 1996; 19: 497-504. Fox PD, Dunn DT, Morris JS, Serjeant GR. Risk factors for 6. proliferative sickle retinopathy. Br J Ophthalmol 1990; 74: 172-6. van Meurs JC. Vision-threatening eye manifestations in pa-7. tients with sickle cell disease on Curacao. Ned Tijdschr Ge- neeskd 1990; 134: 1800-2. Abiose A, Lesi FE. Ocular findings in children with homozy-8. gous sickle cell anemia in Nigeria. J Pediatr Ophthalmol Strabismus 1978; 15: 92-5. Majekodunmi SA, Akinyanju OO. Ocular findings in homozy-9. gous sickle cell disease in Nigeria. Can J Ophthalmol 1978; 13: 160-2. Eruchalu UV, Pam VA, Akuse RM. Ocular findings in chil-10. dren with severe clinical symptoms of homozygous sickle cell anaemia in Kaduna, Nigeria. West Afr J Med 2006; 25: 88-91. Al-Salem M, Ismail L. Ocular manifestations of sickle cell 11. anaemia in Arab children. Ann Trop Paediatr 1990; 10: 199- 202. Babalola OE, Wambebe CO. Ocular morbidity from sickle 12. cell disease in a Nigerian cohort. Niger Postgrad Med J 2005; 12: 241-4. Friberg TR, Young CM, Milner PF. Incidence of ocular ab-13. normalities in patients with sickle hemoglobinopathies. Ann Ophthalmol 1986; 18: 150-3. Fox PD, Vessey SJ, Forshaw ML, Serjeant GR. Influence 14. of genotype on the natural history of untreated proliferative sickle retinopathy: an angiographic study. Br J Ophthalmol 1991; 75: 229-31. Ndiaye PA, Ndoye PA, Seye C, et al. Vitreo-retinal complica-15. tions of hemoglobinopathy SC. Dakar Med 1998; 43: 21-4. Acheson RW, Ford SM, Maude GH, Lyness RW, Serjeant 16. GR. Iris atrophy in sickle cell disease. Br J Ophthalmol 1986; 70: 516-21. Kent D, Arya R, Aclimandos WA, Bellingham AJ, Bird AC. 17. Screening for ophthalmic manifestations of sickle cell disea- se in the United Kingdom. Eye 1994; 8(Pt 6): 618-22. is low because of the relatively low number of eyes with PSR in our study. Kent et al (17) found no correlation between the systemic complications of SCD and PSR. The risk of developing PSR in SCD patients was not evenly spread throughout our study population. While the overall prev- alence was 12.9%, it varied with different subgroups. HbSC adults had the highest prevalence of PSR (31.4%). Overall PSR prevalence in HbSS patients was 3.5%, but it was 25.0% in adults with admission per year of life >0.6. In an environment with limited resources, our analysis supports a PSR screening program which should consider HbSS adults with admissions per year >0.6 and all HbSC adults. This pro- tocol would leave out children, who have a PSR prevalence of 1.5%, and HbSS adults with admission per year of life ≤0.6, who have a PSR prevalence of 3.6%. However, all sickle cell patients with frank iris atrophy on anterior segment examina- tion should have a screening funduscopy to rule out PSR. In this study, we examined the eyes of 201 SCD patients with an overall PSR prevalence of 12.9%. Of 70 HbSC adults exam- ined, 22 had PSR. Of 8 HbSS adults with admission per year >0.6 examined, 2 had PSR. By applying the screening protocol as described above, only 78 of 201 patients, i.e., about 40%, would have to be examined and 24 out of the 26 PSR subjects would be diagnosed. Only 2 out of the 26 subjects with PSR (2 HbSS subjects) would not have been diagnosed correctly. However, the risk of blindness for these 2 patients is still very low due to the higher rate of a self-limiting course of disease in HbSS patients. Although vision loss is uncommon in SCD, it is very impor- tant to those who end up with reduced visual acuity. Therefore, preventive measures are important to minimize this risk. This study gives data that support the development of a screening protocol for PSR in Accra with the focus on HbSC adults and HbSS adults with previous admissions due to sickle cell crisis in order to improve the early detection and management of PSR in Ghana. The authors report no proprietary interest or financial support. Address for correspondence: Martin Nentwich, MD Ludwig-Maximilians-University Department of Ophthalmology Klinikum der Universität München Campus Innenstadt Mathildenstrasse 8 80336 Munich, Germany Martin.Nentwich@med.uni-muenchen.de