Article Text

Download PDFPDF

Severe intraocular inflammation after a change of HAART
  1. Medical Eye Unit, St Thomas's Hospital
  2. London SE1 7EH
  3. Kobler Centre, Chelsea and Westminster Hospital, London
  4. Medical Eye Unit, St Thomas's Hospital
  5. London SE1 7EH
  1. Medical Eye Unit, St Thomas's Hospital
  2. London SE1 7EH
  3. Kobler Centre, Chelsea and Westminster Hospital, London
  4. Medical Eye Unit, St Thomas's Hospital
  5. London SE1 7EH
  1. Medical Eye Unit, St Thomas's Hospital
  2. London SE1 7EH
  3. Kobler Centre, Chelsea and Westminster Hospital, London
  4. Medical Eye Unit, St Thomas's Hospital
  5. London SE1 7EH
  1. Dr E M Graham timljackson{at}

Statistics from

Editor,—Patients with previous cytomegalovirus retinitis (CMVR) have been observed to get vitritis, cystoid macular oedema (CMO), and epiretinal membranes after commencing HAART (highly active antiretroviral therapy).1-3 It is postulated that this is due to the improved CD4+ function that occurs with HAART.4

We present a patient with previously treated CMVR and immune recovery vitritis who developed further severe intraocular inflammation 6 weeks after changing HAART.


A 43 year old man, diagnosed as HIV positive in 1984, was treated in January 1996 for bilateral CMVR that responded to systemic ganciclovir.

In April 1996 he started saquinavir, stavudine, and lamivudine. A viral load was 18 000 copies ×106/l and CD4+ count 40 cells ×106/l. In December 1996 ritonavir was added.

In February 1997 he developed immune recovery vitritis in both eyes. Best corrected visual acuities (BCVA) were 6/6, N5 right; 6/6, N8 left. A viral load was now 7800 copies ×106/l and CD4+ count 100 cells ×106/l. A fluorescein angiogram showed bilateral CMO with macular ischaemia (Fig 1) and he was treated with 250 mg sustained release oral acetazolamide, twice daily. After 1 month's treatment there was a subjective improvement in vision but no objective change in VA. In September 1997 his BCVAs were 6/6, N5 right; 6/6, N6 left. In December 1997 his CMVR was inactive and ganciclovir was stopped.

Figure 1

(A) Fluorescein angiogram of the right eye at 38 seconds demonstrating macular ischaemia. (B) Macular oedema evident in a late phase fluorescein angiogram of the left eye.

In February 1998 he developed marked lipodystrophy and consequently was changed to efavirenz, hydroxyurea, didanosine, and continued on stavudine. By April 1998 his viral load had fallen to less than 500 copies ×106/l and his CD4+ count was 128 cells ×106/l (Fig 2).

Figure 2

CD4+ lymphocyte count and viral load.

Six weeks after changing HAART, he complained of blurred left vision. Visual acuity was 6/6 right, 6/9 left. There was a left vitritis with a white retinal lesion along the inferotemporal vessels at the margin of inactive CMVR (Fig 3). A diagnosis of relapse of CMVR was made and he was treated with intravenous ganciclovir.

Figure 3

Inferotemporal, white retinal lesion at margin of mature CMVR with surrounding haemorrhage, sheathing of neighbouring veins but no obvious involvement of arterioles.

Two weeks later vision was counting fingers, the vitritis was worse, and the retinal lesion had extended with more haemorrhage. He was treated for presumed toxoplasmosis with sulphadiazine, pyramethamine, and folinic acid but 1 week later his vision was hand movements and the lesion was larger (Fig 4). A diagnostic vitrectomy was performed. Polymerase chain reaction (PCR) was negative for CMV, toxoplasmosis, varicella zoster, and Epstein–Barr viruses. Cytology did not reveal lymphoma.

Figure 4

One week later the lesion has increased in size and extended to involve all retinal layers. Retinal vessels are obscured and haemorrhage persists.

One week later the eye was painful and vision was perception of light. A retinal biopsy was performed in an area of active inflammation. Histology showed retinal necrosis and exudates. No viral inclusion bodies were found and stains for acid fast bacilli, fungi, and toxoplasma cysts were all negative. A few atypical lymphocytes were seen and PCR for B cell clonality showed oligoclonal banding. A contrast enhanced magnetic resonance imaging of the brain and orbits was normal.

Eighteen months later the BCVAs are 6/6, N5 right; perception of light, left. A chorioretinal scar persists in the macular and temporal retina of the left eye but there is no active inflammation in either eye.


Investigations in this patient failed to identify any cause for the intraocular inflammation; however, the cessation of the inflammatory process in the absence of any specific treatment is similar to the clinical course of immune recovery vitritis and suggests a non-infectious mechanism. The temporal association with the change in HAART, together with the PCR findings, both imply an immunological cause. It is unclear why this response was uniocular.

The presence of multiple discrete bands on PCR may indicate a premalignant lymphoproliferation but 18 months later the eye is quiet with no recurrence and the patient remains well. Our hypothesis is that the intraocular inflammation occurred because the new combination HAART produced a second, very exaggerated immune response to pre-existing antigens.

Clinicians should be aware that changing the combination of HAART may be associated with aggressive intraocular inflammation with an oligoclonal lymphocyte response. Early recognition and treatment with systemic steroids may have prevented blindness in our patient.


Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.