Key Points
Question
Is there a difference in the rates of retinal thinning among adults with and without sickle cell retinopathy, and what are the ocular and systemic factors associated with retinal thinning?
Findings
In this case-control study of 370 adults with and without sickle cell retinopathy over 20 years, retinal thinning detected through spectral-domain optical coherence tomography occurred at faster rates among participants with sickle cell retinopathy compared with age- and race-matched participants without sickle cell retinopathy. Participant age, stage of retinopathy, history of hypertension, and presence of acute chest syndrome were associated with increased rates of retinal thinning, and receipt of hydroxyurea therapy was associated with decreased rates of retinal thinning.
Meaning
The study’s findings suggest that anatomic worsening of sickle cell maculopathy identified through spectral-domain optical coherence tomography may be a useful parameter to monitor worsening of sickle cell retinopathy.
Abstract
Importance
Determination of retinal thinning rates may help to identify patients who are at risk of progression of sickle cell retinopathy.
Objective
To assess the rates of macular thinning in adults with and without sickle cell retinopathy using spectral-domain optical coherence tomography (OCT) and to identify ocular and systemic risk factors associated with retinal thinning.
Design, Setting, and Participants
This longitudinal prospective case-control study enrolled adult participants from a university-based retina subspecialty clinic between February 11, 2009, and July 3, 2019. The study was designed in autumn 2008 and conducted from February 2, 2009, to July 3, 2020. Participants with sickle cell retinopathy (sickle cell group) were matched by age and race with participants without sickle cell retinopathy (control group). Participants received annual spectral-domain OCT and clinical examinations. Those with at least 1 year of follow-up by July 3, 2020, were included in the analysis. Data were analyzed from February 2, 2009, to July 3, 2020.
Main Outcomes and Measures
The primary outcome was comparison of spectral-domain OCT measurements from early-treatment diabetic retinopathy study subfield rates of retinal thinning between eyes with and without sickle cell retinopathy and between different sickle cell hemoglobin subtypes. The secondary outcome was identification of ocular and systemic risk factors associated with rates of retinal thinning.
Results
Among 370 adults (711 eyes) enrolled in the study, 310 participants (606 eyes) had sickle cell retinopathy, and 60 participants (105 eyes) did not. Of those, 175 of 310 participants (56.5%; 344 of 606 eyes [56.8%]; mean [SD] age, 37.8 [12.8] years; 126 women [72.0%]) in the sickle cell group and 31 of 60 participants (51.7%; 46 of 105 eyes [43.8%]; mean [SD] age, 59 [15.4] years; 22 women [71.0%]) in the control group had at least 1 year of clinical and spectral-domain OCT follow-up data from baseline. The mean (SD) follow-up was 53.7 (32.6) months for the sickle cell group and 54.6 (34.9) months for the control group. Rates of macular thinning in the sickle cell group were significantly higher than those in the control group for the inner nasal (difference, −1.18 μm per year; 95% CI, −1.71 to −0.65 μm per year), inner superior (difference, −1.03 μm per year; 95% CI, −1.78 to −0.29 μm per year), inner temporal (difference, −0.61 μm per year; 95% CI, −1.16 to −0.07 μm per year), and outer nasal (difference, −0.41 μm per year; 95% CI, −0.80 to −0.03 μm per year) quadrants. Patients with sickle cell hemoglobin SC and sickle cell hemoglobin β-thalassemia subtypes had higher rates of retinal thinning than those with the sickle cell hemoglobin SS subtype. Risk factors associated with greater rates of retinal thinning included participant age, stage of retinopathy, previous stroke, and presence of hypertension, acute chest syndrome, or diabetes. Hydroxyurea therapy was associated with decreased rates of retinal thinning and may be a protective factor.
Conclusions and Relevance
In this study, rates of retinal thinning were higher among participants with sickle cell retinopathy compared with those without sickle cell retinopathy, and thinning rates increased with participant age and stage of retinopathy. These findings suggest that identifying anatomic worsening of sickle cell maculopathy through spectral-domain OCT may be a useful parameter to evaluate the progression of sickle cell retinopathy.
This case-control study assesses the rates of retinal thinning in adults with and without sickle cell retinopathy using spectral-domain optical coherence tomography and examines ocular and systemic risk factors associated with retinal thinning.
Introduction
Sickle cell retinopathy is characterized by occlusion of the retinal vasculature, which may produce ischemia and infarction of the retina. Several studies have reported that retinal macular thinning occurs in sickle cell retinopathy and is associated with age, sickle cell hemoglobin subtype, and proliferative sickle retinopathy (PSR) stage measured by the Goldberg classification system.1,2,3,4 One observational study using spectral-domain optical coherence tomography (OCT) found that Early-Treatment Diabetic Retinopathy Study (ETDRS) subfield thickness measurements for the central and inner subfields were lower among individuals with sickle cell retinopathy compared with age- and race-matched individuals without sickle cell retinopathy and that thinning was associated with age and PSR stage.2 Retinal thinning was most severe in patients with the sickle cell hemoglobin SS (HbSS) subtype compared with patients with the sickle cell hemoglobin SC (HbSC) and sickle cell hemoglobin β-thalassemia (HbS-β-thalassemia) subtypes. Using microperimetry, another study found that areas of retinal thinning had decreased retinal sensitivity and thus decreased retinal function.4 The results of a spectral-domain OCT study of healthy eyes indicated that retinal thinning was associated with aging and that most thinning was associated with retinal nerve fiber layer (RNFL) thinning.5 A longitudinal study using spectral-domain OCT imaging reported higher rates of RNFL thinning in eyes with sickle cell retinopathy compared with eyes without sickle cell retinopathy.6,7 A 2018 study of 38 eyes with sickle cell retinopathy (24 patients) and 30 eyes without sickle cell retinopathy (30 patients), with a mean follow-up of 25.5 months, reported that eyes with sickle cell retinopathy had greater rates of thinning than those without sickle cell retinopathy.8 The purpose of the present study was to calculate and compare rates of retinal thinning among eyes with and without sickle cell retinopathy and to identify associated ocular and systemic risk factors. We hypothesized that eyes with sickle cell retinopathy would have greater rates of thinning than eyes without sickle cell retinopathy because of ischemia associated with sickling episodes.
Methods
Participants were prospectively enrolled in this longitudinal observational case-control study from a university-based retina subspecialty clinic between February 11, 2009, and July 3, 2019. The study was designed in autumn 2008 and conducted from February 2, 2009, to July 3, 2020. The institutional review board of the University of Illinois at Chicago approved the study, and written informed consent was obtained from all participants before enrollment. No incentive or compensation was given to participants. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for case-control studies.
Participants with sickle cell retinopathy (sickle cell group) were matched by age and race with participants without sickle cell retinopathy (control group). Individuals were eligible for inclusion in the control group if they had no ocular or retinal disease other than mild refractive error or mild cataracts that did not impact best-corrected visual acuity. Participants in the sickle cell and control groups received full ophthalmic examinations, which consisted of best-corrected Snellen visual acuity testing, slitlamp examination, and dilated ophthalmoscopy. Patients in the sickle cell group were staged using the Goldberg classification system for PSR (range, 1-5, with stage 1 indicating peripheral arterial occlusion and stage 5 indicating tractional retinal detachment). The presence and extent of vascular tortuosity, arteriovenous anastomosis, and vascular stenosis were graded on a scale of 1 to 3, with 1 indicating mild and 3 indicating severe conditions. The retina was carefully examined at each visit for the presence or absence of a foveal depression sign, arterial or venous occlusion, arteriovenous anastomosis, pigmentary abnormality, sunburst lesion, iridescent spot, retinal hemorrhage, salmon patch hemorrhage, peripheral retinal neovascularization (seafan), vitreous hemorrhage, and tractional or rhegmatogenous retinal detachment.
Eyes were evaluated using standard and enhanced-depth spectral-domain OCT imaging (Spectralis; Heidelberg Engineering) using 6-mm scans if the media permitted good-quality images. Spectral-domain OCT images were repeated if artifacts were seen; segmentation boundaries were corrected as needed. The ETDRS measurements for the central, inner (nasal, superior, temporal, and inferior), and outer (nasal, superior, temporal, and inferior) subfields were recorded for each eye at each visit. Electronic medical records from several clinics (internal medicine, hematology, cardiology, orthopedics, kidney, pulmonary, and ophthalmology) were reviewed to obtain past and current medical conditions for each participant at baseline. Patient records were reviewed for the presence of sickle cell disease–associated medical conditions (acute chest syndrome, avascular necrosis of any joint, cholelithiasis, hypertension, liver disease, pulmonary embolism, deep venous thrombosis, pulmonary hypertension, kidney disease, seizure disorder, stroke, and transient ischemic attack) and surgical procedures (cholecystectomy, joint replacement, or splenectomy). In addition, diagnoses of major systemic diseases (hypertension, cardiac disease, cancer, and infectious or inflammatory conditions) and previous surgeries were recorded. At follow-up visits, interim medical history was obtained (electronic medical record review and patient history).
Statistical Analysis
ETDRS subfield measurements were recorded at each follow-up spectral-domain OCT scan. The change in subfield measurements from baseline over time was assessed using mixed-effects linear models to determine the rate of change of retinal thickness, with negative rates indicating thinning over time. Changes in the rates of retinal thickness (measured in μm per year) were compared between eyes with and without sickle cell retinopathy, and rate differences with 95% CIs were reported. Rates of thinning between hemoglobin subtypes were compared using mixed-effects linear models that included follow-up time (in years), hemoglobin subtype, and their interaction, while controlling for baseline spectral-domain OCT measurements. In addition, rates of thinning were compared between each hemoglobin subtype vs the control group.
To identify risk factors associated with rates of retinal thinning in eyes with sickle cell retinopathy, mixed-effects linear models that included follow-up year, one of the risk factors, and their interaction were used. A significant interaction was defined as a statistically significant association between the risk factor and the rate of thinning. These analyses were planned at the start of the longitudinal study. Patients with at least 1 year of follow-up by July 3, 2020, were included in the analysis. Rates of thinning were calculated from data obtained throughout the study and from follow-up appointments, which were, in general, conducted yearly. The statistical significance threshold was 2-tailed P < .05 (without adjustment for multiple tests). Data were analyzed from February 2, 2009, to July 3, 2020.
Results
Among 370 adults (711 eyes) enrolled in the study, 310 participants (606 eyes) had sickle cell retinopathy, and 60 participants (105 eyes) did not. Of those, 175 of 310 participants (56.5%; 344 of 606 eyes [56.8%]; mean [SD] age, 37.8 [12.8] years; 126 women [72.0%]) in the sickle cell group and 31 of 60 participants (51.7%; 46 of 105 eyes [43.8%]; mean [SD] age, 59 [15.4] years; 22 women [71.0%]) in the control group had at least 1 year of clinical and spectral-domain OCT follow-up data from baseline. Sickle cell hemoglobin subtypes included HbSS (219 of 344 eyes [63.7%]; 111 of 175 patients [63.4%]), HbSC (92 of 344 eyes [26.7%]; (47 of 175 patients [26.9%]), and HbS-β-thalassemia (33 of 344 eyes [9.6%]; 17 of 175 patients [9.7%]). The mean (SD) follow-up was 53.7 (32.6) months for the sickle cell group and 54.6 (34.9) months for the control group. Other demographic data were similar between participants in the sickle cell and control groups.
In total, 30 of 31 participants (96.8%) in the control group and 138 of 175 participants (78.9%) in the sickle cell group had never smoked cigarettes. The incidence of current smoking in the sickle cell group was 26 of 175 participants (14.9%; 15 of 111 patients [13.5%] with the HbSS subtype, 9 of 47 patients [19.1%] with the HbSC subtype, and 2 of 17 patients [11.8%] with the HbS-β-thalassemia subtype) and in the control group was 1 of 31 participants (3.2%). Patients with different hemoglobin subtypes differed in their receipt of hydroxyurea therapy. A total of 18 of 111 patients (16.2%) with the HbSS subtype, 6 of 47 patients (12.8%) with the HbSC subtype, and 3 of 17 patients (17.6%) with the HbS-β-thalassemia subtype were currently receiving hydroxyurea therapy; 54 of 111 patients (48.6%) with the HbSS subtype, 5 of 47 patients (10.6%) with the HbSC subtype, and 4 of 17 patients (23.5%) with the HbS-β-thalassemia subtype had previously received hydroxyurea therapy. Most patients with the HbSC subtype (36 of 47 patients [76.6%]) had never received hydroxyurea therapy compared with those with the HbSS subtype (39 of 111 patients [35.1%] never received hydroxyurea) and the HbS-β-thalassemia subtype (10 of 17 patients [58.8%] never received hydroxyurea).
Over time, among patients in the sickle cell group, retinopathy stage remained stable in 115 of 117 eyes (98.3%) at 1 year, 73 of 74 eyes (98.6%) at 2 years, and 34 of 39 eyes (87.2%) at 5 years. Visual acuity decreased at a mean (SD) rate of 0.07 (0.10) lines per year for the sickle cell group and increased at a mean (SD) rate of 0.02 (0.03) lines per year for the control group (P < .001). The rates of thinning in ETDRS subfields were greater in the sickle cell group than in the control group for the inner nasal (difference, −1.18 μm per year; 95% CI, −1.71 to −0.65 μm per year), inner superior (difference, −1.03 μm per year; 95% CI, −1.78 to −0.29 μm per year), inner temporal (difference, −0.61 μm per year; 95% CI, −1.16 to −0.07 μm per year), and outer nasal (difference, −0.41 μm per year; 95% CI, −0.80 to −0.03 μm per year) quadrants (Table 1).
Table 1. Comparison of Rates of Change in Retinal Thickness Between Participants With and Without Sickle Cell Retinopathy Measured by Spectral-Domain Optical Coherence Tomographya.
| ETDRS subfield | Rate of retinal thickness change in eyes without sickle cell retinopathy, μm/y | Difference in rate of retinal thickness change in eyes with vs without sickle cell retinopathy (95% CI), μm/y | P value |
|---|---|---|---|
| Central | 0.29 | −0.47 (−1.04 to 0.10) | .10 |
| Inner | |||
| Nasal | 0.26 | −1.18 (−1.71 to −0.65) | <.001 |
| Superior | −0.35 | −1.03 (−1.78 to −0.29) | .006 |
| Temporal | −0.32 | −0.61 (−1.16 to −0.07) | .03 |
| Inferior | −0.48 | −0.39 (−0.85 to 0.07) | .10 |
| Outer | |||
| Nasal | −0.41 | −0.41 (−0.80 to −0.03) | .04 |
| Superior | 0.10 | −0.35 (−1.45 to 0.75) | .54 |
| Temporal | −0.51 | −0.49 (−1.34 to 0.37) | .27 |
| Inferior | −0.88 | −0.09 (−0.64 to 0.47) | .75 |
Abbreviation: ETDRS, Early-Treatment Diabetic Retinopathy Study.
Because a negative rate indicates retinal thinning, a negative difference between eyes with and without sickle cell retinopathy indicates faster thinning.
A comparison of hemoglobin subtypes indicated that rates of retinal thinning differed for the central (χ2 = 8.78; P = .01), inner nasal (χ2 = 7.25; P = .03), and inner temporal (χ2 = 7.55; P = .02) quadrants but not for other quadrants (Table 2). Patients with sickle cell hemoglobin SC and sickle cell hemoglobin β-thalassemia subtypes had higher rates of retinal thinning than those with the sickle cell hemoglobin SS subtype. Compared with the eyes of participants in the control group, those of patients with the HbS-β-thalassemia subtype indicated faster thinning in the inner nasal (difference, −1.90 μm per year; 95% CI, −3.10 to −0.70 μm per year), inner temporal (difference, −1.36 μm per year; 95% CI, −2.32 to −0.39 μm per year), inner inferior (difference, −0.87 μm per year; 95% CI, −1.65 to −0.09 μm per year), outer nasal (difference, −0.79 μm per year; 95% CI, −1.45 to −0.13 μm per year), and outer superior (difference, −1.23 μm per year; 95% CI, −2.27 to −0.21 μm per year) subfields. The eyes of patients with the HbSS subtype indicated faster thinning rates in the inner nasal (difference, −1.03 μm per year; 95% CI, −1.64 to −0.43 μm per year), and outer nasal (difference, −0.37 μm per year; 95% CI, −0.72 to −0.03 μm per year) subfields compared with the eyes of participants in the control group. In addition, patients with the HbSC subtype differed from those in the control group with regard to thinning rates in the inner nasal (difference, −1.25 μm per year; 95% CI, −1.91 to −0.59 μm per year), inner superior (difference, −1.39 μm per year; 95% CI, −2.42 to −0.36 μm per year), and inner temporal (difference, −0.80 μm per year; 95% CI, −1.42 to −0.18 μm per year) subfields.
Table 2. Comparison of Rates of Change in Retinal Thickness Among Patients With Sickle Cell Retinopathy With Different Hemoglobin Types, as Measured by Spectral-Domain Optical Coherence Tomography.
| ETDRS subfield | Sickle cell hemoglobin subtype | Results of statistical tests comparing rates among the 3 hemoglobin subtypes | |||||||
|---|---|---|---|---|---|---|---|---|---|
| HbSS (219 eyes) | HbSC (92 eyes) | HbS-β-thalassemia (33 eyes) | |||||||
| Difference (95% CI) | P value | Difference (95% CI) | P value | Difference (95% CI) | P value | χ2 | df | P value | |
| Central | −0.31 (−0.96 to 0.34) | .35 | −0.52 (−1.25 to 0.22) | .17 | −1.21 (−2.45 to 0.03) | .05 | 8.78 | 2 | .01 |
| Inner | |||||||||
| Nasal | −1.03 (−1.64 to −0.43) | .001 | −1.25 (−1.91 to −0.59) | <.001 | −1.90 (−3.10 to −0.70) | .002 | 7.25 | 2 | .03 |
| Superior | −0.81 (−1.62 to 0.00) | .05 | −1.39 (−2.42 to −0.36) | .01 | −1.56 (−3.25 to 0.14) | .07 | 5.24 | 2 | .07 |
| Temporal | −0.39 (−0.93 to 0.15) | .16 | −0.80 (−1.42 to −0.18) | .01 | −1.36 (−2.32 to −0.39) | .006 | 7.55 | 2 | .02 |
| Inferior | −0.34 (−0.84 to 0.16) | .18 | −0.33 (−0.82 to 0.15) | .18 | −0.87 (−1.65 to −0.09) | .03 | 2.59 | 2 | .27 |
| Outer | |||||||||
| Nasal | −0.37 (−0.72 to −0.03) | .03 | −0.40 (−0.90 to 0.10) | .11 | −0.79 (−1.45 to −0.13) | .02 | 2.28 | 2 | .32 |
| Superior | −0.19 (−1.09 to 0.71) | .68 | −0.45 (−1.87 to 0.97) | .53 | −1.23 (−2.27 to −0.21) | .02 | 1.78 | 2 | .41 |
| Temporal | −0.43 (−1.45 to 0.59) | .41 | −0.45 (−1.18 to 0.28) | .22 | −1.01 (−2.19 to 0.16) | .09 | 0.81 | 2 | .67 |
| Inferior | −0.08 (−0.57 to 0.42) | .76 | −0.05 (−0.81 to 0.72) | .91 | −0.40 (−1.31 to 0.52) | .40 | 0.72 | 2 | .70 |
Abbreviations: ETDRS, Early-Treatment Diabetic Retinopathy Study; HbS-β-thalassemia, hemoglobin β-thalassemia: HbSC, hemoglobin C; HbSS, hemoglobin SS.
Results from a secondary analysis of the association of ocular and systemic risk factors with the rates of retinal thinning in patients with sickle cell retinopathy are shown in Table 3. Participant age was associated with rates of thinning for all subfields, with the exception of the outer superior subfield. Stage of PSR was associated with greater thinning in the inner temporal, outer nasal, and outer inferior quadrants. Associations between rates of retinal thinning and hypertension, acute chest syndrome, diabetes, and stroke were observed. Among patients with a history of acute chest syndrome, rates of retinal thinning were greater for all subfields, with the exception of the outer temporal quadrant. Hypertension was associated with greater rates of thinning in the outer nasal and outer inferior quadrants. Diabetes was associated with increased rates of thinning in the central, inner temporal, inner inferior, outer nasal, and outer inferior subfields. A history of stroke was associated with greater rates of thinning in all subfields, with the exception of the outer temporal subfield. In contrast, receipt of hydroxyurea therapy was associated with lower rates of retinal thinning in the central and inner temporal quadrants.
Table 3. Association Between Systemic Risk Factors and Retinal Thinninga.
| Interaction between risk factor and follow-up y | ETDRS subfield | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Central | Nasal | Superior | Temporal | Inferior | ||||||
| Coefficient (95% CI) | P value | Coefficient (95% CI) | P value | Coefficient (95% CI) | P value | Coefficient (95% CI) | P value | Coefficient (95% CI) | P value | |
| Inner | ||||||||||
| Hypertension | −1.06 (−0.47 to −0.65) | <.001 | −1.08 (−1.46 to −0.70) | <.001 | −0.85 (−1.37 to −0.34) | .001 | −1.03 (−1.50 to −0.55) | <.001 | −0.71 (−1.11 to −0.31) | <.001 |
| Acute chest syndrome | 0.18 (−0.19 to 0.55) | .39 | −0.09 (−0.44 to 0.25) | .59 | −0.33 (−0.79 to 0.13) | .16 | 0.08 (−0.34 to 0.51) | .70 | −0.34 (−0.69 to 0.02) | .07 |
| Diabetes | −0.78 (−1.42 to −0.15) | .02 | −0.39 (−0.98 to 0.21) | .20 | −0.77 (−1.57 to 0.02) | .06 | −1.25 (−1.99 to −0.51) | .001 | −0.81 (−1.42 to −0.19) | .01 |
| Stroke | 0.15 (−0.29 to 0.60) | .50 | −0.52 (−0.94 to −0.11) | .01 | 0.01 (−0.55 to 0.57) | .97 | −0.05 (−0.57 to 0.47) | .86 | −0.17 (−0.60 to 0.27) | .45 |
| Hydroxyurea status | 0.56 (.16 to 0.97) | .006 | 0.06 (−0.32 to 0.44) | .76 | 0.34 (−0.17 to 0.85) | .19 | 0.76 (0.29 to 1.23) | .002 | 0.14 (−0.25 to 0.54) | .48 |
| Age | −0.03 (−0.05 to −0.02) | <.001 | −0.03 (−0.04 to −0.01) | <.001 | −0.02 (−0.04 to −0.01) | .02 | −0.03 (−0.04 to −0.01) | .002 | −0.02 (−0.04 − 0.01) | .001 |
| Retinopathy stage | −0.01 (−0.29 to 0.27) | .95 | 0.18 (−0.09 to 0.44) | .19 | 0.10 (−0.25 to 0.45) | .58 | −0.33 (−0.66 to −0.01) | .04 | −0.06 (−0.33 to 0.21) | .65 |
| Outer | ||||||||||
| Hypertension | NA | NA | −0.70 (−1.04 to −0.36) | <.001 | −1.44 (−2.42 to −0.46) | .004 | −0.39 (−1.17 to 0.39) | .32 | −0.58 (−1.03 to −0.14) | .01 |
| Acute chest syndrome | NA | NA | −0.41 (−0.72 to −0.11) | .007 | −0.12 (−1.01 to 0.76) | .78 | −0.03 (−0.72 to 0.67) | .94 | −0.42 (−0.81 to −0.02) | .04 |
| Diabetes | NA | NA | −0.73 (−1.26 to −0.22) | .006 | −1.30 (−2.82 to 0.21) | .09 | −1.14 (−2.35 to 0.06) | .06 | −1.28 (−1.96 to −0.60) | <.001 |
| Stroke | NA | NA | −0.46 (−0.82 to −0.09) | .01 | −0.60 (1.66 to 0.47) | .27 | −0.10 (−0.95 to 0.74) | .81 | −0.31 (−0.79 to 0.17) | .21 |
| Hydroxyurea status | NA | NA | −0.20 (−0.53 to 0.14) | .25 | −0.16 (−1.13 to 0.82) | .75 | 0.30 (−0.48 to 1.07) | .45 | −0.09 (−0.53 to 0.35) | .68 |
| Age | NA | NA | −0.02 (−0.03 to −0.01) | .001 | −0.03 (−0.06 to 0.01) | .11 | −0.04 (−0.07 to −0.02) | .002 | −0.03 (−0.05 to −0.01) | <.001 |
| Retinopathy stage | NA | NA | 0.37 (0.15 to 0.60) | .001 | 0.18 (−0.49 to 0.85) | .60 | −0.38 (−0.89 to 0.13) | .14 | 0.51 (0.19 to 0.83) | .002 |
Abbreviations: ETDRS, Early-Treatment Diabetic Retinopathy Study; NA, not applicable.
The association between a systemic risk factor and retinal thinning was assessed using a mixed-effects linear model that included follow-up year, risk factor, and their interaction. A statistically significant interaction indicates that the rate of change was associated with the risk factor of interest. A negative coefficient for an interaction term indicates that the risk factor was associated with faster thinning.
Compared with eyes without PSR stage progression, the rates of thinning in eyes with PSR stage progression were greater in the inner temporal quadrant (difference, −0.81 μm per year; 95% CI, −1.32 to −0.30 μm per year) but lower in the outer nasal quadrant (difference, 0.79 μm per year; 95% CI, 0.43 to 1.15 μm per year) (Table 4). An analysis of baseline thickness by quartile showed a positive association with rates of thinning; eyes with greater baseline retinal thickness had greater rates of thinning. When controlling for the baseline retinal thickness, the association with progression of stage remained for all quadrants in patients with the HbS-β-thalassemia subtype but for only 50% of quadrants in patients with the HbSS and HbSC subtypes.
Table 4. Comparison of Rates of Change in Retinal Thickness Between Patients With Stage-Progressed vs Non–Stage-Progressed Sickle Cell Retinopathy Measured by Spectral-Domain Optical Coherence Tomography.
| ETDRS subfield | Rate of change in eyes with nonprogressed retinopathy, μm/y | Difference in rate of change in eyes with progressed vs nonprogressed retinopathy (95% CI), μm/y | P value |
|---|---|---|---|
| Central | −0.18 | −0.03 (−0.47 to 0.41) | .89 |
| Inner | |||
| Nasal | −0.99 | 0.30 (−0.11 to 0.72) | .15 |
| Superior | −1.39 | 0.01 (−0.54 to 0.56) | .97 |
| Temporal | −0.75 | −0.81 (−1.32 to −0.30) | .002 |
| Inferior | −0.85 | −0.12 (−0.55 to 0.31) | .59 |
| Outer | |||
| Nasal | −1.00 | 0.79 (0.43 to 1.15) | <.001 |
| Superior | −0.25 | −0.02 (−1.07 to 1.04) | .98 |
| Temporal | −0.97 | −0.14 (−0.98 to 0.69) | .73 |
| Inferior | −1.10 | 0.59 (0.10 to 1.07) | .02 |
Abbreviation: ETDRS, Early-Treatment Diabetic Retinopathy Study.
Discussion
Sickle cell stage and visual acuity remained stable for most participants in the sickle cell and control groups. However, despite this stability, spectral-domain OCT measurements revealed that progressive thinning over time was greater among patients with sickle cell retinopathy than among age- and race-matched participants in the control group. Spectral-domain OCT is thus a more sensitive biomarker for the progression of sickle cell retinopathy than clinical staging. Factors associated with greater rates of retinal thinning include HbSC and HbS-β-thalassemia subtypes, age, systemic diseases (hypertension, acute chest syndrome, diabetes, and stroke), progression of sickle cell retinopathy stage, and greater baseline retinal thickness. Only 1 factor, receipt of hydroxyurea therapy, was associated with lower rates of retinal thinning.
A strength of our study was the use of a control group to account for the associations of retinal thinning with age. In a previous OCT study of healthy eyes, retinal thinning was reported to occur at a rate of 0.53 μm per year; most of this thinning was associated with RNFL thinning, which occurred at a rate of 0.44 μm per year.9 In a spectral-domain OCT study of 120 patients aged 18 to 81 years, Demirkaya et al5 found an association between retinal thickness and age. The study reported that peripapillary RNFL, pericentral ganglion cell layer, peripheral inner plexiform layer, and foveal outer segment layer thicknesses decreased substantially with age. Extrapolating their findings to estimate the loss of thickness in retinal layers over 20 years, the authors estimated a cumulative loss of 2.66 μm in the peripapillary RNFL, 2.06 μm in the pericentral ganglion cell layer, 0.92 μm in the peripheral inner plexiform layer, and 1.76 μm in the foveal outer segment layer. In our study, the rates of thinning in eyes with sickle cell retinopathy exceed the rates reported by Demirkaya et al.5
Compared with historical control groups, the extent of central subfield thinning observed among patients in our sickle cell group is much higher. Because participants in our control group were slightly older and therefore had thinner retinas, it is possible that we may have underestimated the difference between eyes with and without sickle cell retinopathy; that is, the inclusion of younger participants in the control group may have produced even greater differences. Our results cannot be explained by racial differences because we included race-matched participants in the control group. This inclusion is important because patients of African descent have thinner central macular thickness and nerve fiber layers.10,11,12
The finding of greater rates of retinal thinning in patients with the HbSC and HbS-β-thalassemia subtypes compared with those with the HbSS subtype is possibly because patients with the HbSS subtype have thinner retinas at baseline and therefore have less retinal tissue to lose (ie, the floor effect). This explanation is supported by the analysis of ETDRS subfield rates of thinning, as measured by baseline retinal thickness quartiles. The highest quartile of retinal thickness did not indicate any difference between the rates of thinning by hemoglobin subgroup, and these rates were greater compared with those of the lowest quartile. Because the eyes of patients with the HbSS subtype have thinner subfield measurements than those with the HbSC or HbS-β-thalassemia subtype, despite similar durations of sickle cell disease, those retinas are less likely to thin over time. The subfield thickness measurements may not be as useful for follow-up of older patients with the HbSS subtype as they are for follow-up of patients with similar ages who have the HbSC or HbS-β-thalassemia subtype. For the HbSC and HbS-β-thalassemia subtypes, a change in thickness observed on spectral-domain OCT suggests progression of disease, which may benefit from intervention that could avert further ischemia and subsequent retinal thinning.
The finding of the highest rate of thinning in patients with the HbSC subtype may reflect more severe continued ocular infarction and ischemia in this subgroup. A previous study reported that patients with the HbSS subtype had the thinnest retinal ETDRS subfield measurements.2 We hypothesize that patients with the HbSS subtype experienced substantial ischemic events at an earlier period of their disease and thus had retinal thinning earlier in life. The association of acute ischemia with subsequent retinal thinning has been reported in patients with sickle cell retinopathy and paracentral acute middle maculopathy who developed macular thinning 2 months after the onset of maculopathy.13 The finding that ischemia is associated with localized thinning in pediatric patients with sickle cell disease supports this finding. The finding that the temporal and superior areas had the fastest rates of thinning also supports the association with ischemia, as the temporal location is a watershed zone.
Our finding of higher rates of thinning in eyes with PSR stage progression is consistent with previous observations of higher rates of severe retinal thinning in more advanced stages of PSR. In a previous study, focal areas of severe retinal thinning occurred in 10 of 16 eyes (62.5%) with PSR stage 3, in 7 of 23 eyes (30.4%) with nonproliferative retinopathy, and in 1 of 16 eyes (6.3%) without sickle cell retinopathy (P = .003).14 Our finding of greater thinning rates in eyes with similar PSR stages may identify those eyes that are at risk of progression. Our study indicated that greater rates of thinning were associated with progression of retinopathy.
Using OCT angiography, several investigators found decreased vascular density in the deep and superficial capillary plexus,15 which are associated with ischemia and infarction of the retina. Furthermore, vascular density was more substantially decreased with attendant retinal thinning. Ong et al16 noted decreased perfusion (ie, flow loss) in pediatric patients with sickle cell retinopathy compared with those without sickle cell retinopathy before the onset of thinning differences identified on spectral-domain OCT. They also found that a higher proportion of patients with the HbSS subtype had thinning on spectral-domain OCT compared with patients with other hemoglobin subtypes.
In this longitudinal study, our finding that hydroxyurea therapy has potentially protective benefits for retinal thinning rates has several implications. Hydroxyurea therapy is associated with increases in fetal hemoglobin, which decreases vaso-occlusive crises and could be associated with decreased retinal ischemia and less retinal thinning over time. Our study suggests that hydroxyurea therapy, which is associated with lower rates of systemic sickling events, may also have consequences for retinal thinning. It is notable that patients with the HbSC subtype had the highest rates of retinal thinning and the lowest rates of hydroxyurea therapy. This difference is not unexpected, as hydroxyurea treatment is not standard care for patients with sickle cell disease. This finding suggests that patients with these hemoglobin subtypes may have benefited from hydroxyurea therapy. Spectral-domain OCT could be used to detect eyes with sickle cell retinopathy that exhibit greater rates of retinal thinning and that may benefit from more frequent monitoring and receipt of hydroxyurea therapy or other therapies to decrease sickling episodes.
Limitations
This study has several limitations. Goldberg staging for sickle cell retinopathy was based on clinical examination without fluorescein angiography. Fluorescein angiography has been reported to allow greater detection of more advanced stages of sickle cell retinopathy, although clinical examination reliably detected stage 3 retinopathy, and widefield angiographic results did not change clinical management.17 Our finding of greater rates of thinning in eyes with stage progression is unlikely to be associated with the lack of angiography. Fluorescein angiography is also not routinely performed for patients with sickle cell disease. Previous research has indicated that macular thickness, especially in the outer temporal area, is negatively associated with hypertension, particularly among patients with high fasting blood glucose levels.18 Although hypertension was a potential confounding variable in our study, it was equally present in both the control and sickle cell groups. We also found changes in all inner, outer nasal, and inferior quadrants in addition to the outer temporal area. Few patients in our sample had diabetes (and all of those with diabetes did not have diabetic retinopathy), which is also a potential confounder in our study.
We did not analyze the association between retinal thinning and disease-modifying therapies, such as chronic exchange transfusions, crizanlizumab therapy, or voxelotor therapy. Patients who received bone marrow transplants were excluded from analysis after their transplants. It is possible that these treatments slowed retinal thinning and that our rates of thinning underestimate the true rates. However, our findings represent the actual retinal thinning rates in patients with sickle cell retinopathy who received medical care. Other than reviewing notes from sickle cell clinics (which included patient responses regarding adherence to hydroxyurea treatment) and blood tests to detect adverse effects associated with hydroxyurea therapy, we could not assess drug adherence. This limitation could produce an underestimation of the protective benefits of hydroxyurea therapy.
Conclusions
This study’s findings indicate that retinal thickness and rates of retinal thinning detected using spectral-domain OCT may be useful biomarkers for the progression of sickle cell disease. The rates of thinning in patients with sickle cell retinopathy exceeded those of age- and race-matched participants without sickle cell retinopathy. These findings suggest that spectral-domain OCT data should be considered in studies investigating the effectiveness of drugs for the treatment of sickle cell disease. Drugs such as p-selectin, which are thought to be associated with decreases in sickling episodes and lower rates of acute chest syndrome, may decrease the incidence of subclinical retinal infarctions.19 When evaluating the benefits of therapies, the different rates of thinning between different hemoglobin subtypes may be important to consider. Because patients with the HbSS subtype have slower rates of retinal thinning over time compared with patients with the HbSC subtype, studies are needed to analyze the results with respect to hemoglobin subtypes when evaluating therapeutic outcomes.
References
- 1.Mathew R, Bafiq R, Ramu J, et al. Spectral domain optical coherence tomography in patients with sickle cell disease. Br J Ophthalmol. 2015;99(7):967-972. doi: 10.1136/bjophthalmol-2014-305532 [DOI] [PubMed] [Google Scholar]
- 2.Lim JI, Cao D. Analysis of retinal thinning using spectral-domain optical coherence tomography imaging of sickle cell retinopathy eyes compared to age- and race-matched control eyes. Am J Ophthalmol. 2018;192(8):229-238. doi: 10.1016/j.ajo.2018.03.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Hussnain SA, Coady PA, Slade MD, et al. Hemoglobin level and macular thinning in sickle cell disease. Clin Ophthalmol. 2019;13(4):627-632. doi: 10.2147/OPTH.S195168 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Chow CC, Genead MA, Anastasakis A, Chau FY, Fishman GA, Lim JI. Structural and functional correlation in sickle cell retinopathy using spectral-domain optical coherence tomography and scanning laser ophthalmoscope microperimetry. Am J Ophthalmol. 2011;152(4):704-711. doi: 10.1016/j.ajo.2011.03.035 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Demirkaya N, van Dijk HW, van Schuppen SM, et al. Effect of age on individual retinal layer thickness in normal eyes as measured with spectral-domain optical coherence tomography. Invest Ophthalmol Vis Sci. 2013;54(7):4934-4940. doi: 10.1167/iovs.13-11913 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Chow CC, Shah RJ, Lim JI, Chau FY, Hallak JA, Vajaranant TS. Peripapillary retinal nerve fiber layer thickness in sickle-cell hemoglobinopathies using spectral-domain optical coherence tomography. Am J Ophthalmol. 2013;155(3):456-464. doi: 10.1016/j.ajo.2012.09.015 [DOI] [PubMed] [Google Scholar]
- 7.Thavikulwat AT, Cao D, Vajaranant TS, Lim JI. Longitudinal study of peripapillary thinning in sickle cell hemoglobinopathies. Am J Ophthalmol. 2019;202(6):30-36. doi: 10.1016/j.ajo.2019.02.006 [DOI] [PubMed] [Google Scholar]
- 8.Cai CX, Han IC, Tian J, Linz MO, Scott AW. Progressive retinal thinning in sickle cell retinopathy. Ophthalmol Retina. 2018;2(12):1241-1248. doi: 10.1016/j.oret.2018.07.006 [DOI] [PubMed] [Google Scholar]
- 9.Alamouti B, Funk J. Retinal thickness decreases with age: an OCT study. Br J Ophthalmol. 2003;87(7):899-901. doi: 10.1136/bjo.87.7.899 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Girkin CA, McGwin G Jr, Sinai MJ, et al. Variation in optic nerve and macular structure with age and race with spectral-domain optical coherence tomography. Ophthalmology. 2011;118(12):2403-2408. doi: 10.1016/j.ophtha.2011.06.013 [DOI] [PubMed] [Google Scholar]
- 11.Girkin CA, Sample PA, Liebmann JM, et al. ; ADAGES Group . African descent and glaucoma evaluation study (ADAGES): II. ancestry differences in optic disc, retinal nerve fiber layer, and macular structure in healthy subjects. Arch Ophthalmol. 2010;128(5):541-550. doi: 10.1001/archophthalmol.2010.49 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Girkin CA, McGwin G Jr, Xie A, Deleon-Ortega J. Differences in optic disc topography between Black and White normal subjects. Ophthalmology. 2005;112(1):33-39. doi: 10.1016/j.ophtha.2004.07.029 [DOI] [PubMed] [Google Scholar]
- 13.Hussnain SA, Coady PA, Stoessel KM. Paracentral acute middle maculopathy: precursor to macular thinning in sickle cell retinopathy. BMJ Case Rep. 2017;2017:bcr2016216124. doi: 10.1136/bcr-2016-216124 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Dell’Arti L, Barteselli G, Riva L, et al. Sickle cell maculopathy: identification of systemic risk factors, and microstructural analysis of individual retinal layers of the macula. PLoS One. 2018;13(3):e0193582. doi: 10.1371/journal.pone.0193582 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Han IC, Tadarati M, Pacheco KD, Scott AW. Evaluation of macular vascular abnormalities identified by optical coherence tomography angiography in sickle cell disease. Am J Ophthalmol. 2017;177(5):90-99. doi: 10.1016/j.ajo.2017.02.007 [DOI] [PubMed] [Google Scholar]
- 16.Ong SS, Linz MO, Li X, Liu TYA, Han IC, Scott AW. Retinal thickness and microvascular changes in children with sickle cell disease evaluated by optical coherence tomography (OCT) and OCT angiography. Am J Ophthalmol. 2020;209(1):88-98. doi: 10.1016/j.ajo.2019.08.019 [DOI] [PubMed] [Google Scholar]
- 17.Han IC, Zhang AY, Liu TYA, Linz MO, Scott AW. Utility of ultra-widefield retinal imaging for the staging and management of sickle cell retinopathy. Retina. 2019;39(5):836-843. doi: 10.1097/IAE.0000000000002057 [DOI] [PubMed] [Google Scholar]
- 18.Kong M, Kwun Y, Sung J, Ham D-I, Song Y-M. Association between systemic hypertension and macular thickness measured by optical coherence tomography. Invest Ophthalmol Vis Sci. 2015;56(4):2144-2150. doi: 10.1167/iovs.14-16080 [DOI] [PubMed] [Google Scholar]
- 19.Ataga KI, Kutlar A, Kanter J, et al. Crizanlizumab for the prevention of pain crises in sickle cell disease. N Engl J Med. 2017;376(5):429-439. doi: 10.1056/NEJMoa1611770 [DOI] [PMC free article] [PubMed] [Google Scholar]