|  Fig. 1 Flow diagram showing all the patients of acute myeloid leukemia treated in the study period and their outcomes with respect to development of invasive fungal infections

The median age of the study population was 17 years (range 1-75 years) with a male:female ratio of 1.6:1 [Table 1]. Most patients (87%) received daunorubicin + cytarabine-based induction therapy.

Table 1

Efficacy and toxicity of voriconazole

ppp fungal infection developed in 5/75 (6.6%) patients. Two of these had proven IFI (blood culture isolated candidiasis, one had Candida krusei, and one had Candida glabrata). Three others had possible IFI based on clinical and radiological features [Figure 2 and Supplementary Table 1].

|  Fig. 2 Lung nodule showing with halo sign (a) and bilateral opacities with cavitations (b)

Supplementary Table 1

Characteristics of patients who developed invasive fungal infection (n = 5) while on voriconazole prophylaxis

Twenty-one patients started empiric anti-fungal therapy. Hence, a total of 26/75 (34%) patients started therapeutic (empiric + pppIFI) antifungals. The mean time to start of therapeutic antifungals was 16 days. The mortality due to IFI was 1/75 while 7/75 patients died due to other causes [Supplementary Table 2].

Supplementary Table 2

Characteristics of patients who expired (n = 8)

On comparing the outcomes between patients receiving voriconazole and fluconazole, it was found that patients treated with voriconazole had a significantly lesser incidence of pppIFI and lesser need for empiric anti-fungal treatment and also had a much longer time to initiate therapeutic antifungal drugs [Table 2].

Table 2

DISCUSSION

This is one of the few studies prospectively evaluating the efficacy of voriconazole as AFP during induction therapy of AML. Voriconazole was a well-tolerated and extremely effective as a prophylactic agent and reduced the incidence of IFI to 6.6% (from 25% in an earlier cohort who received fluconazole). The incidence of pppIFI was only 6.6%. In addition, voriconazole also reduced the number of deaths due to IFI, need of empirical antifungal treatment and delayed the start of anti-fungal treatment.

Three other studies [Table 3] have reported on the use of voriconazole as an antifungal prophylactic agent.[25,26,27] The incidence of IFI ranged from 0% to 4% in these studies. We saw an incidence of 6.6% in our patients which is comparable to that reported by Mandhaniya et al.[25] and Chabrol et al.[26] but higher than that reported by Vehreschild et al. However the study by Vehreschild et al.[27] had very small number of patients which limits interpretation. On comparing secondary end points, the need for empirical antifungal treatment was 28% in our study that was 17% and 22% in studies by Chabrol et al. and Mandhaniya et al. respectively, which is comparable.[25,26]

Table 3

The incidence of IFI with voriconazole among our patients appears to be slightly higher than the remarkable 2% reported by Cornely et al. in the registration trial of posaconazole.[18] However, cross-trial comparisons can be fallacious because of the variability in end points and the variable settings in which they are performed. Cornely et al. considered only “proven” and “probable” cases as IFI, whereas in our study proven, “possible” cases were also included. Moreover, the study by Cornely et al. was conducted in a Western environment with a temperate weather where the incidence of IFI could be much lesser compared to the tropical conditions in India.[28,29] In fact, the incidence of IFI in the study by Cornely et al. in the control arm using fluconazole was only 6% (compared to 25% with the use of fluconazole at our center).

An important concern with voriconazole was its potential for toxicity and drug interactions. However, most of the medicines used during AML therapy (including many antibacterials) don’t have clinically significant interactions with voriconazole, and this is well-established as a therapeutic agent in these patients. We found that voriconazole was very well-tolerated, and most of the side effects were mild and as expected from earlier reports.[30,31,32,33] Drug-related serious adverse events leading to discontinuation of voriconazole prophylaxis was 4% in our study while in the study by Mandhaniya et al. it was 2% and 14% in the study by Chabrol et al.[25,26] Serious adverse events possibly or probably related to posaconazole have been reported in up to 6% of patients.[18]

CONCLUSION

Our patients did not have access to serological tests that could have led to underdiagnosis of “probable” IFI. Furthermore, HRCT scans at baseline were performed routinely in the voriconazole cohort while they were performed only in the symptomatic patients in the historical controls. This may have led to these patients receiving AFP rather than therapy. However, we had a proactive approach to performing diagnostic high-resolution computed tomography scans of the chest and nasal sinuses to enable early diagnosis of fungal infections. We did not monitor voriconazole drug level that is recommended by some authors.[34,35] Though the comparator cohort receiving fluconazole was historical, both groups were comparable and received similar treatment and supportive care medicines, making the comparisons relevant. This is one of the few studies to clearly show the efficacy and safety of voriconazole as a prophylactic agent during AML therapy and clearly shows its superiority versus fluconazole. Only a randomized trial can answer the question of whether it can be equivalent to posaconazole. However, in a resource-constrained setting with a high incidence of IFI, generic voriconazole affords an effective and cheaper alternative.

|  Fig. 1 Flow diagram showing all the patients of acute myeloid leukemia treated in the study period and their outcomes with respect to development of invasive fungal infections

|  Fig. 2 Lung nodule showing with halo sign (a) and bilateral opacities with cavitations (b)

REFERENCES