Cidofovir treatment of HIV-associated cytomegalovirus polyradiculopathy

Abstract

Cytomegalovirus polyradiculopathy is an unusual complication of late-stage HIV infection, with a 100% mortality in the era before effective therapies [1–5]. We present the first report of the successful treatment of this condition with cidofovir with HAART, a previously unreported regimen. In June 2004, a 34-year-old woman with long-standing HIV infection presented with faecal incontinence, urinary hesitancy and low backache. She was diagnosed with HIV in 1998 after presenting with cerebral toxoplasmosis, and since diagnosis had multiple HIV-related complications exacerbated by poor adherence to antiretroviral therapy and multiple drug-resistant HIV. She had suffered severe cytomegalovirus retinitis and experienced ganciclovir-induced myelosuppression while undergoing treatment. At this presentation she had not taken any HIV treatment or cytomegalovirus prophylaxis for over one year. She had pre-existing peripheral neuropathy and vulval ulceration caused by herpes simplex virus infection. Examination revealed symmetrical flaccid weakness (power 3/5 all groups) of her lower limbs, sensory loss affecting all modalities from T11 to L3, and a loss of knee and ankle jerks with upgoing plantar reflexes. There was no evidence of active retinopathy. Her CD4 cell count was 28 × 106 cells/μl on admission and had been consistently less than 70 × 106 cells/μl for the preceding year. The HIV viral load was 1 500 000 copies/ml. An enhanced magnetic resonance imaging scan of the entire spine did not reveal any compressive pathology, but did show meningeal enhancement from L1 to L4. A lumbar puncture revealed 2270 white cells/ml (97% polymorphs) and 415 red cells/ml. A Gram stain and auramine phenol stain of the cerebrospinal fluid (CSF), cryptococcal antigen, herpes simplex, varicella zoster virus, enteroviruses, parechoviruses, Epstein–Barr virus and JC virus polymerase chain reactions and pneumococcal antigen were all negative as were appropriate cultures of both blood and CSF. CSF biochemistry showed a protein level of 2.43 g/dl and glucose 1.5 mmol/l (30% of serum). In the CSF cytomegalovirus was present at the very high level of 2.0 × 107 genome copies/ml compared with the much lower level of cytomegalovirus in the plasma of 5450 genome copies/ml (measured concurrently). Cytomegalovirus treatment was instigated in order to reduce cytomegalovirus viraemia until immune restoration with HAART. Effective therapeutic options for active cytomegalovirus disease were limited, and cidofovir was chosen as the agent least likely to cause complications. Ganciclovir was contraindicated because of previous severe myelosuppression, and foscarnet was avoided in view of her genital ulceration. Initial therapy was based on cidofovir regimens for cytomegalovirus retinitis, with 5 mg/kg intravenously fortnightly, with prehydration and probenicid modified for her reduced glomerular filtration rate. This was continued for 5 weeks, with no complications or deterioration in renal function, after which she received a monthly regimen. From admission to the initiation of therapy on day 4, progression of the polyradiculopathy had resulted in a sensory level at T8 and global power reduction in both lower limbs varying from 0 to 1/5 by muscle group. By the end of the first week of therapy this progression appeared to have halted, with the sensory level remaining at T8 and by week 4 there was a drop in sensory level and an improvement in power in both lower limbs. Despite clinical signs of recovery, the cytomegalovirus viraemia was slow to decrease. At week 4 a double-boosted protease inhibitor HAART regime based on resistance testing was started (lopinavir/ritonavir, atazanavir and zalcitabine), and it was only after starting HAART that cytomegalovirus was undetectable in the blood by polymerase chain reaction. Her sensory level continued to fall and muscle power increased. Rehabilitation was undertaken on our ward with intensive physiotherapy and occupational therapy input. Six months after admission, she was able to mobilize independently with a frame and transfer from bed to chair. She was discharged home, where she remains under follow-up one year after admission. This appears to be the first report of cytomegalovirus polyradiculopathy being successfully treated with cidofovir followed by HAART. Although neurological complications of cytomegalovirus occur in less than 1% of cytomegalovirus infections in HIV [6,7], the frequency of cytomegalovirus infection in HIV is such that clinicians need to be aware of its clinical spectrum, including cytomegalovirus polyradiculopathy. This patient had a classical presentation both clinically and in terms of her CSF and radiology findings, with no evidence of alternative pathology. Untreated cytomegalovirus polyradiculopathy carries 100% mortality from progressive cytomegalovirus disease, and still carries a 22% mortality rate [8] with anticytomegalovirus therapy. Ganciclovir is considered first-line therapy for most cytomegalovirus-related diseases [9], but second-line agents are often needed as a result of frequent toxicity. When ganciclovir therapy is contraindicated or has failed, monotherapy for cytomegalovirus with foscarnet still has a mortality rate as high as 60–70% [8], and cidofovir may be an effective alternative.