A woman with didanosine retinopathy and non-cirrhotic portal hypertension

Introduction

Didanosine (2′,3′dideoxyinosine) is a reverse transcriptase inhibitor used in the treatment of HIV. It has been linked with retinal toxicity in a short series of case reports;^1–7 and well-known extraocular adverse reactions are pancreatitis and peripheral neuropathy.^8,9 A significant association has also been described between didanosine and non-cirrhotic portal hypertension. ¹⁰

Case Report

A 50-year-old HIV-2-positive black woman presented with complaints of gradual difficulty with near vision.

Diagnosis of HIV infection had been made eight years earlier (2004), at which time she presented with oral candidiasis, CD4+ count of 50 cells/μL, viral load 8100 copies/mL and was initiated on zidovudine, lamivudine and indinavir. In 2009, CD4+ count was 103 cells/μL and medications were replaced by didanosine (ddI) 400 mg once daily, abacavir 300 mg twice daily, raltegravir 400 mg twice daily and darunavir 600 mg/ritnoavir 100 mg twice daily. In January 2012 the patient had a diagnosis of histologically proven non-cirrhotic portal hypertension (NCPH) – liver biopsy showed regenerative nodular hyperplasia, duct hyperplasia, mild fibrosis and septal lymphocytic infiltrate – that was attributed to her prior 33 months of ddI exposure, thus motivating suspension of this drug. Other medical history included hysterectomy for leiomyomas and Pseudomonas aeruginosa urinary tract infection.

At presentation, CD4 + T-cell count was 160 cells/μL and viral load was undetectable. Best visual acuity was 20/20 in the right eye; 20/25 in the left eye, and Jl in each eye for near after correction with spectacles – at which point the patient claimed no further difficulty with her vision. Biomicroscopy and intraocular pressures were normal. Dilated fundus examination revealed mottled atrophic changes of the retinal pigment epithelium (RPE) in the periphery of both eyes (Figure 1). OPEN IN VIEWER

Fluorescein angiography showed early diffuse nummular hyperfluorescence, consistent with diffuse atrophy of the RPE reaching just inside the vascular arcades (Figure 2). Vessels, maculae and optic discs were of normal appearance. Optical coherence tomography was normal in both eyes. Full-field elec-troretinography (ERG) showed mild non-selective photoreceptor damage and the pattern ERG was normal. OPEN IN VIEWER

Dilated fundoscopy performed in 2004 was reportedly normal. There was no history of treatment with clofazimine or tenofovir. Chest X-ray was normal; bacteriological analysis for mycobacteria was negative in ascitic fluid; venereal disease research laboratory and serology for viral hepatitis were negative; and immunoglobulin M for cytomegalovirus was positive. There was no history of peripheral neuropathy, pancreatitis or lipoatrophy.

Discussion

Retinal toxicity in connection with ddI was first documented by Whitcup^1,2 in 1992. Three out of 43 children on ddI developed mottled atrophy of the peripheral RPE in an apparently cumulative dose-related fashion, with electrophysiological evidence for selective involvement of the RPE in one patient. After drug discontinuation, progression of lesions seemed to halt clinically; and in the latter patient, the electrooculography reverted to normal. Histopathological analysis of one patient was available later, where were described areas of RPE loss in the peripheral retina, specially in the midperiphery, with surrounding areas of hyper- and hypopigmentation. The territory between the vascular arcades was unaffected. Electronic microscopy showed numerous lamellar inclusions and cytoplasmic bodies in the RPE cells. ³

One year later a similar clinical picture was described in an adult, ⁴ who had unaffected visual acuity and normal electro-physiology. Four other cases in adults have been described since.^5–7 In one patient, follow-up data one year after suspension of ddI showed improvement of symptoms and electrophysiological performance. ⁷

Our report of asymptomatic retinal changes matches previous descriptions of ddI-associated ocular toxicity, in the absence of other likely causes. Furthermore, we raise the question of this patients’ particular sensitivity to ddI, based on the concomitant history of NCPH.^10,11 There was no history of treatment with tenofovir, known to increase plasma availability of ddI. ¹² This being said, we must also acknowledge that ddI NCPH may be an idiosyncratic reaction to ddI exposure. ¹³

Mitochondrial dysfunction has been suggested as a possible mechanism for RPE injury in this setting, although the latter remains unclear. ¹⁴ Whitcup ³ observed normal mitochondriae in his clinicopathological analysis. The inclusion bodies seen in the defective RPE cells resemble those found in lysosomal storage diseases – this has led to speculation over the possibility of impaired cell metabolism. An underlying tendency for flawed metabolism would explain our particular patients’ history.

DdI is one of the two antiretroviral drugs which have been linked with retinal toxicity, the other being zidovudine.^15,16

Our case adds to a short series of similar reports, thus providing further grounds for routine ophthalmological examination of patients on ddI. A dilated fundus observation is key in detecting the characteristic peripheral changes in this condition, which may be asymptomatic and altogether missed. In turn, this also leads us to consider that the real prevalence of ddI retinopathy may be considerably higher than reported so far. Additional studies are needed to confirm the benefit of routine eye exams in patients exposed to ddI, and to ascertain whether retinal lesions precede and predict extraocular complications.