Table 2.

Studies (from 2015 to 2019) on the relationship between CYP2C19 polymorphisms and voriconazole pharmacokinetics, clinical outcomes and adverse effects

Study design, author, year	Age, race, sample size (n)	Treatment or prophylaxis use, voriconazole dose reported	CYP2C19 genotype (n)	Voriconazole PK, clinical outcomes and adverse effects
Association between CYP2C19 genotype and voriconazole PK
Case series, Weigel 2015 [27]	Adult Caucasian, n = 17	Treatment of critically ill adults 6 mg/kg IV q12h for first 24 h, then 4 mg/kg IV q12h	UM/RM (n = 3) NM (n = 3) Unknown (n = 11)	Subtherapeutic troughs occurred more frequently in UM/RM (77%) than NM (33%).
Retrospective study, Gautier-Veyret 2015 [28]	Adult Caucasian, n = 33	Prophylaxis, post-allogeneic HSCT No dose reported.	UM/RM (n = 11) NM (n = 10) IM (n = 8)	IM had higher initial trough levels than UM/RM (p <0.01)
Retrospective study, Lamoureux 2016 [29]	Adult Caucasian n = 35 genotyped n = 30 controls	Treatment 400 mg PO q12H for first 24 h, then 200 mg PO q12H. Adjust by 50–100 mg q12H with goal trough 1–5 mg/L	UM (n = 4) RM (n = 13) NM (n = 11) IM (n = 6) PM (n = 1)	UM/RM associated with subtherapeutic levels PM associated with supratherapeutic levels
Case report, Hicks 2016 [30]	Pediatric (10 years old) Caucasian, n = 1	Treatment 9.5 mg/kg q12H then 8 mg/kg q12H	RM (n = 1)	Pediatric RM patient whose final dose that achieved target trough was 14 mg/kg q12H for first 24 h then 12 mg/kg IV q12H
Prospective observational study, Chuwongwattana 2016 [31]	Adult Asian, n = 115	Treatment 6 mg/kg IV q12H then 4 mg/kg IV q12H or 200–250 mg PO BID	NM (n = 59) IM (n = 42) PM (n = 14)	IM and PM had higher trough levels than NM
Retrospective study, Niioka 2017 [32]	Adult Asian n = 65	Prophylaxis (n = 54) and treatment (n = 11), Dose ranged from 100 to 300 mg IV/PO BID	NM (n = 24) IM (n = 29) PM (n = 12)	Subtherapeutic levels occurred more frequently in NM (58%), IM (21%), than PM (8%). Voriconazole to metabolite ratio was impacted by route of administration, C-reactive protein, age, and CYP2C19 genotype.
Prospective cohort study, Hamadeh 2017 [33]	Adults 81% Caucasian n = 70	Treatment 6 mg/kg IV q12H × first 24 h, then 4 mg/kg q12H.	UM (n = 3) RM (n = 24) NM (n = 28) IM (n = 14) PM (n = 1)	Subtherapeutic levels occurred more frequently in UM/RM (52%) than NM/IM/PM (16%) (p <0.01). Two UM and one RM achieved target level after dose increased to 5 mg/kg q12h.
Cross-sectional study, Ebrahimpour 2017 [34]	Adult Iranian n = 37	Treatment 400 mg IV BID × first 24 h then 200 mg IV/PO BID. Target level: 1–5.5 mg/L	RM (n = 8) NM (n = 18) IM (n = 9) PM (n = 2)	Genotype correlated with trough levels. RM: 25% at target level, 75% subtherapeutic level NM: 94% at target level, 6% supratherapeutic level IM: 78% at target level, 22% supratherapeutic level PM: 50% at target level, 50% supratherapeutic level
Retrospective study, You 2018 [35]	Adult Asian n = 64	Treatment IV or oral voriconazole per package insert	NM (n = 29) IM (n = 27) PM (n = 8)	PM had higher odds of supratherapeutic troughs than non-PM (p < 0.02). Variables that influenced troughs include age, CYP2C19 genotype, and liver function.
Retrospective cohort study, Miao 2018 [36]	Adult Asian n = 106	Treatment 400 mg q12H for 1st 24 h then 200 mg BID	NM (n = 48) IM (n = 44) PM (n = 14)	Median troughs were higher in PM, IM than NM (4.22, 3.12, and 1.87 mg/L respectively (p < 0.05)
Retrospective study, Shao 2017 [10]	Pediatric (≥ 15 years old) and adult Asian n = 86	Treatment or prophylaxis in hematological malignancy patients Dose per label with maintenance dose range from 200 to 600 mg daily. Target level: 1–4 mg/L	NM (n = 32) IM (n = 36) PM (n = 18)	IM/PM had higher troughs than NM (4.1 ± 2.5 vs. 2.8 ±1.9 mg/L) (p = 0.002) Additional variables for IM/PM included CYP3A4, age, and body mass index
Association between CYP2C19 genotype and voriconazole-related clinical and safety outcomes
Retrospective study, Wang 2014 [37]	Adult Asian n = 144	Treatment Voriconazole dosed based on package insert	RM (n = 3) NM (n = 62) IM (n = 62) PM (n = 17)	Levels and clinical outcomes and safety: Trough levels correlated with efficacy and hepatoxicity. CYP2C19 genotype correlated with trough levels No association between CYP2C19 genotype and hepatotoxicity Response to treatment was ~ 90% when trough was 1.5–4 mg/L
Prospective study, Trubiano 2015 [38]	Adult Caucasian n = 19	Treatment, hematological malignancy patients, 6 mg/kg IV q12h × first 24 h then 4 mg/kg IV/PO q12h.	UM (n = 1) RM (n = 5) NM (n = 8) IM (n = 5)	Trough correlated with CYP2C19 phenotype. Median trough higher in IM, NM followed by UM/RM (5.23, 3.3, and 1.25 mg/L respectively) UM/RM/NM (50%) had subtherapeutic levels. IM took longest time to reach target level. Adverse events was highest in IM (60%) and was mainly associated with hepatotoxicity or photopsia
Case reports, Beata 2017 [39]	Adult Caucasian n = 4	Prophylaxis, post-allogeneic-HSCT patients who developed adverse drug reaction so genotyping was done Dose was based on product label.	IM (n = 4)	No voriconazole levels drawn Adverse events: GI, dermatological, neurological, hepatobiliary, and renal adverse reactions reported in these IM cases.
Case report, Danion 2018 [40]	Adult Caucasian n = 2	Treatment of cerebral aspergillus UM case: initially 200 mg BID RM case: initially 250 mg BID (4 mg/kg BID). Target trough: 2–5 mg/L	UM (n = 1) RM (n = 1)	UM: subtherapeutic troughs despite increasing dose to 350 mg BID (8 mg/kg BID). After genotyping showed UM phenotype, the drug was changed to isavuconazole with a treatment response. RM: subtherapeutic troughs and eventually increased to 400 mg PO TID (7 mg/kg TID) to reach target trough. Patient responded to treatment.
Application of PK-PD modeling to CYP2C19 genotype based voriconazole dose adjustments
Retrospective study, Wang 2014 [37]	Adult Asian n = 144	Treatment Voriconazole dosed based on package insert	RM (n = 3) NM (n = 62) IM (n = 62) PM (n = 17)	Simulated-derived treatment voriconazole dose showed: For PM: 200 mg PO/IV BID For non-PM: 300 mg PO BD or 200 mg IV BID
Retrospective study, Lamoureux 2016 [29]	Adult Caucasian n = 35genotyped n = 30 controls	Treatment 400 mg PO q12H for first 24 h, then 200 mg PO q12H. Adjust by 50–100 mg q12H with goal trough 1–5 mg/L	UM (n = 4) RM (n = 13) NM (n = 11) IM (n = 6) PM (n = 1)	UM/RM associated with subtherapeutic levels PM associated with supratherapeutic levels Model-derived oral dose: For UM: 6 mg/kg BID For RM: 4 mg/kg BID For NM: 2.5 mg/kg BID
Retrospective study, Mangal 2018 [24]	Adult 81% Caucasian n = 68	Treatment PK data from infected adults with known 2C19 genotypes receiving weight-based voriconazole dosing. Monte Carlo simulations performed to determine probability to achieve therapeutic level and response	UM (n = 3) RM (n = 24) NM (n = 27) IM (n = 14)	RM/UM had low probability to achieve therapeutic level for aspergillosis (trough > 2 mg/L) using 200 mg q12h (23.2%), improved by pantoprazole coadministration (46.5%). RM/UM: Monte Carlo simulation found labeled dosing appropriate for candidiasis, but higher dosing of 500–600 mg q12h needed for invasive aspergillus infections. Simulation generated different voriconazole doses in the presence and absence of different proton-pump inhibitors.
Prospective PK study, Lin 2018 [41]	Adult Asian n = 106	Treatment or prophylaxis in renal transplant patients Voriconazole dose based on package insert.	RM (n = 1) NM (n = 44) IM (n = 49) PM (n = 12)	Trough correlated with CYP2C19 phenotype Median troughs in RM, NM, IM and PM were 1.90, 2.19, 2.32, and 3.86 mg/L (p < 0.001) Model-derived doses with target level 2–6 mg/L NM: 300 mg IV q12H IM: 200 mg IV q12H or 350 mg PO BID PM: 150 mg IV q12H or 250 mg PO BID.
Prospective PK study, Kim 2019 [42]	Adult Asian n = 193 (93 volunteers and 100 patients)	Healthy volunteers and non-infected patients using different dosing regimens. Dose ranged from 200 to 400 mg IV or PO, as single or multiple doses every 12 h; to 6 mg/kg IV or oral 400 mg q12H for first 24 h, then 4 mg/kg or 200 mg PO q12H	NM (n = 75) IM (n = 70) PM (n = 48)	Voriconazole clearance was reduced in 17% (IM) and 53% (PM). Modeled dosing 400 mg q12h day 1,then 200 mg q12h predicted target troughs (2–5.5 mg/L) on day 7 in 39% (74% NM subtherapeutic and 48% PM supratherapeutic). Dose based on PK model: NM: 400 mg q12h IM: 200 mg q12h PM: 100 mg q12h
Prospective CYP2C19-guided voriconazole dosing
Prospective study, Teusink 2016 [43]	Pediatric Race not reported Pilot group (n = 25) Genotype-guided group (n = 20)	Prophylaxis, post-HCT patients Pilot study: fixed dose 5 mg/kg q12H Genotype-guided dosing NM: 7 mg/kg q12H IM: 6 mg/kg q12H PM: 5 mg/kg q12H Target level: 1–5.5 mg/L	Pilot study RM (n = 2) NM (n = 17) IM (n = 3) PM (n = 1) Unknown (n = 1) Genotype-guided RM (n = 1) NM (n = 10) IM (n = 7) PM (n = 2) Unknown (n = 2)	Median time to target level was shorter in IM, than NM and RM (4, 6.5, and 9 days respectively) Genotype-guided dosing reached target level faster (6.5 vs. 29 days, P < 0.01) when all patients were started on the same dose regardless of CYP2C19 genotype.
Case reports prospective genotyping, Fulco 2019 [44]	Adult n = 2 Caucasian (50%) AA (50%)	Treatment, aspergillosis in HIV patients Genotype-guided initial dose adjustment NM: initial 200 mg BID IM: initial 350 mg BID with dose-adjusted per TDM.	NM (n = 1) IM (n = 1)	NM: final dose remained at 200 mg BID with TDM IM: final dose adjusted to 75 mg BID with TDM Preemptive genotyping guided initial voriconazole dose adjustment as there would be interaction with HIV medication (CYP3A4 inhibitor) that could explain the increased trough levels.
Prospective study, Patel 2019 [45•]	Adult n = 89 Caucasian (73%)	Prophylaxis, post-allogeneic HCT UM/RM: 300 mg PO BID (intervention dose) NM/IM/PM: 200 mg PO BID (standard dose)	UM (n = 3) RM (n = 29) NM (n = 30) IM (n = 23) PM (n = 4)	Less patients with subtherapeutic levels in genotype-guided cohort than historical control (29% vs. 50%, p < 0.001) Lower subtherapeutic rates in UM/RM on intervention dose (15.6%) than NM and IM on standard dose (50% and 26.1% respectively). PM had no subtherapeutic rates (0%) at standard dose. The supratherapeutic rates in UM/RM on intervention dose was (6.3%) compared to NM/IM/PM (0%). Clinical success defined as no treatment failure or continued treatment until day 100 post-HCT showed higher success rate in genotype-guided vs. historical cohort group (78% vs. 54%, p < 0.001) Adverse events (56% transaminitis, 8% neurological symptoms) did not correlate with trough levels, with PM (33%), IM (17%), NM (13%), and UM/RM (14%).
Prospective study, Hicks 2019 [46•]	Adult n = 202 76% Caucasian	Prophylaxis, acute myeloid leukemia patients UM: use alternative agent RM: 300 mg BID (intervention dose) NM/IM/PM: 200 mg BID (standard dose)	176 patients dosed per protocol UM (n = 3) RM (n = 46) NM (n = 64) IM (n = 56) PM (n = 7)	RM: median trough was higher with intervention vs standard dose (2.7 vs. 0.6 mg/L, p = 0.001) More RM achieved target level via CYP2C19-guided dosing than standard dosing (84% vs. 46%, p = 0.02) Clinical outcome: non-significant decrease in hospital-acquired nodular pneumonia incidence (p = 0.46) Adverse events: no difference in neurotoxicity or transaminitis rates between intervention and standard dosing groups.

UM, ultrarapid metabolizer (CYP2C19*17/*17); RM, rapid metabolizer (CYP2C19*1/*17); NM, normal metabolizer (CYP2C19*1/*1); IM. intermediate metabolizer (CYP2C19*1/*2, *1/*3, *2/*17); PM. poor metabolizer (CYP2C19*2/*2, *2/*3, *3/*3); HCT, hematopoietic cell transplant; HIV human immunodeficiency virus; TDM, therapeutic drug monitoring