Whitcup SM, Fortin E, Lindblad AS, Griffiths P, Metcalf JA, Robinson MR, Manischewitz J, Baird B, Perry C, Kidd IM, Vrabec T, Davey, Jr RT, Falloon J, Walker RE, Kovacs JA, Lane HC, Nussenblatt RB, Smith J, Masur H, Polis MA. Discontinuation of Anticytomegalovirus Therapy in Patients With HIV Infection and Cytomegalovirus Retinitis. JAMA. 1999;282(17):1633-1637. doi:10.1001/jama.282.17.1633
Author Affiliations: The Clinical Branch (Drs Whitcup and Smith and Ms Perry) and Laboratory of Immunology (Drs Fortin, Robinson, and Nussenblatt), National Eye Institute; Laboratory of Immunoregulation, The National Institute of Allergy and Infectious Diseases (Ms Metcalf and Drs Davey, Falloon, Walker, Lane, and Polis); and The Critical Care Medicine Department, Clinical Center, (Ms Baird and Drs Kovacs, and Masur), National Institutes of Health, Bethesda, Md; The EMMES Corporation, Potomac, Md (Dr Lindblad); Laboratory of Virology/Retrovirology, CBER, Food and Drug Administration, Rockville, Md (Ms Manischewitz); Departments of Virology, The Royal Free Hospital School of Medicine, London, England (Drs Griffiths and Kidd); and the Department of Ophthalmology, Wills Eye Hospital, Philadelphia, Pa (Dr Vrabec).
Context Persons with cytomegalovirus (CMV) retinitis and acquired immunodeficiency
syndrome (AIDS) have required lifelong anti-CMV therapy to prevent the progression
of retinal disease and subsequent loss of vision.
Objective To determine whether patients who were taking highly active antiretroviral
therapy (HAART) and who had stable CMV retinitis could safely discontinue
anti-CMV therapy without reactivation of their retinitis or increase in human
immunodeficiency virus (HIV) viral load.
Design Prospective nonrandomized interventional trial performed from July 1997
to August 1999.
Setting Clinical Center of the National Institutes of Health, Bethesda, Md.
Patients Fourteen patients with stable CMV retinitis and HIV infection and CD4+ cell counts higher than 0.15 × 109/L and being treated
with systemic anti-CMV medications and HAART.
Interventions Discontinuation of specific anti-CMV therapy.
Main Outcome Measures Reactivation of CMV retinitis, development of extraocular CMV infection,
detection of CMV in blood and urine, HIV burden, immunologic function, quality
of life, morbidity, and mortality.
Results Twelve (89.7%) of 14 patients had evidence of immune recovery uveitis
before anti-CMV drugs were discontinued. No patient had reactivation of CMV
retinitis or development of extraocular CMV disease during mean follow-up
of 16.4 months (range, 8.3-22.0 months) without anti-CMV therapy. Human immunodeficiency
viral load remained stable following cessation of anti-CMV medications. Blood
and urine assays for CMV were briefly positive in 9 patients but did not predict
reactivation of CMV disease. Worsening immune recovery uveitis was associated
with a substantial (>3 lines) vision loss in 3 patients.
Conclusions Maintenance anti-CMV medications were safely stopped in those patients
who had stable CMV retinitis and elevated CD4+ cell counts and
who were taking HAART. The study demonstrates that immune recovery following
potent antiretroviral therapy is effective in controlling a major opportunistic
infection, even in patients with a history of severe immunosuppression.
Cytomegalovirus (CMV) retinitis is the most common intraocular infection
in patients with the acquired immunodeficiency syndrome (AIDS). Prior to the
widespread use of potent antiretroviral therapy, CMV retinitis occurred in
up to 40% of patients during the course of their disease,1
usually when CD4+ cell counts fell below 0.10 × 109/L. Without therapy, CMV retinitis leads to diffuse retinal necrosis
and blindness, with a median time from detection of active CMV retinitis to
progression of disease of 2 to 3 weeks.2 Systemic
therapy slows the progression of retinitis only moderately to a median time
to progression with 2 months with intravenous ganciclovir sodium or foscarnet
sodium2,3 and is associated with
considerable toxic effects, inconvenience, and financial cost. Several case
series have suggested that CMV might not reactivate after stopping maintenance
anti-CMV therapy in some patients with AIDS who had elevated CD4+
cell counts and were taking combination antiretroviral therapy.4- 10
The purpose of this clinical trial was to investigate prospectively whether
anti-CMV therapy could be safely discontinued in patients who have some degree
of immune recovery while receiving highly active antiretroviral therapy (HAART).
The effect of discontinuing anti-CMV therapy on plasma human immunodeficiency
virus (HIV) viral load, the ability to detect CMV in the urine or blood, and
intraocular inflammation was also evaluated.
The study protocol complied with the principles of the Declaration of
Helsinki and was approved by the National Eye Institute Data and Safety Monitoring
Committee and the National Eye Institute Institutional Review Board. Informed
consent was obtained from all study participants. All patients continue to
be followed up until progression of CMV retinitis or death.
Eligible patients included women and men at least 18 years of age with
a diagnosis of CMV retinitis and HIV infection with a CD4+ cell
count higher than 0.15 × 109/L.11
Patients had inactive CMV retinitis that was not immediately sight-threatening
based on examination by 2 ophthalmologists and confirmed by centralized grading
of retinal photographs by an independent reading center. Retinitis was defined
as not immediately sight-threatening if located at
least 1000 µm away from the optic disc or fovea (periphery of zone 1)
or if the visual acuity in the affected eye was already severely reduced from
any cause (fewer than 19 letters as measured on an Early Treatment of Diabetic
Retinopathy Study [ETDRS] chart).12 Patients
had to be receiving systemic anti-CMV therapy with ganciclovir, foscarnet,
or cidofovir at maintenance doses. Patients receiving immunotherapy for AIDS,
such as interleukin 2, were assessed for eligibility at least 1 month after
the last infusion. Patients were excluded from the study if they had received
a sustained-release ganciclovir intravitreal implant.
Systemic anti-CMV medications were discontinued in enrolled patients.
Study visits were performed at baseline, every 2 weeks for the first 3 months
of the study, every 3 weeks from 3 to 6 months after baseline, and at least
every 4 weeks thereafter.
All study visits included a complete ophthalmologic examination, including
slit lamp biomicroscopy and dilated retinal examination, and 60° retinal
photographs graded at a centralized reading center (Fundus Photograph Reading
Center, University of Wisconsin, Madison). Best corrected visual acuity was
measured using logarithmic charts and a standardized protocol. Immune recovery
uveitis was defined as the presence of 2+/4+ vitritis associated with greater
than a 2-line loss in visual acuity or clinically documented macular edema.
A medical history and physical examination were performed at each visit.
Human immunodeficiency virus antibody using enzyme-linked immunosorbent
assay and Western blot were obtained at baseline. A quality of life assessment
using the National Eye Institute 25-Item Visual Function Questionnaire(VFQ-25,
Copyright 1996, RAND, Santa Monica, Calif) was performed at baseline and every
3 months thereafter. CD4+ cell counts on whole blood and HIV load
obtained on plasma were taken at baseline and every 1 to 2 months thereafter
using a branched DNA (bDNA) assay (Chiron, Emeryville, Calif). Cytomegalovirus
shell vial and standard cultures on blood were performed at each study visit.13 The presence of CMV in whole blood was detected and
quantified using nonnested polymerase chain reaction (PCR) to amplify 149
base pairs of the human cytomegalovirus glycoprotein B gene (gpUL55). Positive samples from the qualitative assay were quantitated
from 5 µL of DNA extracted from 200 µL of whole blood as previously
evaluated in clinical studies.14
The primary end point for the study was reactivation of CMV retinitis
determined by centralized grading of retinal photographs by an independent
reading center. Reactivation of CMV retinitis was defined as the time from
baseline when (1) a new retinal CMV lesion developed, (2) the border of an
existing CMV lesion reactivated as evidenced by opacification and whitening,
or (3) the border of any CMV lesion advanced by 750 µm over a 750-µm
front into previously normal-appearing retina. Secondary end points included
the development of extraocular CMV disease, detection of CMV by culture and
PCR, HIV burden, adverse events, quality of life measures, and mortality.
Fourteen patients were enrolled in this study between July 1997 and
August 1998 and were seen at more than 95% of the expected patient visits.
Patient demographics, duration of and treatment for HIV disease and CMV retinitis,
length of follow-up, and CD4+ cell counts and maximal HIV viral
loads are listed in Table 1.
Duration of HIV infection prior to study enrollment ranged from 1.8 years
to 13.2 years with a median of 6.3 years. All patients were taking at least
3 antiretroviral agents with at least 1 protease inhibitor. The length of
time between first CMV retinitis diagnosis and enrollment ranged from 1.1
years to 3.3 years (mean 1.9 years). The median CD4+ cell count
at the time of diagnosis of CMV retinitis was 0.025 × 109/L
(range, 0.0-0.11 cells × 109/L). All patients had CD4+ cell counts in excess of 0.15× 109/L at screening
for eligibility. The CD4+ cell counts at the time of discontinuation
of anti-CMV medications ranged from 0.08 to 1.3 × 109/L (median,
0.032 × 109/L). The CMV retinitis was uniocular in 8 patients;
3 patients had retinitis less than 1500 µm from the optic disc or less
than 3000 µm from the fovea. Twelve (85.7%) of 14 patients had evidence
of immune recovery uveitis in an eye with CMV retinitis prior to discontinuation
of anti-CMV medications, based on a decline in visual acuity in 10 patients
and macular edema in 2 patients.
Cytomegalovirus retinitis did not progress in any patient after a mean
of 16.4 months (range, 8.3-22.0 months) after stopping anti-CMV therapy. During
the study, 1 patient had an improvement in visual acuity of at least 15 letters.
Four patients had a decrease in visual acuity of at least 15 letters from
baseline in at least 1 eye on at least 2 consecutive visits. Three of these
4 patients had an increase in immune recovery uveitis characterized by worse
vitritis and/or macular edema; all had intraocular inflammation at the time
of entry into the study. Vision loss in the fourth patient followed removal
of silicone oil that was placed prior to study enrollment for a retinal detachment.
None of the visual loss was attributed to progression or reactivation of CMV.
There were no new retinal detachments in any patient. There was no significant
difference in the National Eye Institute's Visual Function Questionnaire composite
score at baseline (79) and at 12 months (73) (P>.25).
The score ranges from 0-100. The CD4+ cell counts significantly
increased over time (P<.001).
Nine patients were culture positive for CMV in their urine by both conventional
and shell vial specimens at various times during the study. Qualitative PCR
was intermittently positive in 10 patients, but the viral load was too low
for formal quantification in 9 of these 10 patients. The 10th patient had
a qualitatively positive CMV PCR at week 2 that persisted through week 32.
Quantitatively increasing CMV viral load was detected by PCR beginning at
week 8 with 5200 copies/mL and increased to 15,600 copies/mL by week 12 but
then decreased to levels below detection at weeks 15 through 21.
Standard therapy for CMV retinitis in persons with HIV infection has
required lifelong treatment with systemic ganciclovir, foscarnet, or cidofovir
or with an intravitreal ganciclovir implant. In this prospective cohort study,
all 14 patients with AIDS, inactive CMV retinitis, and CD4+ cell
counts higher than 0.15 × 109/L were able to discontinue
maintenance anti-CMV therapy without progression of retinitis after a mean
follow-up of 16.4 months and a minimum follow-up of 8.3 months. Benefits of
stopping anti-CMV medications include avoiding toxic effects associated with
treatment and decreasing the cost of treatment.
Previous studies have shown that activation of the immune response in
patients with AIDS caused by opportunistic infections such as CMV could lead
to increased HIV replication.15 In this study,
discontinuation of maintenance anti-CMV therapy causes no significant changes
in plasma HIV.
Potent combination antiretroviral therapy has decreased the incidence
of CMV retinitis16 and altered its clinical
presentation. Although mild vitreous inflammation (vitritis) was reported
in patients with severe immunosuppression and with CMV retinitis,17 more profound inflammation is now occurring in patients
with inactive CMV retinitis receiving HAART.6,8,16,18- 21
This immune recovery uveitis is characterized by significant vitritis, increased
permeability and leakage of the central retinal blood vessels causing retinal
edema (cystoid macular edema), fibrotic tissue adherent to the retina (epiretinal
membranes), and swelling of the optic nerve head (disc edema). Immune recovery
uveitis was documented in 12 of 14 patients at baseline, and substantially
increased in 3 patients during the course of the trial. Although there is
variability in the occurrence of immune recovery uveitis reported in the literature,
other investigators have reported immune recovery uveitis in about 90% of
patients with CMV retinitis on HAART.22 The
reason for the high prevalence of immune recovery uveitis in this study remains
unclear, but it may be related to the strict adherence and a strong response
to antiretroviral therapy. The median CD4+ cell count was above
0.30 × 109/L at baseline. Since immune recovery uveitis has
been reported in patients with increasing CD4+ cell counts who
continue to receive anti-CMV medications,19
larger studies will be needed to determine if stopping anti-CMV medications
exacerbates the uveitis.
No new retinal detachments occurred during the study, although the risk
in a similar population before the use of combination antiretroviral therapy
was 19% at 6 months.23 Intraocular inflammation
and hypertrophy of the retinal pigment epithelium may be associated with increased
adherence of the retina and a decreased risk of retinal detachment.
In this study, maintenance anti-CMV therapy was stopped in patients
with CD4+ cell counts higher than 0.15 × 109/L
receiving HAART and with CMV retinitis that was not immediately sight threatening.
Stopping anti-CMV medications in patients with immediate sight-threatening
retinitis should only be considered in patients who can have frequent examinations.
All patients in this trial were receiving HAART for at least 4 months prior
to stopping anti-CMV medications. Since patients may not have adequate immunologic
protection during the first several weeks while receiving HAART as CD4+ cell counts are rising,24 we cannot
make recommendations on stopping anti-CMV medications in patients who have
recently started HAART.
Although this study lacks a control group, absence of progression of
retinitis in any of the 14 patients is significant since other well-controlled
randomized clinical trials prior to the use of HAART report a median time
to CMV retinitis progression of 3 weeks in patients without specific anti-CMV
therapy25- 27 and
2 months in patients taking specific anti-CMV therapy and with CD4+
cell counts lower than 0.050 × 109/L.2,28
The lack of reactivation of retinitis or development of extraocular CMV disease
in our patients is likely due to improved immunologic control of CMV infection.
Because of the clonal nature of the antigen-specific immune response,
there was concern that increased T-cell numbers following potent antiretroviral
therapy may not restore adequate protection against opportunistic pathogens
such as CMV. This study provides clinical proof that immune recovery in patients
receiving HAART is effective in controlling a major opportunistic infection,
even among those with a history of severe immunosuppression.