Fosfomycin

Limits on what this shows

• Null findings: Milan: no association with daily dose or continuous versus intermittent infusion [11]. Turkish prospective cohort: daily sodium and potassium mEq loads were not associated with electrolyte disorder status (unadjusted, n=54) [10]. The licensing trial rate (3.4%) is far below the ICU cohorts. No sodium result was reported in the healthy-volunteer IV studies [7][8], and none occurred with oral dosing in the labels [2][3]. Reported associations of electrolyte disorders with mortality [10] are unadjusted and confounded by illness severity; a retrospective ICU abstract linking hypernatremia to stay and survival [31] was not used because its group labels were ambiguous. Neither is used for grading.
• At-risk subgroups: ICU patients and septic shock (for any adverse event) [11], positive fluid balance (associated with any electrolyte disorder) [10], higher APACHE II score, enteral nutrition and albumin replacement [13], age 70 or over (for any adverse event) [12], heart failure, renal impairment, cirrhosis, hypertension, hyperaldosteronism, and neonates under sodium restriction [1][4]. In the pooled analysis of two Staphylococcus aureus bacteremia trials, treatment stopped for acute heart failure in 5 of 178 combination patients versus 1 of 191 on monotherapy [16]; heart failure is a clinical outcome and is not filed as a nutrient finding.
• Counter-evidence and confounding: ICU cohorts were treated with several other antibiotics (83% to 99% combination use), saline-based fluids and feeds, so cohort rates overstate the drug's independent effect; the randomized trial (rate 3.4%) and the pharmacovigilance disproportionality are the less confounded lines. Definitions differ (over 145, 150 or above, over 150). The two Turkish series and the Taiwan series are small and unadjusted.

Regulatory status

US IV label: 1,980 mg sodium per 6 g vial (330 mg per gram); a low-sodium diet is recommended; monitor electrolytes, fluid status and edema, especially in cardiac insufficiency, renal impairment, cirrhosis, hypertension, hyperaldosteronism, hypernatremia or pulmonary edema [1]. EU text: 0.32 g sodium per gram; hypernatremia listed as common; low-sodium diet recommended [4]. Oral label: no sodium statement [3]; EU text: no sodium warning for oral products [4].

Population evidence

Licensing RCT (label, 233 fosfomycin versus 231 piperacillin-tazobactam; hypernatremia defined as sodium over 150 mEq/L): 3.4% versus 0.9% [1]. Cohorts, all IV: German/Austrian prospective, n=209, 14.8% (10.5% judged drug-related; 6 already hypernatremic before) [9]; Milan retrospective, n=224, 23.7% at sodium 150 or above; for any adverse event (not hypernatremia alone) adjusted RR was 1.59 (95% CI 1.09 to 2.31) in ICU and 1.73 (1.27 to 2.37) in septic shock, and daily dose was not associated [11]; Leipzig retrospective, n=143, 42.6% (sodium over 145) [12]; Turkish ICU retrospective, n=62, 41.9% with baseline hypernatremia excluded [13]; Taiwan ICU, n=42, 33.3% [14]; southern Italy, n=62, 17.7% [15]; Turkish prospective, n=54, 18.5% [10]. Pharmacovigilance: 25 EudraVigilance reports of IV fosfomycin with hypernatremia, reporting odds ratio 24.96 (95% CI 5.91 to 105.41) against gentamicin [15]. Plasma fosfomycin trough in the Milan cohort was 419.5 versus 140.6 mg/L with and without hypernatremia (p=0.012, AUROC 0.91; small subgroup) [11].

Mechanism

Present and direct. The product is a disodium salt; 6 g every 8 hours supplies about 5.9 g sodium a day (3 x 1,980 mg, [1]). A narrative review places fosfomycin above 3 g sodium a day at standard doses [18]. Diluent adds to it: in the Turkish ICU series every hypernatremia case was diluted in 0.9% saline [13], and in southern Italy 63.7% of cases were attributed to saline reconstitution in place of glucose [15].

Counter-evidence held on record

Framing caution This is an excess/retention risk of the IV product, not a dietary deficiency. Added salt or sodium supplements would be the hazard, not the remedy. Both labels recommend a low-sodium diet during IV treatment and monitoring by the prescriber. Does not apply to the oral single-dose sachet.

Limits on what this shows

• Null findings: Taiwan: no association with dose or time after exposure [14]. Turkish prospective: potassium load not associated [10]. In the pooled bacteremia trials, treatment stopped for hypokalemia under 3 mmol/L in 3 of 178 versus 1 of 191, and one trial advised supplementary potassium and furosemide [16]. A systematic review of 53 studies described comparably "low" incidence in the SAFO trial (4-hour infusions per [16]) and could not analyze electrolytes in detail [17]. Single-dose IV (1 g and 8 g) and oral 3 g in volunteers reported no potassium result [8]. Oral labels list no hypokalemia [2][3].
• At-risk subgroups: Albumin replacement [13] (ICU populations dominate the cohorts); people on digitalis or other QT-prolonging drugs (EU leaflet text; US label advises avoiding QT-prolonging drugs) [1][4].
• Counter-evidence and confounding: Cohorts include other antibiotics, diuretics, steroids and feeds that lower potassium. This is bounded by two less confounded lines: the randomized trial against an active comparator (comparator rate 1.7%), and the healthy-volunteer study with no illness or co-medication. Confounding by indication is not a plausible explanation for 49% in healthy adults.

Regulatory status

EU text: "Fosfomycin may decrease potassium levels in serum or plasma"; potassium supplementation "should be always considered"; hypokalemia listed as common; sodium and potassium to be monitored regularly [4]. US IV label: hypokalemia (potassium under 3.0 mEq/L and/or supplementation needed) is a warning; QTcF rose over 60 ms in 3.6% and torsades de pointes was reported postmarketing [1].

Population evidence

Licensing RCT: 9.9% versus 1.7% with piperacillin-tazobactam (233 versus 231; ratio about 5.8 from the reported percentages; the label gives no CI) [1]; the trial's own report [6] states hypokalemia and aminotransferase elevations were mostly mild and transient (abstract only). Healthy volunteers, 6 g every 8 hours for 3 doses (18 g in about 24 hours), fasting, n=39: 19 (49%) had hypokalemia, all mild; not taking significant concomitant medications and without acute or chronic illness [7]. Cohorts: Leipzig n=143, 39.9% (potassium under 3.6) [12]; Turkish ICU n=62, 33.9% at under 3.5, onset 3.9 days, baseline hypokalemia excluded [13]; Taiwan ICU n=42, 28.6%, with a significant change between treatment start and stop (p under 0.001) [14]; Turkish prospective n=54, 51.8% (mild 20, moderate 6, severe 2; onset 2.7 days; treatment stopped in 2) [10]; Milan n=224, 9.8% (potassium under 3 with symptoms) [11]; German/Austrian n=209, 6.2% (2.4% judged drug-related; one severe) [9]. The spread (6.2% to 51.8%) tracks the threshold used (under 3.0 to under 3.6).

Mechanism

Proposed, not measured in the sources retrieved. Authors of the Turkish ICU series state that the loss is "likely due to increased urinary potassium excretion in the distal tubules" [13]; a review states fosfomycin "may increase potassium renal excretion" citing the EU product information and a pharmacovigilance letter [19] (the letter is [32], not opened). The healthy-volunteer report proposes no mechanism [7]. A short infusion time has been linked to the effect: no hypokalemia with 4-hour infusions [13], fewer events with prolonged infusion and none with 3-hour infusions in two cohorts described in a systematic review [17].

Counter-evidence held on record

Framing caution A transient serum disturbance during IV treatment, managed in hospital by monitoring and prescriber-directed replacement. It is not a dietary deficiency and gives no basis for a consumer potassium supplement. Does not describe the oral single-dose sachet.

Limits on what this shows

• Null findings: Single-dose IV and oral volunteer study: phosphate not reported [8]. EU text: not listed [4]. Oral series of 7 doses: electrolytes not reported [27].
• At-risk subgroups: Not defined. Hypophosphatemia is common in ICU populations for other reasons, which confounds the cohort setting.

Regulatory status

US IV label lists hypophosphatemia (phosphate under 1.0 mg/dL) as a warning-level event [1]. The EU product text retrieved does not list it [4]. The oral labels do not mention phosphate [2][3].

Population evidence

Licensing RCT: 2.1% versus 0% (233 versus 231) [1]. Healthy volunteers at 18 g in about 24 hours (n=39): hypophosphatemia was the most serious laboratory finding, moderate or severe in 3 (8%), asymptomatic; 2 participants (5%) were withdrawn for it and one severe event was rated treatment-related [7]. Baseline and post-dose values were not reported. The large cohorts [9][10][11][12][13][14] either did not report phosphate or did not measure it.

Mechanism

Absent from every retrieved source. Fosfomycin is described as a phosphonic acid antibiotic [26], but no study linking that structure to phosphate handling was found; a structural analogy is not a mechanism and is not used as one. The grade rests on direct measurement in two independent Tier A sources.

Why orange and not red

One source is a single-arm volunteer study, frequency in patients is low (2.1%), and no mechanism is named. The boundary is IV exposure, high dose, treatment window.

Counter-evidence held on record

Framing caution A lab disturbance during IV therapy, not evidence of dietary phosphate shortfall. No supplement is implied. Customer-visible wording must state "IV fosfomycin, high dose, during treatment."

Limits on what this shows

• Null findings: No magnesium result was retrieved from the cohort papers [9][10][11][13][14] (one states it was not reported), and magnesium is not in the licensing trial table [1]. Single-dose and 7-dose oral volunteer studies reported no magnesium result [8][27].
• At-risk subgroups: Not defined.
• Counter-evidence and confounding: No control arm, fasting confinement and unknown thresholds. One source; no replication; no effect size beyond incidence.

Regulatory status

Not named in the US IV label sections retrieved [1] or the EU text [4]. Not in oral labels [2][3].

Population evidence

29 of 39 healthy volunteers (74%) had hypomagnesemia, all mild; baseline and post-dose values were not reported; reference ranges were not given [7].

Mechanism

None named in the source [7]. A shared renal electrolyte-wasting mechanism with potassium and calcium is possible but unstated, and is not claimed.

Counter-evidence held on record

Framing caution Monitoring prompt during prolonged high-dose IV therapy only. Not a deficiency claim; no magnesium supplement is implied.

Limits on what this shows

• Null findings: The 1 g dose produced more hypocalcemia than 8 g [8]. The label comparator rate (2.6%) is close to the fosfomycin rate [1]. Not reported in the cohorts.

Regulatory status

US IV label lists hypocalcemia (corrected calcium under 8.0 mg/dL) [1]. Not in the EU text retrieved [4] or the oral labels [2][3].

Population evidence

Licensing RCT: 3.9% versus 2.6% with piperacillin-tazobactam [1]. Healthy volunteers, 18 g in about 24 hours (n=39): 16 (41%), of which 11 mild and 5 moderate; 4 withdrawn for hypocalcemia [7]. Single-dose crossover (n=30): hypocalcemia 17.9% after 1 g IV and 10.7% after 8 g IV, none treatment-related after oral 3 g; one participant had unresolved mild hypocalcemia after 8 g [8]. In the pooled bacteremia trials one of 178 combination patients stopped for hypocalcemia (under 2.0 mmol/L) [16].

Mechanism

None named in any source.

Why yellow

Three Tier A sources agree on direction, but the controlled difference is 1.3 percentage points, the single-dose crossover shows an inverse dose pattern, the measure (total versus corrected calcium, albumin) is not consistent, and no mechanism exists. Infusion volume or hydration effects are not excluded.

Counter-evidence held on record

Framing caution A laboratory observation during IV dosing. Not a calcium deficiency; no calcium supplement is implied.

Regulatory status

US label: "can be taken without regard to food" [2][3]. EU text: "Rate and extent of absorption are reduced by food"; take on an empty stomach or about 2 to 3 hours after a meal [4]. IV dosing has no food interaction.

Population evidence

Absolute bioavailability 37% fasting versus 30% fed; Cmax 26.1 versus 17.6 mcg/mL; time to peak about 2 versus 4 hours; urinary concentration peak 706 versus 537 mcg/mL; cumulative urinary excretion 1,140 versus 1,118 mg; concentrations of at least 100 mcg/mL lasted 26 hours in both states [2][3]. Later volunteer studies dosed fasting only [8][27].

Mechanism

Delayed and reduced absorption; no named transporter in the labels.

Separation interval

US label: none. EU text: about 2 to 3 hours after meals, or an empty stomach [4]. Metoclopramide lowers serum and urinary fosfomycin and should be avoided with the sachet [2][4].

Direction check

The relationship is food affecting the drug, so it is filed as a timing interaction (Hard Rule 4), capped at yellow.

Counter-evidence held on record

Framing caution Follow the dispensed product's instruction. The two regulators differ, but both report unchanged urinary recovery.