In this study, we evaluated the background of patients who developed optic neuropathy during voriconazole treatment. Well-known symptoms of visual disturbances caused by voriconazole include foggy vision, photophobia, and color blindness, which are reported to occur within about two weeks after starting treatment with voriconazole and often disappear spontaneously in about a week.9 However, optic neuropathy, which, necessitates the discontinuation of voriconazole, is rarely observed. In the present study, optic neuropathy appeared after 72–292 days of voriconazole use, and not only in the early stages of treatment.
It has been reported that voriconazole side effects are related to trough value and increase at trough concentrations greater than 5.5–6.0 µg/ml.10,11 While there are reports that visual symptoms tend to appear at trough levels of voriconazole of 4.2 µg/ml or higher,12 there are also contradictory reports that suggest they are not related to blood concentration.13 In our cases, the trough value at the onset of vision loss was 0.59–2.37 µg/ml, which is lower than 4.2 µg/ml, and no relationship between trough value and side effects was demonstrated.
Five of the six cases were treated with voriconazole and ethambutol, and it was necessary to consider both drugs as potential causes of optic neuropathy. Most of these patients using voriconazole also used ethambutol. About 33% of patients with chronic pulmonary aspergillosis treated with antifungal agents in our hospital were patients with chronic pulmonary aspergillosis complicated non-tuberculous mycobacteriosis, which in recent years has been reported to have a worse prognosis than non-tuberculous mycobacteriosis alone.14,15 In fact, all five patients were using ethambutol for treating pulmonary non-tuberculous mycobacteriosis. It has been reported that ethambutol-induced optic neuropathy occurs 1–12 months after the start of ethambutol treatment.16,17 In all five cases of ethambutol use in our study, ethambutol was used for more than 22 months, a median duration of 70 months, which is much longer than the time linked to ethambutol-induced optic neuropathy, suggesting that voriconazole is associated with the optic neuropathy in these patients. In patients with non-tuberculous mycobacteriosis-associated aspergillosis, voriconazole-associated optic neuropathy may be masked by ethambutol. In reality, discontinuation of ethambutol is inevitable upon the appearance of vision loss, but ethambutol is an important drug in the treatment of non-tuberculous mycobacterial infections to prevent macrolide resistance.18,19 Therefore, failure to reinstate treatment with ethambutol may have a seriously negative impact on treatment. If voriconazole is strongly suspected as the cause of optic neuropathy in a patient with non-tuberculous mycobacteriosis, re-administration of ethambutol may be considered after visual recovery.
Although there were many cases of concomitant use of ethambutol and voriconazole, case 2 in the present study and two of the three cases (Table 1, Table 2) reported in the literature featured optic neuropathy when voriconazole was used alone. The Pharmaceuticals and Medical Devices Agency, which provides information on adverse drug reactions in Japan, reported 33 cases of vision loss as an adverse reaction to voriconazole and included only three cases related to ethambutol use.20 French pharmacovigilance centers reported six cases of voriconazole-associated toxic optic neuropathy, three of which were related to voriconazole alone.7 Although voriconazole is not listed in reviews as the causative agent of toxic optic neuropathy, several cases of optic neuropathy due to voriconazole alone have been observed, and physicians prescribing voriconazole should be aware of this fact.21
In addition to our six cases, three other cases (Table 1, Table 2) of voriconazole-associated optic neuropathy have been reported in other studies. Since neither the amount of voriconazole used per body weight nor the blood concentrations have been included in previous reports, it is significant that those indicators were found to be less relevant to optic neuropathy in the present study. This study used the CFF test in addition to visual acuity tests in the diagnosis of optic neuropathy and evaluated optic nerve function more accurately.
One report indicates that the blood concentration of voriconazole increased with concomitant use of ethambutol, inducing vision loss.7 Voriconazole is primarily metabolized by Cytochrome P450 (CYP) 2C19 and partly by CYP 3A4 and CYP 2C9. Ethambutol moderately inhibits CYP 2C19,22 while clarithromycin and erythromycin inhibit CYP 3A4. Therefore, these drugs may cause an increase in the blood concentration of voriconazole.23 In the present study, macrolide antibiotics were used in combination in cases 1, 2, 4, and 5. CYP 2C19, the main metabolizing enzyme of voriconazole, is a polymorphic gene, and blood levels of voriconazole are known to vary among individuals, but it has been reported that blood levels of voriconazole also fluctuate within individuals, for example, increasing during severe inflammatory episode.24,25 In our cases, the blood concentration of voriconazole was normal at the time of onset of optic neuropathy, but there was an increase in blood concentration during treatment, which may have triggered the optic neuropathy.
Numbness of the lower extremities was observed in four of our six cases and in one of three cases from previous reports. Peripheral neuropathy can be caused by drugs, diabetes mellitus, vitamin deficiency, folic acid deficiency, paraproteinemia, autoimmune diseases, renal failure, and hypothyroidism. Triazole antifungal agents are reported to cause axonal neuropathy with sensory predominance, and peripheral neuropathy is reported to occur in 9% of patients treated with voriconazole.26 The blood levels of voriconazole were reported as normal at the time when voriconazole-associated peripheral neuropathy appeared,26 matching this study’s findings. Among our four cases of peripheral neuropathy, case 3 had a history of Hashimoto's disease, but thyroid function was in the normal range. Although ethambutol is considered an unlikely causative agent of peripheral neuropathy,27–29 its involvement could not be ruled out. However, in three of our four cases, peripheral neuropathy appeared at the same time as optic neuropathy, which might be associated with voriconazole.
Five out of six cases in our study, and all three cases in a previous report, showed recovery of visual acuity after discontinuation of voriconazole. The time from drug discontinuation to visual acuity improvement varied from 1 to 20 months in this study. In six cases of voriconazole-associated optic neuropathy reported in French pharmacovigilance centers, one patient recovered visual acuity, three did not improve, and two had an unknown follow-up.7 Because the mechanism of voriconazole-associated optic neuropathy is unknown, voriconazole is often discontinued when optic neuropathy is observed, as is the case with ethambutol, in order to avoid irreversible vision loss.
Although the mechanism of voriconazole-associated optic neuropathy is currently unknown, it has been reported that in patients with voriconazole-associated visual impairment the b-wave is attenuated in electroretinograms. This is similar to complete congenital stationary night blindness (cCSNB).30 Since the b-wave is composed of ON-bipolar cell neurotransmission, the impairment of ON-bipolar cells may constitute a factor in voriconazole-associated optic neuropathy. However, symptoms of visual impairment caused by cCSNB differed from those related to voriconazole treatment. CCSNB can cause night blindness, but rarely causes photophobia or optic neuropathy. Furthermore, night blindness was present in only one of the six cases in this study. The differences in symptoms suggest that these diseases are not the same, but it may be possible to determine whether voriconazole is the cause of optic neuropathy by using electroretinograms. Further research in this field is needed.
This study was subject to several limitations. The research was conducted only at a single center, and the number of patients was small. The total number of cases, including those reported previously, was only nine, and a greater sample size would be desirable. To select cases, we checked the medical records for voriconazole use and optic neuropathy, but a selection bias may have been present since optic neuropathy may not have been on the list of diseases if the patients were diagnosed at other hospitals. Further, since some of our cases were concomitantly treated with ethambutol, we were not able to make a strict diagnosis of optic neuropathy associated with voriconazole, and there is a need to investigate the possibility of optic neuropathy associated with drug interaction. Finally, although CYP 2C19 is polymorphic, we did not examine this aspect in our study nor consider the evaluation of blood concentration by the metabolizing enzyme.