The Impact of Allogeneic Hematopoietic Stem Cell Transplantation on Sickle Cell Retinopathy and Maculopathy: A Prospective, Observational Study

Abstract

Sickle cell retinopathy (SCR) represents a significant ocular manifestation of sickle cell disease (SCD) that can dramatically impair visual acuity. SCR can be divided into non-proliferative SCR and proliferative SCR. Non-proliferative SCR involves abnormalities of the peripheral retina such as vascular occlusion without neovascularization, typical scarring (“black sunbursts”) and intraretinal hemorrhages (“salmon patches”). Proliferative SCR represents the more severe form of SCR, characterized by neovascularization. This can lead to vitreous hemorrhage or retinal detachment, which can temporarily or permanently impair visual acuity. In more recent years, modern imaging techniques such as spectral-domain optical coherence tomography (SD-OCT) and optical coherence tomography angiography (OCTA) have also revealed subclinical changes in the central part of the retina (the macula) in asymptomatic patients with SCD [1]. These changes are referred to as sickle cell maculopathy (SCM) and include macular thinning, enlargement of the foveal avascular zone (FAZ) and lower vessel densities. Both SCR and SCM are progressive in nature and associated with increasing age [2].

Allogeneic hematopoietic stem cell transplantation (HSCT) is the only established curative treatment option for patients with SCD. Improvements in non-myeloablative regimens have led to excellent disease-free and overall survival in transplanted adults with SCD, also in the setting of haploidentical HSCT [3]. SCD-related organ damage is one of the main indications for HSCT, but data on long-term outcomes regarding the progression of organ complications is scarce [4]. This is particularly true for adult patients, who frequently already have developed (sub-)clinical organ damage before HSCT.

Evidence concerning the behavior of SCR and SCM after HSCT is especially scarce. The only study on SCR after HSCT was conducted in children with SCD and did not include SCM [5]. This leaves a significant gap in understanding how these conditions evolve in adults with SCD undergoing HSCT. We hypothesized that HSCT has the potential to stabilize pre-existing SCR and will prevent the development of new SCR in patients with SCD. Therefore, the aim of this study was to evaluate the natural course of SCR and SCM in an adult cohort of SCD patients undergoing HSCT.

In this prospective observational study in adults with SCD (HbSS, HbSC, HbSβ⁰ and HbSβ⁺ genotype), retinopathy and maculopathy were assessed before and at least 1 year post-transplant. All SCD patients undergoing non-myeloablative matched sibling donor (MSD) and haploidentical bone marrow transplantation were included [3, 6]. Patients undergoing MSD transplantation received a 3-month preconditioning with azathioprine/hydroxyurea, followed by alemtuzumab (1 mg/kg) and 3 Gy total body irradiation (TBI) [6]. The conditioning regimen for haploidentical bone marrow transplantation consisted of anti-thymocyte globulin (4.5 mg/kg), thiotepa (10 mg/kg), fludarabine (150 mg/kg), cyclophosphamide (29 mg/kg), TBI (2 Gy), and post-transplant cyclophosphamide (100 mg/kg) [3]. Indications for HSCT included: frequent vaso-occlusive episodes (87.5%), history of acute chest syndrome (56.3%), pulmonary hypertension (19.4%), and history of stroke (12.9%). All consecutive non-transplant SCD patients who underwent routine ophthalmic screening (with available SD-OCT and OCTA scans) during the same period (January 2018–August 2023) as the first ophthalmic screening of transplanted patients were included as the control group. Prior to transplantation, 21 (65.6%) patients were treated with hydroxyurea, and 10 (31.3%) received chronic red blood cell exchange transfusions. In the control group, 17 patients (29.8%) were treated with hydroxyurea, and 3 patients (5.3%) received chronic transfusions. Nineteen (59.4%) patients had a donor with sickle cell trait (HbAS).

Patients with hereditary or acquired retinopathy (e.g., diabetic retinopathy, retinal vascular occlusions), any ocular media opacities preventing detailed imaging, high myopia (> 6 diopters) or vitreomacular interface abnormalities were excluded. SCR stage was determined for each eye based on fundoscopic examination as: [1] no signs of SCR, [2] signs of non-proliferative SCR, or [3] proliferative SCR. Progression of SCR was defined as an increase in SCR stage and/or number of lesions or the occurrence of vitreous hemorrhage or retinal detachment.

Macular SD-OCT and OCTA scans were obtained with the Optovue RTVue XR Avanti. For each eye, a 3 × 3 mm and a 6 × 6 mm scan centered on the fovea were acquired. Foveal avascular zone measurements (area, perimeter and acircularity index) were automatically calculated. Figure S1 illustrates representative normal and abnormal OCTA scans. The macular thickness color-coded map was qualitatively evaluated. Macular thinning was defined as the presence of blue areas in at least one of the standard ETDRS (Early Treatment of Diabetic Retinopathy Study) macular subfields (Figure S2). Progression of macular thinning was defined as new areas of macular thinning in previously unaffected subfields. Linear mixed models and generalized estimating equations analysis were used to compare ophthalmic outcomes between groups, considering interocular correlation. Data from HbSS patients in the control group were also analyzed separately, given their progression rate of SCR and SCM would likely align more closely with the HSCT patient group, which predominantly consisted of HbSS patients.

Between January 2018 and August 2023, 32 HSCT patients (29 of African origin, 3 of Middle-Eastern origin) and 57 consecutive non-transplant control patients were included (56 of African origin and 1 of Chinese origin). Baseline characteristics are outlined in Table 1. Median age at transplantation was 26.5 years (range 18–49). Median age at baseline of control patients was 31 years (range 17–66). In the transplant group, 16 patients (50%) were male and 16 (50%) female. The control group consisted of 25 male patients (43.9%) and 32 female patients (56.1%). Median follow-up time was 28.5 months (range 12–74) in the transplant group and 35 months (range 8–70) in the control group (p = 0.728).

TABLE 1. Baseline characteristics.

	HSCT patients (n = 32)	Control patients (n = 57)	p	Control patients (HbSSonly, (n = 20)	p
Median age, years (range)	26.5 (18–49)	31 (17–66)	0.116*	30.5 (18–61)	0.239*
Hb genotype
HbSS	24 (75%)	20 (35.1%)	< 0.001†	20 (100%)	N/A
HbSC	3 (9.4%)	31 (54.4%)		0 (0%)
HbSβ0	3 (9.4%)	3 (5.3%)		0 (0%)
HbSβ+	2 (6.3%)	3 (5.3%)		0 (0%)
Sex
Male, n (%)	16 (50%)	25 (43.9%)	0.577†	9 (45%)	0.726†
Female, n (%)	16 (50%)	32 (56.1%)		11 (55.0%)
Systemic treatment
Hydroxyurea, n (%)	21 (65.6%)	17 (29.8%)	0.001†	11 (55%)	0.444†
Chronic transfusion, n (%)	10 (31.3%)	3 (5.3%)	0.001‡	3 (15%)	0.188‡
VA < 20/25, n eyes (%)	1/64 (1.6%)	0/100 (0%)	N/A	0/34 (0%)	N/A
Median follow-up time in months (range)	28.5 (12–74)	35 (8–70)	0.728*	30.5 (10–74)	0.781*

• Abbreviations: HSCT, hematopoietic stem cell transplantation; VA, visual acuity.
• * Mann–Whitney U Test.
• ^† Pearson Chi-Square.
• ^‡ Fisher's Exact Test.

Twenty-three (71.9%) patients underwent MSD HSCT and 9 (28.1%) patients underwent haploidentical HSCT. Thirty of the 32 patients engrafted successfully and are alive and disease-free with normalized hemoglobin and hemolysis parameters. Due to the death of two patients in the HSCT group, follow-up examinations were performed in 30 HSCT patients. Except for one eye with impaired visual acuity due to pre-existing amblyopia, all other eyes maintained normal visual acuity throughout. No cases of cataract secondary to the HSCT were observed.

Non-proliferative SCR was present in 13 eyes (21.7%) and proliferative SCR in 4 eyes (6.7%) of the HSCT patients (Table 2). During follow-up examination, the left eye of one patient (1.7%) progressed from no SCR to proliferative SCR (26 months post-transplant). The retinopathy status of the other 59 eyes remained stable. In the control group, non-proliferative SCR was present in 24 eyes (24%) and proliferative SCR in 19 eyes (19%) at baseline. This increased to 29 eyes (29%) and 26 eyes (26%) respectively after a median of 35 months, resulting in a progression rate of 15%. This was higher than the progression rate in the transplant group (1.7%, p = 0.026). Among control patients with only HbSS genotype, non-proliferative SCR was present in 10 eyes (29.4%) that increased to 15 eyes (44.1%) at follow-up. Proliferative SCR was present in 1 eye (2.9%) that remained stable upon follow-up. The progression rate in this group was therefore 14.7%, which was also higher than the progression rate in the transplant group (p = 0.049).

TABLE 2. Progression of sickle cell retinopathy and maculopathy in HSCT and control patients.

	HSCT patients (n = 30)		Control patients (n = 57)		p *	Control patients (HbSS only), (n = 20)		p *
	Baseline	Follow-up	Baseline	Follow-up		Baseline	Follow-up
Retinopathy
None, n eyes (%)	43/60 (71.7%)	42/60 (70%)	57/100 (57%)	45/100 (45%)		23/34 (67.6%)	18/34 (52.9%)
NPSCR, n eyes (%)	13/60 (21.7%)	13/60 (21.7%)	24/100 (24%)	29/100 (29%)		10/34 (29.4%)	15/34 (44.1%)
PSCR, n eyes (%)	4/60 (6.7%)	5/60 (8.3%)	19/100 (19%)	26/100 (26%)		1/34 (2.9%)	1/34 (2.9%)
Progression of retinopathy, n eyes (%)	—	1/60 (1.7%)	—	15/100 (15%)	0.026	—	5/34 (14.7%)	0.049
Maculopathy
Macular thinning, n eyes (%)	12/28 (42.9%)	14/28 (50%)	40/100 (40%)	51/100 (51%)		20/34 (58.8%)	23/34 (67.6%)
FAZ area (mean ± SD)	0.352 (±0.130)	0.328 (±0.106)	0.371 (±0.153)	0.387 (±0.152)		0.393 (±0.189)	0.422 (±0.183)
FAZ perimeter (mean ± SD)	2.27 (±0.47)	2.24 (±0.43)	2.33 (±0.49)	2.42 (±0.48)		2.39 (±0.59)	2.58 (±0.53)
FAZ AI (mean ± SD)	1.10 (±0.03)	1.12 (±0.07)	1.10 (±0.03)	1.12 (±0.07)		1.11 (±0.03)	1.16 (±0.10)
Progression of macular thinning, n eyes (%)*	—	2/28 (7.1%)	—	26/100 (26%)	0.046	—	10/34 (29.4%)	0.047
Ocular complications
VH, n eyes (%)	2/60 (3.3%)	2/60 (3.3%)	8/100 (8%)	11/100 (11%)		0/34 (0%)	0/34 (0%)
RD, n eyes (%)	0/60 (0%)	0/60 (0%)	1/100 (1%)	1/100 (1%)		0/34 (0%)	0/34 (0%)
Laser treatment, n eyes (%)	3/60 (5%)	3/60 (5%)	7/100 (7%)	14/100 (14%)		1/34 (2.9%)	1/34 (2.9%)

• Abbreviations: AI, acircularity index; FAZ, foveal avascular zone; HSCT, hematopoietic stem cell transplantation; NPSCR, non-proliferative sickle cell retinopathy; PSCR, proliferative sickle cell retinopathy; RD, retinal detachment; VH, vitreous hemorrhage.
• * GEE population-averaged model.

At baseline, a history of vitreous hemorrhage was present in two eyes (3.3%) in the transplant group and eight eyes (8%) in the control group. During follow-up, three eyes (3%) of control patients, all with HbSC genotype, were affected. Retinal detachment was not observed in any of the patients during the study. Although laser treatment for proliferative SCR was performed in three eyes (5%) in the transplant group prior to HSCT, no laser treatment was needed post-transplant. In the control group, seven eyes (7%) had received laser treatment at baseline. This increased to 14 eyes (14%) during follow-up.

SD-OCT and OCTA scans of sufficient quality for both baseline and follow-up evaluations were available for 28 eyes (14 patients) and 25 eyes (14 patients), respectively, in the transplant group. In the control group, scans were available for 100 eyes from all (57) patients. Macular thinning was observed in 12 eyes (42.9%) in eight HSCT patients at baseline. This increased to 14 eyes (50%) in nine patients at follow-up (progression rate 7.1%; Table 2). One patient who initially had macular thinning only in the right eye at baseline also developed macular thinning in the left eye 2 years post-transplant (Figure S3). This patient had non-proliferative SCR and a normal visual acuity in both eyes at baseline and at follow-up. Another patient, without macular thinning or retinopathy at baseline, developed new-onset macular thinning in the left eye during follow-up. This was detected 1 year post-transplant (Figure S4). Visual acuity remained unaffected, and no signs of retinopathy were present at the follow-up examination. In the control group, 40 eyes (40%) in 26 patients had macular thinning at baseline, increasing to 51 eyes (51%) in 34 patients upon follow-up evaluation (progression rate 26%, p = 0.046). The progression rate of macular thinning among HbSS control patients separately (29.4%) was also higher than the progression rate in the transplant group (p = 0.047). Table 3 outlines the differences in the area, perimeter, and acircularity index of the FAZ between baseline and follow-up. The FAZ perimeter increased in the HbSS control group but not in the transplant group (p = 0.018).

TABLE 3. Comparison of changes in FAZ parameters between HSCT and control patients (all genotypes).

			p *
	HSCT (n = 25 eyes)	Control patients (n = 100 eyes)
FAZ area change (mean ± SD)	−0.024 ± 0.072	0.016 ± 0.090	0.064
FAZ perimeter change (mean ± SD)	−0.04 ± 0.17	0.09 ± 0.32	0.066
FAZ AI change (mean ± SD)	0.02 ± 0.06	0.02 ± 0.07	0.985

HSCT versus controls (HbSS only)			p *
	HSCT (n = 25 eyes)	HbSS controls (n = 34 eyes)
FAZ area change (mean ± SD)	−0.024 ± 0.072	0.028 ± 0.120	0.102
FAZ perimeter change (mean ± SD)	−0.04 ± 0.17	0.19 ± 0.42	0.018
FAZ AI change (mean ± SD)	0.02 ± 0.06	0.04 ± 0.11	0.282

• Abbreviations: AI, acircularity index; FAZ, foveal avascular zone; HSCT, hematopoietic stem cell transplantation.
• * Mixed-effects ML regression.

We show that the progression rates of SCR and SCM in adult SCD patients undergoing non-myeloablative HSCT were significantly lower than in the non-transplant control group. These results suggest that HSCT has the potential to limit the progression of existing or the development of new SCD-related ocular complications. Progression of SCR occurred in only one eye of one transplanted patient. Retinal neovascularization is typically the result of angiogenesis secondary to ischemia. As the transplanted patient who developed new-onset proliferative retinopathy had full donor chimerism with normalized erythrocyte phenotype and hemoglobin levels, it remains unclear what caused the progressive retinopathy in this patient. Notably, the progression of SCR in this eye may have already been present prior to HSCT. The baseline examination of this patient was conducted unusually early (16 months prior to HSCT). A study in 247 pediatric patients undergoing HSCT reported two cases of new-onset retinopathy and one case of progression of pre-existing retinopathy, reflecting a similar progression rate to our findings [5]. However, this study did not provide data on the time interval between the baseline examination and HSCT, leaving the possibility that SCR progression was present prior to transplantation unaddressed. Our study is the first prospective evaluation of SCR and SCM after HSCT in adult patients, including a substantial follow-up period. However, there remains a need for longer-term studies, as SCR can take several years to develop and its incidence tends to increase with age.

In the transplanted patients, progression of SCM was more prevalent than progression of SCR. However, macular vascular abnormalities were less progressive in patients who had undergone HSCT compared to control patients. This suggests that, while HSCT may halt further macular microvascular occlusion, macular thinning can still occur post-transplant. The post-transplant macular thinning might be a consequence of macular microvascular occlusions that developed prior to transplantation. This was also supported by our finding that the area and perimeter of the FAZ did not increase in the transplant group. Close collaboration between ophthalmologists and hematologists is essential to ensure timely detection and monitoring of ocular complications and to integrate ophthalmic findings into the overall clinical management of patients with SCD.

In conclusion, our study demonstrates that non-myeloablative HSCT effectively diminishes the progression of SCR and SCM in adults with SCD. This underscores the potential of HSCT in limiting the progression of existing and the development of new SCD-related organ complications. Further research with longer follow-up time is needed to assess long-term outcomes following HSCT.