Fluconazole

Limits on what this shows

• Null findings: Beck 2020 [13] (fluconazole inactive on the mineralocorticoid pathway, preclinical). Molloy 2018 [10]: no QTc over 500 ms at 1200 mg/day. Glasmacher 2005 [41]: hypokalemia was more frequent on itraconazole than on fluconazole (no counts). Malik 1998 [12]: fluconazole arm had less than half the hypokalemia rate of the amphotericin B arm.
• At-risk subgroups: Label: hypokalemia, advanced cardiac failure; quinidine and pimozide contraindicated; erythromycin to be avoided; amiodarone with high-dose (800 mg) fluconazole; hydrochlorothiazide raises fluconazole concentrations by 40% (no dose change required per label) [1].

Direction

Two directions are mixed in the sources and are kept apart here. (a) Possible potassium loss caused by the drug (disturbance, not a dietary deficiency). (b) Low potassium amplifying fluconazole's own QT toxicity (nutrient state modifying drug toxicity, Hard Rule 4). QT prolongation and torsades are drug outcomes, not nutrient deficiencies (Hard Rule 2). Neither direction is an indication to increase potassium intake.

Regulatory status

Tablet label: "Some azoles, including fluconazole, have been associated with prolongation of the QT interval"; "Fluconazole causes QT prolongation via the inhibition of Rectifier Potassium Channel current (Ikr)"; "rare cases of QT prolongation and torsade de pointes"; "Patients with hypokalemia and advanced cardiac failure are at an increased risk" for life-threatening ventricular arrhythmias. Hypokalemia is listed among postmarketing metabolic events with hypercholesterolemia and hypertriglyceridemia; no potassium monitoring instruction was found in the text read [1]. The IV label repeats the QT precaution and lists hypokalemia only as a postmarketing reaction [2]. Wikipedia lists hypokalaemia as rare (under 0.1%) [4] (Tier D, cross-check only).

Population evidence

Malik 1998 [12]: randomized comparison in cancer patients with prolonged fever and neutropenia, 100 analyzed (48 amphotericin B, 52 fluconazole 400 mg orally daily). "Severe hypokalemia (25 versus 12)", amphotericin B versus fluconazole. Full text opened on the repair pass: Table 5 splits these into correctable hypokalemia (13 amphotericin B, 9 fluconazole) and refractory hypokalemia (12 amphotericin B, 3 fluconazole), so 12 of 52 (23%) in the fluconazole arm against 25 of 48 (52%) in the amphotericin B arm; the denominators are the analyzed arm sizes. The paper gives no numeric potassium threshold (severity is defined by whether replacement allowed therapy to continue), no baseline potassium and no P value for the hypokalemia difference; overall adverse effects were 19 of 52 (36%) versus 32 of 48 (67%), P Glasmacher 2005 [41]: open-label randomized trial, about 494 neutropenic haematology patients (itraconazole oral solution 248, fluconazole oral solution 246), 8 weeks; the abstract states "more hypokalaemia occurred in the itraconazole group" without counts, and total adverse events were similar; fluconazole was the comparator arm. Molloy 2018 [10]: fluconazole 1200 mg/day in HIV-associated cryptococcal meningitis, mean QTc change to day 7 +10.1 ms (IQR -28 to 46) versus -12.6 ms (IQR -39 to 13.5) without fluconazole, P = 0.04, none over 500 ms; the authors note it "remains important to correct any electrolyte imbalance"; n not stated in the abstract. Case reports [9]: torsades with fluconazole in at least nine single patients (Tholakanahalli 2001, Khazan 2002, Pham 2006 at 200 mg/day IV, Tatetsu 2006, Esch 2008, McMahon 2008 with QTc 556 ms after resuscitation, Gandhi 2003, Tilton 2019, Unal Yuksekgonul 2021); in the one with a reported value, potassium was 3.6 mEq/L alongside levofloxacin (Gandhi); two others mention low potassium without values. Menaka 2025 [14]: prospective, 116 adult ICU patients with drug-induced QTc prolongation; fluconazole implicated in 10.34% of cases; potassium changes significant, magnesium not; multi-drug cohort not stratified to fluconazole, so Tier C for this question. Zeuli 2013 [11]: 94 hematology patients on a fluoroquinolone plus an azole, mean QTc change 6.1 ms (95% CI 0.2 to 11.9); the fluconazole share was not confirmed, so Tier C and not used to grade.

Mechanism

QT: direct IKr block, stated in the label. Potassium loss: no mechanism is named in the label. The pseudohyperaldosteronism route described for posaconazole and itraconazole does not extend to fluconazole in the one preclinical test found: fluconazole "showed no effect on mineralocorticoid targets" (Beck 2020, listing level) [13].

Counter-evidence held on record

Counter-evidence and confounding The 12 events in the fluconazole arm of Malik (12 of 52) are from an active-comparator trial in neutropenic cancer patients with chemotherapy, fever and gastrointestinal losses; an active-comparator design does not test fluconazole against non-use, so it cannot attribute hypokalemia to fluconazole, and the comparator (amphotericin B) is a known potassium-wasting drug. The same limit applies to Glasmacher 2005. Orange was re-examined on the repair pass and not met: no drug-specific rate against a placebo or untreated group exists in anything opened. Two ectopic Cushing syndrome case reports with fluconazole describe hypokalemia at presentation, which is a feature of the disease [40]. Hypokalemia in the QT case reports preceded or accompanied other QT-prolonging drugs (levofloxacin, amphotericin B). Framing caution Monitoring prompt for people with QT risk factors, not a potassium deficiency claim and not a reason to take potassium. Hypokalemia here is mainly a risk modifier for the drug's own arrhythmia toxicity; potassium changes belong with a prescriber.

Limits on what this shows

• Null findings: None found, because no study measured 25-hydroxyvitamin D in fluconazole users with normal calcitriol. The Filler and Ahsan reports are multi-intervention single cases.
• At-risk subgroups: People with calcitriol-driven hypercalciuria or hypercalcemia (HHRH, CYP24A1 loss, granulomatous disease) where the effect is the target. No subgroup with a vitamin D deficit has been identified.

Direction

Reduction of the active hormone through an enzyme effect (functional, hormone-level), intended in disorders with excess calcitriol. Not a dietary vitamin D deficiency and not an indication to supplement vitamin D (Hard Rule 2). Case reports of calcium change point in both directions (below).

Regulatory status

The tablet label text read contains no vitamin D or calcium statement; its only vitamin entry is the vitamin A case report (see the vitamin A entry) [1].

Population evidence

Filler 2022 [15]: one 12-year-old boy (SLC34A3 HHRH), fluconazole titrated to 100 mg once daily alongside growth hormone, phosphate, potassium and salt restriction. 1,25(OH)2D above 480 pmol/L at baseline (reference 60 to 208) then 462 and 426 pmol/L on treatment; urinary calcium/creatinine 1.66 to 0.73 mol/mmol (reference below 0.6); serum phosphate 0.65 to 0.87 mmol/L; 25-hydroxyvitamin D not reported; abstract level. Ahsan 2025 [18]: 26-year-old man with a homozygous CYP24A1 variant; serum calcium 12.7 to 9.4 mg/dL over a period that included 3 months of fluconazole, vitamin D and calcium restriction and magnesium; 1,25(OH)2D value not in the abstract. Roztoczynska 2025 [17]: newborn with Barakat syndrome and congenital hypoparathyroidism, calcium normalized on fluconazole, mechanism proposed via vitamin D metabolism; no vitamin D levels in the abstract. Akman 2024 [19]: hypercalcemia (13 mg/dL) on all-trans-retinoic acid plus fluconazole 400 mg/day, no vitamin D measured. FLUCOLITH [16] (Bertholet-Thomas 2022): published protocol for a randomized placebo-controlled trial of fluconazole in hypercalciuria with 1,25(OH)2D above 150 pmol/L, planned n = 60, primary outcome 24-hour urinary calcium at 16 weeks; no results were found. A retrospective series of 11 SLC34A3/SLC34A1 patients says azoles appear promising, with no fluconazole-specific numbers in the abstract (Naciri Bennani 2025, listing level) [39].

Class extrapolation

CLASS EXTRAPOLATION, capped at yellow. Ketoconazole in hypercalcemic sarcoidosis [20]: Glass 1990 (2 men, 600 to 800 mg/day for 4 to 6 days, 1,25(OH)2D fell about 40%, calcium unchanged); Adams 1990 (1 patient, 800 mg/day, 1,25(OH)2D fell 73% in 4 days, calcium fell 15%); Bia 1991 (1 patient, normalized on 800 mg/day, no numbers). Cohort test: fluconazole was not in any of these cohorts; they are uncontrolled and the patients had raised calcitriol from granulomas.

Mechanism

Proposed: azole inhibition of the vitamin D hydroxylases, usually described as CYP27B1 (1-alpha-hydroxylase) inhibition. A 2024 review lists fluconazole as a possible 1-alpha-hydroxylase inhibitor (listing level, full text not read) [21]. No source read gives enzyme data for fluconazole on CYP24A1 or CYP27B1, so the CYP24 and CYP27 route is not demonstrated here.

Counter-evidence held on record

Counter-evidence and confounding All fluconazole reports are single cases or a protocol, with co-interventions (growth hormone, phosphate, diet, vitamin D restriction). The calcium reports run in opposite directions (hypercalcemia with retinoic acid, hypocalcemia dependence in Barakat syndrome, lowering in CYP24A1 deficiency), which is not what one steady enzyme effect predicts. Ketoconazole data are a different molecule. Framing caution Hormone-level and indication-specific. Not a vitamin D deficiency claim, and vitamin D supplements could oppose the intended effect in calcitriol-excess disorders. No supplement recommendation.

Limits on what this shows

• Null findings: Thompson 2019: retinoic acid change not causal for alopecia. No study of fluconazole on serum retinol was found.
• At-risk subgroups: Patients prescribed ATRA (tretinoin) by mouth. No evidence that people not taking a retinoid are affected.

Direction

Retention or excess of an exogenous retinoid drug through inhibited metabolism (a drug-drug interaction). Not a vitamin A deficiency; adding vitamin A would be the hazard, not the remedy (Hard Rules 4 and 5).

Regulatory status

Tablet label, drug interactions: "Vitamin A: Based on a case report in one patient receiving combination therapy" with all-trans-retinoic acid, pseudotumor cerebri that resolved after fluconazole was discontinued; the label says the combination may be used but CNS effects should be considered [1].

Population evidence

Schwartz 1995 [22]: pharmacokinetics in 2 patients with acute promyelocytic leukemia plus human liver microsome work. Patient 1: ATRA plasma AUC fell from 277 to 91 ng/mL/hr on ATRA alone, then rose to 401 ng/mL/hr after fluconazole; patient 2 showed a smaller effect (no value in the record read). Abstract level. Case reports [23]: Vanier 2003 (ATRA toxicity), Yarali 2008 (acute renal failure), and Akman 2024 (hypercalcemia) [19], each one patient. Thompson 2019 [23]: fluconazole-induced alopecia in an animal model and a human cohort; plasma and tissue retinoic acid was measured and judged not causal (no numbers in the abstract). Embryo studies in rat and mouse report altered CYP26 expression and branchial defects with fluconazole (Menegola 2004, Di Renzo 2007, Tiboni 2009, listing level, animal) [39].

Mechanism

Named: inhibition of retinoic acid hydroxylation (CYP26) by fluconazole, supported by the in vitro human liver microsome result in the Schwartz title and by animal expression data. The label lists fluconazole as a CYP2C9, CYP3A4 and CYP2C19 inhibitor and does not name CYP26.

Counter-evidence held on record

Counter-evidence and confounding Two patients, abstract-level reading, and ATRA exposure falls with repeated dosing by autoinduction, which confounds the sequence in the Schwartz patient. ATRA is a pharmaceutical retinoid, not dietary vitamin A, so this is a drug interaction rather than a nutrient-status finding. Framing caution Excess-risk signal for people on ATRA, not a vitamin A deficiency. Do not recommend vitamin A. This is a prescriber and pharmacist interaction check.

Limits on what this shows

• Null findings: Magill 2004 (placebo-controlled, 400 mg/day, no cortisol or adrenal dysfunction difference); Washton 1989 (no hormone effects at 50 to 400 mg/day).
• At-risk subgroups: Patients on chronic prednisone or other glucocorticoids stopping fluconazole (label); people with limited adrenal reserve; the reported adrenal crisis was a single case.

Direction

Hormone suppression (steroid synthesis), not a deficiency of any nutrient. Under Hard Rule 2 it is recorded as a hormone effect so a reader can see why no sodium, potassium or glucocorticoid-related nutrient is graded from it.

Regulatory status

Label: "Adrenal insufficiency has been reported in patients receiving azoles, including fluconazole"; "Reversible cases of adrenal insufficiency have been reported in patients receiving fluconazole"; in healthy volunteers, fluconazole 200 to 400 mg once daily for up to 14 days had "small and inconsistent effects" on testosterone, endogenous corticosteroids and the ACTH-stimulated cortisol response (subject numbers and values not given in the text read); patients on prednisone "should be carefully monitored for adrenal cortex insufficiency when fluconazole is discontinued" [1].

Population evidence

Magill 2004 [5]: double-blind placebo-controlled trial, 154 critically ill surgical patients (79 fluconazole 400 mg/day, 75 placebo); median plasma cortisol 15.75 versus 16.71 ug/dL (P = 0.52); adrenal dysfunction odds ratio 0.98 (confidence interval not retrieved); abstract level. Liao 2025 [6]: retrospective study of 22 patients with Cushing syndrome on 112.5 to 450 mg/day orally; 24-hour urinary free cortisol fell from 717.6 +/- 1219.4 to 184.1 +/- 171.8 ug/day (P = 0.035) (intended effect). Choo 2023 [7]: one adult with acute adrenal crisis precipitated by fluconazole (early-morning cortisol below 8 nmol/L, ACTH 3.2 ng/L) presenting with hypercalcemia. Washton 1989 [8] (review, listing level): no effect on basal or ACTH-stimulated cortisol, testosterone, estrogen or progesterone at 50 to 400 mg/day. Two Cushing syndrome case reports describe cortisol normalizing on fluconazole 400 mg/day [40].

Mechanism

Not named in the label text read. The therapeutic use in Cushing syndrome implies inhibition of adrenal steroid synthesis at higher exposure; the label states that mammalian sterol demethylation is "much less sensitive" to fluconazole than the fungal enzyme.

Counter-evidence held on record

Counter-evidence and confounding The one controlled trial is null at 400 mg/day, and the large cortisol falls come from patients treated for hypercortisolism where lowering cortisol is the goal. The adrenal crisis case is a single report with its own confounders. Framing caution Hormone effect outside the nutrient scope. It is shown only so that fluconazole is not credited with sodium, potassium or vitamin findings that would really be adrenal effects. No nutrient recommendation follows.

Limits on what this shows

• Null findings: No fluconazole-specific zinc measurement found.
• At-risk subgroups: Undernourished older inpatients (from the study population).

Direction

Indication-linked context flag (confounding by indication presumed, Hard Rule 6). No evidence that fluconazole changes zinc.

Regulatory status

No zinc statement in the label text read [1].

Population evidence

Paillaud 2004 [33]: observational study of 97 hospitalized elderly patients, 36 with oral candidiasis; hypozincemia (serum zinc below 12.5 umol/L) was more frequent in the candidiasis group, exact figure not in the abstract; fluconazole-treated patients were mentioned without a number; no analysis by fluconazole exposure. Listing level.

Mechanism

None for fluconazole. Zinc(II) coordination complexes of fluconazole exist in chemistry papers and are not a host effect [31].

Counter-evidence held on record

Counter-evidence and confounding Hypozincemia in candidiasis patients predates or accompanies the infection; it cannot be attributed to the antifungal. Framing caution Context about the underlying condition. No zinc supplement advice.

Limits on what this shows

• Null findings: No fluconazole B vitamin study was located.
• At-risk subgroups: None identified.

Direction

Reverse: nutrients modifying drug toxicity in an animal model (Hard Rule 4). It does not show that fluconazole lowers any B vitamin.

Regulatory status

No B vitamin statement in the label text read [1].

Population evidence

Al-Abbasi 2019 [34]: randomized animal study, 54 Wistar rats in 9 groups of 6; a vitamin B combination (B1, B2, B3) reduced fluconazole-induced liver enzyme and histology changes; outcomes were liver markers, not vitamin levels. PubMed listing for fluconazole with folate, thiamine, cobalamin, pyridoxine, riboflavin, niacin or biotin (35 results, first 10 read): no record measured a B vitamin level in fluconazole-exposed subjects. Paillaud 2004 measured folate and B12 in a mixed inpatient cohort without reporting values or fluconazole analysis [33].

Mechanism

None named for depletion.

Counter-evidence held on record

Counter-evidence and confounding Rat hepatotoxicity at the doses used does not translate to human B vitamin status, and the supplementation arm is not a depletion test. Framing caution Not a B vitamin deficiency claim. No vitamin recommendation.

Limits on what this shows

• Null findings: Food interaction: none found on the label or on Drugs.com ("no known alcohol/food interactions") [38].
• At-risk subgroups: Children (vomiting 5.4% in 577 pediatric patients); people with poor baseline intake.

Direction

Context flag (reduced intake or fluid and electrolyte loss during illness), not a deficiency.

Regulatory status

Label: in 4,048 patients treated 7 or more days, treatment-related nausea 3.7%, headache 1.9%, rash 1.8%, vomiting 1.7%, abdominal pain 1.7%, diarrhea 1.5%; single 150 mg dose in 448 patients: headache 13%, nausea 7%, abdominal pain 6%, diarrhea 3% [1]. The label states exposure "is not affected by food" and tablets may be taken without regard to meals [1].

Population evidence

Label trial frequencies only; no study linked these events to a measured nutrient change.

Mechanism

Gastrointestinal intolerance (not characterized in the label text read).

Counter-evidence held on record

Counter-evidence and confounding Rates are low and apply mostly to infection and background disease; they do not identify a nutrient. Framing caution Monitoring prompt for persistent vomiting or diarrhea only.

Limits on what this shows

• Why no grade: Dual condition met: no Tier B or C evidence connects fluconazole to magnesium, and no circulating claim was found. Disclosed gap, not reassurance. The potassium entry covers the electrolyte findings that do exist.

What was searched

PubMed listing "fluconazole AND magnesium" (18 results; both pages, all 18 records, now read) [14][43]: no study measured magnesium in fluconazole users. Page 2 (repair pass, records 11 to 18): the Zeuli 2013 QT cohort (snippet truncated, electrolytes not seen), an EDTA Cryptococcus biofilm study, an alkyl-tertiary-alcohol chemistry paper, a drug-metal complex and platelet aggregation chemistry paper, the Tatetsu 2006 torsades case (intravenous magnesium sulfate given as treatment, no serum magnesium reported), a 1990 study of oral antacid on fluconazole pharmacokinetics (antacid-to-drug direction, abstract not opened, outcome is fluconazole absorption and not magnesium status), a voriconazole renal case, and a Cryptococcus ecto-ATPase paper; none measures magnesium in fluconazole-exposed people. The nearest record is Menaka 2025, a 116-patient ICU cohort of drug-induced QTc prolongation in which fluconazole was implicated in 10.34% of cases; potassium changes were significant and magnesium changes were not, with no fluconazole-stratified result. Cochrane amphotericin B versus fluconazole reviews mention magnesium only as supplementation given with amphotericin B (no values) [39]. Label text read: no magnesium statement [1]. MedlinePlus asks people to report low magnesium (among calcium, sodium and potassium) before use, a general caution rather than a depletion claim [38].

Channel outcomes

The listing is now fully read at snippet level (18 of 18). Europe PMC and a second phrasing were not run because of rate limits, and the Zeuli abstract was not opened for electrolytes. Open-web search unavailable. Result: a genuine gap on the channels that worked, with the tooling limits stated.

Limits on what this shows

• Why no grade: Dual condition met: no Tier B or C evidence supports a yellow grade for fluconazole and sodium, and no circulating claim was found. One retrospective null (Cave 2023, a Tier A cohort with fluconazole in it but an outcome of clinical hyponatremia without retrievable numbers) is not proof of safety and is not a graded finding. Disclosed gap.

What was searched

PubMed listing "fluconazole AND hyponatremia" (21 results, first 10 read) [35]: Cave 2023, a retrospective chart review (Omaha, 2013 to 2021) of vincristine side effects with concomitant fluconazole in pediatric ALL induction, reported "no correlation between fluconazole use and increased incidence of hyponatremia or peripheral neuropathy" (re-read on the repair pass through an OpenAlex single-record lookup: the abstract still gives no figures, n or odds ratio, and the outcome is clinical hyponatremia, not a sodium level). Other records on page 1 describe hyponatremia as a baseline or disease feature in cryptococcal or coccidioidal infection (n = 47, 83, 236 and single cases), with no sodium change attributed to fluconazole. Page 2 (repair pass, records 11 to 20 of 21) held voriconazole and transplant cases, cryptococcal meningitis with hyponatremia or SIADH (sodium 120 to 125 mmol/L, disease feature), disseminated cryptococcosis with adrenal insufficiency, cerebral salt wasting in cryptococcal meningoencephalitis, and one liver-transplant cryptococcosis case in which hyponatremia preceded fluconazole maintenance therapy; none measures sodium on fluconazole against a comparator [43]. Label text read: no sodium statement for the tablet; the IV product contains 9 mg of sodium chloride per mL with 2 mg of fluconazole and carries no sodium load warning in the excerpt read [2] (by arithmetic, a 200 mg bag of 100 mL carries about 900 mg of sodium chloride; this figure is calculated here, not stated by the label).

Channel outcomes

Twenty of 21 records are now read; page 3 (1 record) was not read. Page 2 needed four attempts with waits of about 5 minutes because of HTTP 429 refusals. Open-web search unavailable. Sodium load from the IV diluent is a formulation fact rather than a nutrient effect.