Characteristics of Pentosan Polysulfate Sodium–Associated Maculopathy and Similarities With Other Maculopathies Commonly Managed in a Retina Practice

Abstract

Purpose:

This work describes characteristics of pentosan polysulfate sodium (PPS)–associated maculopathy and its similarities with common maculopathies in a retina practice cohort.

Methods:

Thirty-two patients were identified through electronic medical record query who were exposed to PPS. One patient was excluded for lack of retinal imaging. Thirty-one patients (62 eyes) were included. A retrospective review was used to obtain patient characteristics, examination findings, and retinal imaging of the study patients. Classification into “likely,” “unlikely,” or “possible” to have PPS-associated maculopathy groups was based on the fundus photography and retinal imaging. Main outcome measures were best-corrected visual acuity, age, sex, diagnosis of reason for referral, allocation into designated maculopathy group, and presence of choroidal neovascularization.

Results:

Of 31 patients (62 eyes), the median age was 70 years (range, 24-104 years) and the majority were women (87%). Mean best-corrected visual acuity was 0.3 ± 0.4 logMAR at presentation. The most common reason for referral was age-related macular degeneration (29%). Maculopathy grades were “likely” (29%, 9 total patients), “possible” (26%, 8 total patients), or “unlikely” (45%, 14 total patients). Choroidal neovascularization was noted in 9.7% of all eyes and 11% of eyes in the “likely” group. The “possible” and “likely” groups had older ages of presentation (P < .05) compared with the “unlikely” group.

Conclusions:

A high percentage (55%) of patients with a history of chronic PPS exposure showed features of “likely” or “possible” maculopathy. Similarities with common maculopathies such as age-related macular degeneration and the importance of screening and recognizing at-risk patients are highlighted.

Introduction

Pentosan polysulfate sodium (PPS; Elmiron, Janssen Pharmaceuticals) is one of only a few US Food and Drug Administration–approved treatments of interstitial cystitis, a chronic and debilitating but poorly understood retropubic pain syndrome. ^{1 -3} PPS is a semisynthetic heparin-like compound possessing both anticoagulant and fibrinolytic effects with structural similarities to glycosaminoglycans. ^4,5 Although not fully understood, its mechanism of action is believed to involve a buffering effect secondary to direct binding to bladder epithelial glycosaminoglycans. ⁴ Generally well tolerated, PPS became a widely used oral treatment of bladder pain associated with interstitial cystitis since its Food and Drug Administration approval in 1996. ^4,5

However, recent reports have described a progressive pigmentary maculopathy associated with chronic PPS therapy, calling its safety into question. ⁶ Subsequent studies have shown progression of the maculopathy despite cessation of the drug, as well as an association with the development of choroidal neovascularization (CNV). ^7,8 Cumulative PPS dose has recently been shown to be a significant risk factor for the development of maculopathy. ⁹

We have described a cohort of 32 patients on PPS therapy who were seen at the Retina Group of Washington (RGW; Washington, DC; Virginia; and Maryland) from February 2016 to January 2020 and displayed varying levels of pigmentary maculopathy. This report adds to the emerging evidence of an association between chronic PPS and a toxic maculopathy, and it highlights the difficulties that may exist in distinguishing between PPS maculopathy and other maculopathies commonly encountered in a retina practice.

Methods

This is a retrospective case series of patients taking PPS who were examined at RGW from February 2016 to January 2020. All patients were evaluated and examined by vitreoretinal fellowship–trained physicians. Three reviewers reviewed the imaging of all patients included in the study: 1 senior ophthalmology resident and 2 vitreoretinal fellowship–trained physicians. The reviewers were blinded to the examination findings, diagnosis, and grading of the other reviewers. There was a correlation in grading between the reviewers. There was an agreement on the grading of “likely,” “possible,” and “unlikely” PPS-associated maculopathy in 77% of patients (n = 24) between the 3 reviewers. The final grading for those patients not in agreement between the 3 reviewers was performed by the 2 vitreoretinal fellowship–trained physicians.

Patient Selection and Data Collection

The electronic medical record (EMR) of RGW (Electronic Medical Assistant, Modernizing Medicine) was queried for patients examined between February 2016 and January 2020 who had “pentosan polysulfate,” “pentosan polysulfate sodium,” or “Elmiron” (ALZA Corp) in their medication list. This search returned 32 patients followed by multiple providers at RGW for a variety of diagnoses. Baseline patient characteristics, diagnoses, and treatment history were collected (Table 1). Each patient had undergone a complete baseline ophthalmic evaluation and imaging including color fundus photography, fundus autofluorescence (FAF), and/or macular spectral-domain optical coherence tomography (OCT) (Zeiss Cirrus 5000, Carl Zeiss Meditec, Spectralis, Heidelberg Engineering). The majority of patients had follow-up imaging.

Table 1.

Demographic and Examination Findings, Characteristics, and Pentosan Polysulfate Sodium Exposure.

	Grade								P vs L vs U	P vs L	P vs U	L vs U
	Overall		P		L		U
	No.	(%)	No.	(%)	No.	(%)	No.	(%)	P value
Sex
Female	54	(87.1)	16	(100.0)	14	(77.8)	24	(85.7)	.18	.11	.28	.69
Male	8	(12.9)	0	0	4	(22.2)	4	(14.3)
Smoker
Current	2	(3.2)	0	0	0	0	2	(7.1)	.25	.11	.46	.42
Former	14	(22.6)	6	(37.5)	2	(11.1)	6	(21.4)
Never	46	(74.2)	10	(62.5)	16	(88.9)	20	(71.4)
Race
Asian	4	(6.5)	0	0	0	0	4	(14.3)	.09	NA	.14	.12
Black	2	(3.2)	0	0	0	0	2	(7.1)
White	54	(87.1)	16	(100.0)	18	(100.0)	20	(71.4)
Hispanic	2	(3.2)	0	0	0	0	2	(7.1)
PPS dose, mg/d
100	32	(51.6)	6	(37.5)	12	(66.7)	14	(50.0)	.41	.17	.53	.55
200	14	(22.6)	6	(37.5)	2	(11.1)	6	(21.4)
300	16	(25.8)	4	(25.0)	4	(22.2)	8	(28.6)
Peripapillary atrophy
360°	16	(25.8)	4	(25.0)	12	(66.7)	0	0	<.0001	.0003	.006	<.0001
<360°	14	(22.6)	2	(12.5)	6	(33.3)	6	(21.4)
Right eye only 360°	2	(3.2)	2	(12.5)	0	0	0	0
Absent	30	(48.4)	8	(50.0)	0	0	22	(78.6)
RPE atrophy, DD
<1	8	(12.9)	4	(25.0)	2	(11.1)	2	(7.1)	<.0001	.10	.003	<.0001
>1	10	(16.1)	2	(12.5)	8	(44.4)	0	0
>1 OD, <1 OS	2	(3.2)	2	(12.5)	0	0	0	0
Absent	42	(67.7)	8	(50.0)	8	(44.4)	26	(92.9)
CNV
No	56	(90.3)	14	(87.5)	16	(88.9)	26	(92.9)	.07	.46	.04	.21
Yes	6	(9.7)	2	(12.5)	2	(11.1)	2	(7.1)

Abbreviations: CNV, choroidal neovascularization; DD, disc diameter; L, likely PPS maculopathy; NA, not available; OD, right eye; OS, left eye; P, possible PPS maculopathy; PPS, pentosan polysulfate sodium; RPE, retinal pigment epithelium; U, unlikely PPS maculopathy.

Expert reviewers evaluated all fundus imaging for characteristic features of the PPS-associated pigmentary maculopathy first described by Pearce et al in 2018. ⁶ These included (1) macular hyperpigmented spots, yellow-orange deposits, and/or patchy retinal pigment epithelium (RPE) atrophy on color fundus photography; (2) dense array of macular and peripapillary alternating hyperautofluorescent and hypoautofluorescent spots and reticular changes on FAF; (3) nodular RPE thickening on OCT corresponding to hyperreflectance on near-infrared reflectance imaging; and (4) bilaterality. Additional features, including but not limited to the presence of cystoid macular edema and CNV, were also noted. Based on the characteristics described by Pearce et al, ⁶ patients were classified in 1 of 3 groups as “unlikely,” “likely,” or “possible” to have PPS-associated maculopathy.

Statistical Analysis

Statistical Analysis Software version 9.4 (SAS Institute Inc) was used to perform the analysis. Summary statistics on all characteristics were obtained for the overall sample and by study groups. The Kruskal-Wallis test was used to examine differences in the averages of continuous variables between the 3 maculopathy groups. The Wilcoxon rank sum test was used to examine differences in the averages of continuous variables between the groups. Fisher exact and χ² tests were used to investigate differences for categorical variables. Statistical significance was defined as a P value less than .05.

Results

EMR query returned 32 patients with exposure to PPS. One patient was excluded because of a lack of retinal imaging. All other patients included in this study had FAF and spectral-domain OCT imaging obtained. Overall patient characteristics can be found in Table 1. Of 31 patients (62 eyes) included in the study, the median age was 70 years (range, 24-104 years) and the majority were women (87%). History of tobacco use was reported as current (3%), former (23%), and never (74%). Daily PPS dose was 100 mg (52%), 200 mg (22%), and 300 mg (26%). Mean best-corrected visual acuity (in logMAR) was 0.3 ± 0.4 OD and 0.3 ± 0.5 OS (Snellen equivalent, 20/40 OU) at presentation, and 0.4 ± 0.6 OD (Snellen, 20/50 OD) and 0.6 ± 0.9 OS (Snellen, 20/80 OS) at final visit. The average time elapsed between presentation and final visit was 930 days (SD = 467). The most common reason for provider referral was age-related macular degeneration (AMD) (29% of patients). PPS-associated maculopathy grades were assigned as “likely” (29% of total patients), “unlikely” (45% of total patients), and “possible” (26% of total patients).

Intergroup analysis (Table 2) revealed a statistically significant older age at presentation in the “likely” group compared with the “unlikely” group (P < .003). Neither race, sex, nor smoking status showed significant differences between any of the 3 maculopathy groups (P < .05). Of the 9 patients in the “likely” maculopathy group, 7 (78%) received a diagnosis of AMD after initial evaluation. CNV was noted in 9.7% of all eyes, including 1 eye of 2 patients likely to have PPS-associated (11.1%). Both eyes were treated with serial antivascular endothelial growth factor (anti-VEGF) treatment. Documentation of PPS exposure or high-risk medication screening was present in 29% of all patient visit notes at the time of EMR review.

Table 2.

Age, Subjective Visual Acuity Measurements, and Follow-up Time.

	Grade												P vs L vs U	P vs L	P vs U	L vs U
	Overall			P			L			U
	No.	Mean	SD	No.	Mean	SD	No.	Mean	SD	No.	Mean	SD	P value
Age at presentation, y	62	67	16.4	16	75	17	18	70.8	4.9	28	59.4	17.5	.003	.05	.003	.02
BCVA right eye at presentation	62	0.3	0.4	16	0.5	0.7	18	0.2	0.1	28	0.2	0.2	.29	.12	.22	.83
BCVA left eye at presentation	62	0.3	0.5	16	0.4	0.6	18	0.3	0.3	28	0.2	0.5	.08	.79	.21	.02
BCVA right eye at last follow-up	52	0.4	0.6	16	0.6	1.0	16	0.3	0.3	20	0.2	0.2	.21	.55	.20	.09
BCVA left eye at last follow-up	52	0.6	0.9	14	0.6	1.0	16	0.5	0.7	20	0.6	1.0	.43	.95	.52	.15
Time elapsed, d	52	902	451	16	825	457	16	930.4	467	20	942	448	.71	.72	.35	.91

Abbreviations: BCVA, best-corrected visual acuity; L, likely PPS maculopathy; P, possible PPS maculopathy; U, unlikely PPS maculopathy.

Conclusions

In this retrospective review of patients taking PPS, 55% were found to have “likely” or “possible” retinal toxicity. However, at the initial evaluation, 78% of patients (24 total patients) were diagnosed with AMD. Additional diagnoses included pattern dystrophy (1 patient), idiopathic CNV (1 patient), polypoidal choroidal vasculopathy (1 patient), toxic maculopathy (1 patient), and macular dystrophy not otherwise specified (1 patient). The overlap of clinical findings in PPS maculopathy with more common retinal diseases demonstrates the importance of identifying patients at risk for PPS maculopathy. Most patients were initially diagnosed with AMD because imaging findings resembled this clinical disease; however, many of the clinical findings of PPS-associated maculopathy overlap with inherited retinal dystrophies as well and should be included in the differential diagnosis in the correct clinical context.

One imaging finding that may help differentiate between pattern dystrophy and PPS-associated maculopathy is the pattern of peripapillary atrophy found on FAF. ¹⁰ There is typically a peripapillary halo of hypoautofluorescence with a surrounding ring of hyperautofluorescence in patients with PPS-associated maculopathy, whereas patients with ABCA4-related retinopathies showed no findings on the peripapillary area on FAF. ¹⁰

While the clinical and imaging findings of PPS-associated maculopathy can be striking, difficulty exists in differentiating those patients who may be classified in the “possible” or “unlikely” category. The criteria to select patients in the “likely,” “possible,” or “unlikely” categories were used from the peer-reviewed and published article “Strength of Association Between Pentosan Polysulfate and a Novel Maculopathy” by Hanif and colleagues. ¹¹ The patients were placed into the “likely” group if they displayed all the criteria described in the paper by Hanif et al (Figure 1). Patients were placed into the “possible” group if they had nonspecific pigmentary changes that did not closely resemble the “likely” group or had some but not all the criteria of the “likely” group (Figure 2). The “unlikely” category was clearly dissimilar and did not have any of the criteria of the “likely” group (Figure 3).

Figure 1.

Imaging of the left eye of patient 5, who was likely to have PPS-associated maculopathy. (A) Fundus photograph depicting patchy retinal pigment epithelium atrophy with yellowish deposits. (B) Fundus autofluorescence showing macular and peripapillary alternating hyperautofluorescent and hypoautofluorescent spots and reticular changes. (C) Optical coherence tomography showing nodular retinal pigment epithelium thickening.

Figure 2.

(A) Fundus photograph of the left eye (patient 8, “possible” maculopathy group) depicting macular pigmentary changes and yellowish macular lesions centered around the fovea. (B) Fundus autofluorescence showing alternating hyperautofluorescent and hypoautofluorescent spots in the macula centered around the fovea. Note the lack of the peripapillary halo of alternating hyperautofluorescent and hypoautofluorescent spots typically found in pentosan polysulfate sodium–associated maculopathy. (C) Optical coherence tomography of the left eye showing nodular retinal pigment epithelium deposits but also the presence of drusen. This patient displays certain features of pentosan polysulfate sodium maculopathy but lacks other key features such as the presence of peripapillary changes on fundus autofluorescence.

Figure 3.

(A) Fundus photograph of the right eye (patient 4, “unlikely” maculopathy group) depicting yellowish lesions scattered throughout the macula. (B) Fundus autofluorescence showing a large, confluent area of hypoautofluorescence superior to the fovea with scattered areas of punctate hyperautofluorescence around the macula and nasal to the optic nerve. There is no peripapillary ring of alternating hyperautofluorescence and hypoautofluorescence changes. (C) Optical coherence tomography of the right eye showing drusen and the absence of retinal pigment epithelium excrescences or deposits. This patient lacks many key features of pentosan polysulfate sodium–associated maculopathy.

Some clinical features of the patients placed in the “possible” group included asymmetry of disease, a mix of drusen and RPE thickening (the presence of drusen may indicate AMD), and asymmetric or less than 360° peripapillary hyperautofluorescence changes in both eyes, whereas patients in the “unlikely” group featured significantly more drusen with minimal RPE thickening, noncentral macular involvement, or absent peripapillary hyperautofluorescence changes.

Few studies have reported the presence of CNV or cystoid macular edema in patients suspected of PPS-associated maculopathy. ^8,10 However, of the 62 eyes included in this study, 2 eyes in the “likely” group (3.12%) (patients 1 [Figure 4] and 17) and 2 eyes in the “possible” group (3.12%) (patients 10 and 20) displayed evidence of CNV for a total of 6.5%. Patient 1 received 12 intravitreal injections of aflibercept (Eylea, Regeneron) at 4- to 8-week intervals with poor overall response, eventually discontinuing therapy because of chronic subretinal and intraretinal fluid with progressive submacular fibrosis despite treatment. Patient 17 had received more than 25 intravitreal bevacizumab (Avastin, Roche Pharmaceuticals) injections with good response but suddenly developed subretinal fluid with progressive atrophy despite monthly treatment. Patients 10 and 20 similarly showed good response to anti-VEGF early, but both suddenly became poorly responsive to therapy despite monthly treatment and changing to alternative anti-VEGF agents. The 2 eyes in the “possible” group that developed a CNV membrane had drusen indicative of AMD. CNV was diagnosed by correlating examination findings, OCT, and intravenous fluorescein angiography.

Figure 4.

Imaging of the left eye of patient 1, who was in the “likely” maculopathy group. (A) Fundus photograph showing scattered yellowish deposits around the fovea and central geographic atrophy. (B) Fundus autofluorescence showing a large area of central macular hypoautofluorescence with bordering hyperautofluorescence and an identical halo of hypoautofluorescence around the optic nerve with bordering hyperautofluorescence. (C) Fluorescein angiogram at 29 seconds. (D) Fluorescein angiogram at 5 minutes showing area of hyperfluorescence indicating leakage. (E) Optical coherence tomography showing cystic intraretinal hyporeflective changes indicating retinal edema with adjacent hyperreflective elevation of the retinal pigment epithelium.

Intergroup analysis showed a higher likelihood of CNV only when the “possible” group was compared with the “unlikely” group (P = .0379). This was perhaps related to many of the eyes in the “possible” group having advanced disease with significant atrophic changes both on fundus examination and retinal imaging, making differentiation between AMD more difficult. Our data also suggested an age-related component of the maculopathy; however, it is likely that this was confounded by duration of PPS exposure. Despite this, it is worth noting that the maculopathy had a higher incidence in older patients and may have a higher association with CNV than previously thought.

It was determined that the progress notes of only 29% of patients acknowledged PPS exposure or high-risk medication screening at the time of EMR review. While all patients were contacted by their physician following the initial report of a possible maculopathy per the description of Pearce et al, ⁶ this highlighted the need for an improved intake screening and an EMR alert system that would ensure physicians are notified of high-risk medication exposure.

A significant limitation to this study was the absence of patients' PPS-dosage information. RGW's EMR system does not automatically link to pharmacy databases and upload or update drug information. It therefore depends on patients self-reporting daily and cumulative dose on initial visit intake surveys, as well as any dose changes at subsequent visits, which we found to often be missing or incomplete. Despite attempts to obtain this information retrospectively for the purpose of this series, patients were unable to confidently provide accurate dose or duration of therapy to justify inclusion. This study found a high incidence of “likely” or “possible” PPS maculopathy. In a retina practice there is an inherent selection bias because many patients are referred to the practice because of the presence of funduscopic or visual changes. Additional limitations included the retrospective nature of the study, a small patient cohort, and a lack of electrophysiologic testing and genetic analysis.

In conclusion, this case series adds to the growing evidence of a pigmentary maculopathy associated with chronic PPS use. While previous reports have provided insights into a variety of patient cohorts, our data represents that of one of the largest retina-only practices in the United States. Using systematic grading criteria, clinicians can definitively classify those suffering from PPS-associated maculopathy. As reported by Barnes et al, the imaging characteristics of PPS maculopathy can have as high as 100% sensitivity and 99.6% specificity for the identification of PPS maculopathy. ¹² This review allowed our practice to identify patients whose diagnosis of PPS maculopathy was missed prior to the initial publication by Pearce and colleagues. ⁶ Efforts were made to contact these patients for further evaluation and counseling. We encourage other practices to similarly review their records if feasible to help identify at-risk patients.