|
Ophthalmologic side effects during
alpha-interferon therapy for viral hepatitis
Journal of Hepatology Jan 2006
Louis d'Alterochea, Samuel Majzoubb,
Anne-Isabelle Lecuyerc, Marie-Paule Delplaceb,
Yannick Bacqa
ABSTRACTS
Background/Aims: Ophthalmologic side effects
have been reported during interferon therapy,
particularly retinal lesions and neurovisual
impairment. The aim of this prospective study
was to assess the nature and the frequency of
such lesions during alpha-interferon therapy for
viral hepatitis.
Methods: Between 1995 and 2003, 156 patients
treated with standard or pegylated
alpha-interferon, with or without ribavirin, had
a regular ophthalmologic examination before and
during treatment. No patient had signs of
retinopathy before treatment.
Cotton-wool spots were found in 31 patients and
retinal hemorrhage in nine patients during
treatment (24% of patients). These lesions
remained asymptomatic and disappeared in all
patients. A previous history of arterial
hypertension (RR 4.60, 95% CI 1.95-10.85), age
above 45 years (RR 2.80, 95% CI 1.36-5.85), and
use of pegylated alpha-interferon (RR 2.75, 95%
CI 1.41-5.38) were significantly associated with
retinopathy. Neurovisual impairment was present
in 31 patients (20%) before treatment and in 74
patients (47%) during treatment.
In studies including a significant number of
diabetic patients [5,7] diabetes mellitus has
also been associated with retinopathy.
Furthermore, improvement of retinopathy is
delayed in hypertensive and diabetic patients
after ending treatment [7]. This relationship
emphasizes that IFN-induced retinopathy can
result from physiopathological mechanisms in
common with other retinopathies related to
microvascular abnormalities.
Conclusions: In conclusion, this study showed
that signs of retinopathy and neurovisual
impairment were common in patients receiving
alpha-interferon therapy but were rarely
symptomatic. It suggests that alpha-interferon
may usually be continued in asymptomatic
patients as long as there is careful fundoscopic
examination.
….considering the high frequency of
ophthalmologic damage and the risk of
potentially severe symptomatic ocular lesions
under IFN therapy, we recommend performing
regular fundoscopic examinations in all
patients, particularly during the first three
months of treatment. In cases of abnormalities,
the treatment may usually be continued,
providing that there are no symptoms and there
is close ophthalmologic supervision. Reduction
in doses of IFN may be necessary, especially
when other risk factors for retinopathy are
present.
Introduction
Alpha-interferon (alpha-IFN) is widely used in
the treatment of viral hepatitis, either alone
or combined with other antiviral therapies such
as ribavirin [1]. It can induce various side
effects, particularly systemic effects such as
fever, influenza-like symptoms, thrombocytopenia
and neutropenia [2]. Ophthalmologic side effects
were first reported in 1990 by Ikebe et al. in
the form of retinal injury [3]. IFN-induced
retinopathy was thereafter described as an
association of cotton-wool spots, retinal
hemorrhage and micro aneurysms [4-8]. The
incidence varies according to country and study
design, higher frequencies being found in Japan
where values of more than 50% are common [9]. On
the other hand, several authors have reported
optic tract neuropathy [10-13], thus emphasizing
the possibility that patients treated for
hepatitis may develop neurovisual impairment
(NVI). Isolated case reports of severe
ophthalmologic complications have also been
reported under IFN therapy such as acute
exophthalmia [14], sub-conjunctival hemorrhage
[4], papilledema [15], retinal artery occlusion
[12] and retinal vein thrombosis [16]. The aim
of this prospective study was to assess the
nature and frequency of ophthalmologic side
effects, particularly retinopathy and NVI during
alpha-IFN therapy for viral hepatitis. We also
report the management of these ophthalmologic
side effects and their outcome.
Results
The 156 patients had a median of 4
ophthalmologic examinations (range 2-12). No
patient experienced symptoms except for mild
non-specific ocular tiredness in certain
patients.
Ophthalmologic examinations before treatment
No abnormalities were found on slit lamp
microscopy, fundoscopic examination, corrected
visual acuity or color vision before treatment.
By contrast, 31 patients (20%) had NVI (19
increased VER, 11 VF abnormalities and 7
contrast abnormalities). The mean VER C15 was
118±6ms (92-129ms) without NVI compared to
138±7ms (131-158ms) with NVI, and the VER C60
was 111±6ms (88-120ms) without NVI compared to
126±6ms (121-140ms) with NVI. Four patients had
increased VER on both checks.
Ophthalmologic examinations during treatment
During treatment, signs of retinopathy were
found on fundoscopic examination in 38 patients
(24%) (isolated cotton-wool spots in 29
patients, isolated retinal hemorrhage in 7
patients and both lesions in 2 patients). No
micro aneurysms occurred. The retinopathy was
unilateral in 28 cases (74%) and bilateral in 10
cases. The mean interval to appearance of
lesions was 13±12 weeks (range 5-56).
Retinopathy was discovered at the time of the
first examination in 26 cases (68%).
Signs of NVI were found in 74 patients (47%),
either occurrence (n=59) or increased (n=15).
Relative deficits in central vision on visual
fields and decrease in contrast sensitivity were
observed in 11 and 12 patients, respectively.
VER were increased in 59 patients (up to the
normal level in 13 patients, up to 10ms compared
with the initial value in 23 patients, and both
abnormalities in 23 patients). VER C15 and C60
increased from 119±7ms before treatment to
123±6ms (121±5ms without NVI vs. 128±6ms with
NVI) and from 112±6 to 116±6ms (114±5ms without
NVI vs. 120±5ms with NVI), respectively. The
mean interval to appearance of NVI was 24±10
weeks (range 10-52 weeks).
Risk factors
Risk factors for retinopathy on univariate
analysis are indicated on Table 2. Only patient
age, a previous history of arterial hypertension
and the use of pegylated IFN were significantly
associated with a risk of retinopathy on
multivariate analysis. Relative risks reached
4.60 (95% CI 1.95-10.85; beta coefficient:
1.526) for previous history of hypertension,
2.80 (95% CI 1.36-5.85; beta coefficient: 1.037)
for age above 45 years and 2.75 (95% CI
1.41-5.38; beta coefficient: 1.013) for use of
pegylated IFN.
Risk factors for NVI on univariate analysis
adjusted on gender were limited to patient age
only for women. The mean age of men with NVI was
41.5±8.5 years compared to 38.5±9.6 years
without NVI (P=0.11) and 48.7±9.7 years compared
to 42.9±11.5 years (P=0.029) for women. The
relative risk on multivariate analysis for
patients older than 45 years was 2.45 (95% CI
1.52-3.96).
Management of ophthalmologic abnormalities
Mild retinal hemorrhage was discovered the day
after ending treatment in one patient. The
treatment was immediately stopped in 3 patients
with minimal histological liver lesions, and
cotton-wool spots regressed in 15 days in all 3
patients. Dosage of IFN was reduced by half
until the end of treatment in 27 patients: IFN
was stopped 15 days later because of aggravation
of cotton-wool spots and retinal hemorrhage in 2
patients. Cotton-wool spots regressed in 15 days
in 12 patients and in two months in 13 patients.
IFN was subsequently withdrawn in one patient
because of the reappearance of cotton-wool spots
with retinal hemorrhage. IFN dosage was
maintained in 7 patients: mild hemorrhage
without cotton-wool spot regressed in 15 days in
2 patients, mild hemorrhage and cotton-wool
spots remained stable until the end of treatment
then regressed in 15 days in 4 patients, and 1
patient had cotton-wool spots with secondary
aggravation justifying the withdrawal of
treatment. Finally, retinopathy necessitated
withdrawal of treatment in 4 of 38 patients (2
men aged 55 and 60 years and 2 women aged 56 and
57 years, 3 treated for hepatitis C and 1 for
hepatitis B). Two patients had arterial
hypertension, and one had hypertension
associated with diabetes mellitus. All these
lesions disappeared after ending the treatment.
Because the ophthalmologic surveillance was
ceased on normalization of fundoscopic
examination in all cases of NVI, the evolution
of NVI is known in only 33 of 74 patients: NVI
normalized during treatment in 17 patients
(52%), and after stopping treatment in 15
patients (45%). NVI persisted for two months
after ending treatment in one patient.
Recurrence of ophthalmologic abnormalities
during subsequent treatment
After complete ophthalmologic recovery, 19
relapsers or non-responders to antiviral
treatment were treated again. Of the 3 patients
having signs of retinopathy during the initial
treatment, one had recurrence of cotton-wool
spots with retinal hemorrhage. By contrast, 2 of
the 16 patients without signs of retinopathy
during the initial treatment had cotton-wool
spots. Of the 6 patients with NVI during the
initial treatment, 4 patients had a recurrence
of increased VER. By contrast, of the 13
patients without NVI during the initial
treatment, 2 had impairment of visual fields and
2 had increased VER.
Discussion
Systematic three monthly ophthalmologic
follow-up revealed signs of retinopathy in 24%
of this large series of consecutive patients.
These lesions were clearly associated with IFN.
We did not observe any signs of retinopathy
before treatment as reported in recently
published studies [4,6-8,10]. Moreover, in other
indications for IFN therapy, such as in the
large clinical trial testing alpha-IFN in
age-related macular degeneration,
interferon-induced retinopathy was found in 5%
of patients but was not observed in the placebo
group [17]. The frequency of retinopathy
reported during IFN therapy for viral hepatitis
varies from 18 to 86% [5,7,9], depending on the
study design, especially in those including a
high induction dose of IFN [4,5], and on the
frequency of fundus examinations. Indeed,
cotton-wool spots can appear in the first weeks
of treatment [5,8] and can spontaneously
disappear thereafter [5-8].
We found that patient age, arterial hypertension
and the use of pegylated-IFN were risk factors
for retinopathy. Age and arterial hypertension
are well-defined risk factors for retinopathy,
both induced [5,7] and not induced [18] by
alpha-IFN. In studies including a significant
number of diabetic patients [5,7] diabetes
mellitus has also been associated with
retinopathy. Furthermore, improvement of
retinopathy is delayed in hypertensive and
diabetic patients after ending treatment [7].
This relationship emphasizes that IFN-induced
retinopathy can result from physiopathological
mechanisms in common with other retinopathies
related to microvascular abnormalities. This is
the first time to our knowledge that a study has
evaluated the impact of pegylated alpha-IFN on
induced retinopathy. More than twice the number
of cases of retinopathy occurred with pegylated
alpha-IFN than with standard alpha-IFN. The
doses of standard alpha-IFN and pegylated
alpha-IFN are not comparable and it is therefore
not possible to attribute causality to the doses
used. We found no relationship between dosage of
standard alpha-IFN and the frequency of
retinopathy.
Retinal damage was not symptomatic and always
resolved without sequelae. By contrast, in some
case reports cotton-wool spots (indicating a
precapillary arteriolar occlusion) were
symptomatic [19] or associated with other
symptomatic ischemic signs of retinopathy such
as papilledema [15], retinal artery occlusion
[12] and retinal vein thrombosis [16], and were
sometimes responsible for a definitive decrease
in visual acuity [19-21]. The frequency of
symptomatic ophthalmologic adverse events was
evaluated at 0.4% in a series including 4800
patients [22]. In this large series, the most
common findings on ophthalmologic examination
were cotton-wool spots, intraretinal hemorrhage
and papilledema. Of the 16 cases with an
available diagnosis for the ophthalmologic
injury, IFN-induced retinopathy was observed in
8 patients, retinal artery thrombosis in 2
patients and retinal vein occlusion, optic nerve
infarct, optic neuritis, presbyopia, retinal
detachment and focal vasculitic event involving
the retina in isolated cases [22]. Thus, though
ophthalmologic changes are relatively common,
serious ophthalmologic complications remain
uncommon.
We chose to reduce the dosage of IFN in more
than two thirds of cases of retinopathy. The
lesions regressed and treatment was completed in
most patients. We have recently preferred to
maintain the initial dosage and treatment was
completed in 6 of 7 patients. As at the
beginning of our experience, few authors
withdrew treatment in the presence of
asymptomatic lesions [4,8]. Some authors chose
to continue with reduced dosages [6] and
complete regression of retinopathy has been
described after completing treatment. In fact,
continuing IFN without reducing dosage is more
often preferred [5,7,8], with the risk of having
to stop treatment in a few cases for secondary
aggravation of the lesions [5]. We had to stop
treatment in one third of patients having other
risk factors for retinopathy (age, arterial
hypertension and diabetes mellitus). The choice
between withdrawing IFN or not should take into
account the presence of other risk factors for
retinopathy, the evolution of the retinopathy,
and the severity of the liver disease. In our
experience, the recurrence of cotton-wool spots
during subsequent treatment was not the rule and
several treatments were completed under close
ophthalmologic surveillance.
Using complete functional ophthalmologic
explorations, we found signs of NVI in 20% of
patients before treatment and 47% during
treatment. NVI was thus the most frequently
observed ophthalmologic injury.
Electrophysiological examinations assess the
entire neuronal tractus for visual information,
including the retina. However, Manesis et al.,
who first reported NVI [10], considered that the
site of toxicity is probably located beyond the
retina because of the lack of fundoscopic
findings and the normal electroretinograms. No
patients with NVI in our study had retinopathy
before treatment and it occurred in only a few
patients during treatment, thus emphasizing the
potential direct neurotoxic mechanism. Moreover,
ischemic retinopathy and NVI did not occur in
the same patients or during the same period. The
median interval for appearance of NVI in our
study was 24 weeks and nearly half that for
retinopathy. In other studies, retinopathy was
also observed within the first weeks of
treatment [5,8], whereas NVI appeared
progressively and was observed twice as often at
the end of treatment than half way through
treatment [23]. However, numerous questions
remain concerning the physiopathology, the parts
of the neural pathway affected, and the clinical
significance of NVI. We found only patient age
to be a risk factor for NVI. Others have
reported that NVI is more frequent during
treatment of hepatitis B than hepatitis C [24].
They had no explanation for this phenomenon and
the over-representation of cases of hepatitis B
in their series might have induced a statistical
bias. As for retinopathy, all cases of NVI that
we observed remained asymptomatic. Manesis et
al. also emphasized the usual lack of clinical
significance of NVI, although one third of
patients still had NVI more than 9 months after
ending treatment [24]. Thus, although some cases
of optic neuritis with visual field damage have
been reported [10-12], we believe that
systematic visual neurophysiologic evaluation
should not be recommended, particularly because
of their unwieldy nature. Indeed, these
examinations have no significant clinical impact
in the absence of symptoms. By contrast, such
examinations are essential for symptomatic
patients to complement fundoscopic evaluation
and assist in decision-making.
In conclusion, considering the high frequency
of ophthalmologic damage and the risk of
potentially severe symptomatic ocular lesions
under IFN therapy, we recommend performing
regular fundoscopic examinations in all
patients, particularly during the first three
months of treatment. In cases of
abnormalities, the treatment may usually be
continued, providing that there are no symptoms
and there is close ophthalmologic supervision.
Reduction in doses of IFN may be necessary,
especially when other risk factors for
retinopathy are present.
Patients and methods
Protocol design
Since 1995, we have offered ophthalmologic
surveillance for all our patients treated with
alpha-IFN for viral hepatitis. Between October
1995 and January 2003, 170 consecutive patients
suffering from viral hepatitis and treated by
two of the authors (YB and LA) with IFN had a
pre-treatment examination and agreed to regular
ophthalmologic surveillance. Five patients were
excluded because treatment was withdrawn within
the first weeks of therapy, and nine other
patients were excluded because of non-compliance
with the ophthalmologic surveillance (Table 1).
Finally 156 patients were included in the
present study. The ophthalmologic assessment was
performed before beginning treatment, then every
three months until the end of treatment.
Patients
Mean age was 42±10 years (range 18-70 years),
and 54 patients were female and 102 were male.
Twelve patients had chronic hepatitis B and 144
had hepatitis C (142 chronic hepatitis and 2
acute hepatitis) (Table 1). Histological data
were available for all patients except for 3
hemophilic patients, 1 with Willebrand's
disease, 2 with acute hepatitis C and 2 patients
with hepatitis C-related extra-hepatic
manifestations (1 cryoglobulinemia and 1
polyneuropathy). Cirrhosis was present in 13
patients. None was HIV-positive. Ten patients
had arterial hypertension and another had
diabetes mellitus and arterial hypertension.
Modalities of alpha-IFN treatment
Types and dosage of IFN varied according to the
general recommendations and protocols for
patients included in clinical trials. For the 12
patients with hepatitis B, alpha 2a-IFN (8
patients) or alpha 2b-IFN (4 patients) were used
at 3-8 MIU 3 times a week for 3-17 months
(median duration 6 months). Fifty-four of the
144 patients with hepatitis C were treated with
alpha-IFN alone (35 alpha 2a-IFN and 19 alpha
2b-IFN) and 91 with alpha-IFN combined with
ribavirin (3 alpha 2a-IFN, 48 alpha 2b-IFN, 4
pegylated alpha 2a-IFN and 35 pegylated alpha
2b-IFN). The initial dosage of standard
alpha-IFN varied from 1.5 MIU, 3 times a week,
to 10 MIU daily, and 180ìg of alpha 2a-pegylated
IFN or 1.5ìg per kg per week for alpha
2b-pegylated IFN. The duration of treatment
varied from 2 to 18 months (median 11 months).
Dosages of ribavirin varied from 800 to 1200mg
according to patient weight. It was the first
course of treatment for 129 patients, the second
for 24 and the third for 3 patients.
Ophthalmologic assessments
Ophthalmologic examinations consisted of visual
acuity testing (on the Monoyer scale), slit lamp
microscopy, fundoscopic examination with a
mydriatic agent, and neurophysiologic
assessments including standardized visual fields
(stat 94 protocol Metrovision monitor: central
and pericentral static visual field at 94 points
located at ±30 central degrees at a
suprathreshold level, with initial research of
foveolar threshold and basic level at 4 points),
pattern reversal visual evoked responses (VER),
color vision (15 HUE desaturated test), and
contrast sensitivity. Visual fields were
considered to be abnormal for losses more than
4dB of the basic level, losses at 2 contiguous
points of more than 4 dB or a decreased foveolar
threshold. VER were measured with pattern
stimulation generated with a high-resolution
vision monitor (Metrovision). Pattern contrast
was 98%, and the reversal period was 1800ms. The
waveform analyzed was obtained by averaging the
sum of 30-50 responses. The P100 implicit time
for 15-min checks (C15) and 60-min checks (C60)
were measured. Under these conditions, the
normal levels for our unit were 120±10ms for C15
and 110±10ms for C60. VER were considered to be
abnormal for values above the normal and for
increases in value >10ms compared to initial
values. Color vision was considered abnormal for
scores higher than the age average +2 SD.
Contrast sensitivity was studied on spatial
frequencies between 0.8 and 25.6 cycles per
degree with a static pattern and a 10Hz temporal
modulated pattern. Contrast sensitivity was
considered abnormal for decreases at medium and
low frequencies. Abnormalities of fundoscopic
examination defined retinopathy, and
abnormalities of visual fields, VER, color
vision and contrast sensitivity were grouped
together in the NVI group.
2.5. Approach to retinal abnormalities
In cases of ophthalmologic abnormalities,
treatment was either stopped or reduced or
maintained at the initial dosage after
discussion between the hepatologist and the
ophthalmologist. The therapeutic decision was
made according to the benefit/risk ratio,
comparing the severity of the liver disease and
evolution of the retinopathy. In such cases,
ophthalmologic examination was performed every
two or three weeks until normalization of
fundoscopic examination.
Statistical analysis
The database was set up on Excel software then
analyzed on SAS software using the ÷2-test,
MantelHaenszel weighted ÷2-test, Fischer's exact
test and ANOVA for univariate analysis and a Cox
regression model for multivariate analysis. All
values are expressed as mean±SD with range in
brackets or as median. Risk factors are
expressed as relative risk (RR, 95% confidence
intervals, beta coefficients). A P-value <0.05
was considered statistically significant.
|
