Polymyxin B–Induced Bartter‑Like Tubulopathy: A Case Report of Severe Potassium and Magnesium Wasting

Introduction

Bartter syndrome refers to a group of rare autosomal recessive salt wasting tubulopathies affecting the thick ascending limb (TAL) of the loop of Henle, characterized by hypokalemia, metabolic alkalosis, and variable disturbances in magnesium handling. Reported prevalence differs across genetic subtypes, reflecting the heterogeneity of the disorder rather than a single uniform incidence [¹].

Bartter-like electrolyte disturbances have also been described in association with nephrotoxic medications, such as antibiotics, including polymyxins, used to treat multidrug-resistant Gram-negative infections such as Pseudomonas aeruginosa, Klebsiella pneumoniae, Enterobacteriaceae, and Acinetobacter baumannii [²,³]. Polymyxins are internalized by proximal tubular epithelial cells through megalin and PEPT2 mediated endocytosis, leading to oxidative and mitochondrial injury [⁴,⁵]. Although this mechanism primarily injures the proximal tubule, the resulting impairment of sodium and chloride reabsorption, along with activation of tubuloglomerular feedback, can produce downstream functional changes that mimic TAL dysfunction, thereby generating a Bartter-like clinical phenotype.

We describe a 77 year old woman who developed a Bartter-like syndrome shortly after receiving polymyxin B, with complete resolution following discontinuation of the drug.

Case Report Methods

Clinical information was obtained through review of the electronic medical record, including physician and nursing documentation, laboratory data, imaging studies, and the medication administration record. Electrolyte replacement was quantified by daily potassium(40-160 meq/day) and magnesium(1-3 grams/day) doses administered via intravenous and enteral routes. No diuretics, laxatives, beta agonists, insulin infusions, corticosteroids, amphotericin, aminoglycosides, or nutrition-related factors (including changes in tube feed formulation) were present during the period of interest unless otherwise specified. Institutional review board (IRB) approval was waived for this single-patient case report, and consent was obtained from the patient in accordance with institutional policy.

Case Presentation

A 77-year-old female patient with a past medical history of COPD, hypotension, and respiratory failure status post tracheostomy and PEG tube placement was nonverbal but responsive to noxious stimuli. She was sent to the ED from a nursing home after findings on a HIDA scan. Nursing home records indicated elevated liver function tests (LFTs) and a HIDA scan showing cystic duct obstruction.

On initial presentation, the patient was afebrile (temperature 98.2°F), with a heart rate of 83 bpm, blood pressure of 148/75 mmHg, and oxygen saturation of 100% on mechanical ventilation (MV) settings of 18/400/40/5. Physical examination was unremarkable; the abdomen was non-tender and had a PEG tube in place. The initial laboratory values are shown in Table 1.

Gastroenterology and General Surgery teams evaluated the patient. HIDA Scan images were reviewed, and it suggested chronic cholecystitis versus cystic duct obstruction due to a stone. The next day, the patient became septic with a BP in the 80/50 mm Hg, HR: 108 bpm, and temperature: 95.5°F. Septic workup was sent, and the patient was started on norepinephrine after receiving 3 liters of IV fluids. Chest X-ray Figure 1 showed bilateral multifocal patchy opacities, left greater than right, in the lower lobe. The infectious disease team was consulted, and the patient was started on broad-spectrum antibiotics with meropenem 1 g IV q8h and vancomycin 750 mg IV q12h.

A few days later, sputum culture grew Acinetobacter and Stenotrophomonas, sensitive to Bactrim, which was added (5 mg/kg IV q8h). Urine culture grew E. coli and Proteus, both of which were sensitive to meropenem. Two days later, due to worsening clinical status, meropenem was changed to ampicillin/sulbactam 3 G IV q6h and polymyxin B (2 million units IV once, then 1 million units IV q12h), patient’s weight was 106kg and Polymyxin B dosing was selected according to standard weight based recommendations, and no renal function based adjustment was required as the patient’s kidney function remained stable throughout therapy. At that time, potassium was 4.2 mmol/L, and magnesium was 2.0 mg/dL. Five days after initiation of polymyxin, serum potassium decreased to 2.9 mmol/L, and by the sixth day, magnesium had fallen to 1.11 mg/dL, necessitating replacement.

For this patient, drug induced context, “Bartter-like syndrome” was defined as the presence of: Hypokalemia, Hypomagnesemia, Metabolic alkalosis, Polyuria in the absence of osmotic diuresis, and inappropriately high urinary potassium excretion despite hypokalemia. Serial electrolytes during the period of interest showed stable serum sodium (136 – 140 mmol/L) and chloride (98 – 102 mmol/L), but metabolic alkalosis developed, with bicarbonate rising to 35 mmol/L (baseline 22 mmol/L) and serum chloride decreasing to 90 mmol/L. The patient developed polyuria of 3 – 4 mL/kg/h, corresponding to a daily urine output of 4-5.8 L/day. Fluid balance was otherwise neutral, and vasopressor requirements were stable. Serum glucose, urea, and osmolar gap were not elevated, excluding osmotic diuresis. Kidney function remained stable throughout polymyxin B therapy, with creatinine ranging from 0.76 to 0.98 mg/dL (baseline 0.74 mg/dL), and no evidence of acute kidney injury.

This hypokalemia and hypomagnesemia worsened during polymyxin B therapy (Figure 2, Figure 3). A spot urine sample obtained during documented hypokalemia (serum potassium 2.6 mmol/L) showed a urine potassium of 48.9 mmol/L, urine creatinine of 31.8 mg/dL, and urine chloride of 23 mmol/L. The calculated urine K/Cr ratio was 154 mmol/g, consistent with inappropriate renal potassium wasting in the setting of metabolic alkalosis. All antibiotics were discontinued upon resolution of sepsis. Urinary potassium significantly decreased after polymyxin B was discontinued on day 14. Baseline serum potassium was 2.3 – 3.0 mmol/L during the course of polymyxin B therapy and rose to 3.9 mmol/L after discontinuation. Daily potassium and magnesium levels, replacement amounts, polymyxin exposure dates, and electrolyte nadirs are summarized in Table 2.

48 hours after stopping polymyxin B, serum potassium improved significantly, and polyuria resolved. After electrolyte levels returned to baseline, the patient was discharged to a short-term rehabilitation facility with daily potassium and magnesium supplementation.

Discussion

Bartter syndrome is classically an inherited renal tubular disorder characterized by defective sodium – chloride reabsorption in the thick ascending limb (TAL), leading to renal salt wasting, hypokalemia, hypomagnesemia, and metabolic alkalosis. In contrast, “Bartter-like syndrome” in adults refers not to a true TAL channelopathy but to a phenocopy of these electrolyte and acid – base abnormalities arising from secondary tubular injury. Several nephrotoxic agents, including aminoglycosides and polymyxins, have been implicated in producing this phenotype through proximal tubular toxicity rather than primary TAL dysfunction [²,³].

Polymyxin B is filtered at the glomerulus and taken up by proximal tubular epithelial cells via megalin- and PEPT2-mediated endocytosis, triggering oxidative and mitochondrial injury [⁴]. Although this mechanism primarily affects the proximal tubule, downstream impairment of sodium and chloride handling, activation of tubuloglomerular feedback, and secondary alterations in distal potassium and magnesium transport can collectively mimic the electrolyte pattern of Bartter syndrome. Thus, the term “Bartter-like” in this case reflects a functional resemblance rather than a mechanistic equivalence.

Our patient developed progressive hypokalemia, hypomagnesemia, metabolic alkalosis, and polyuria shortly after initiation of polymyxin B, despite previously stable electrolytes. The early onset on day 5, earlier than the median 10 day onset reported in the literature (range 7 – 18 days) [²], may be explained by several susceptibility factors: advanced age, critical illness with sepsis, baseline hypotension, chronic ventilator dependence, and cumulative exposure to other nephrotoxic stressors (including vasopressors and broad spectrum antibiotics). Reduced renal reserve in older adults may also accelerate tubular accumulation of polymyxin B, precipitating earlier toxicity.

A structured adverse drug reaction assessment further supports the assessment of causality. Using the Naranjo algorithm, the event scored 8, consistent with a “probable” adverse drug reaction. Supporting elements included:

This structured approach strengthens the attribution beyond temporal association alone.

Polymyxin-associated nephrotoxicity is reported in 30 – 60% of treated patients, depending on cohort characteristics and the AKI definition used [⁶]. Importantly, these estimates primarily refer to serum creatinine – based AKI, whereas electrolyte-wasting phenotypes – such as Bartter-like presentations – may occur with or without overt AKI. Our patient maintained stable creatinine throughout therapy, underscoring that tubular electrolyte loss can be an isolated manifestation of polymyxin toxicity.

Differentiating Bartter-like syndrome from other drug induced tubulopathies is essential. Gitelman like features (hypocalciuria, hypomagnesemia with metabolic alkalosis) and Fanconi type proximal tubular dysfunction (phosphaturia, glycosuria, aminoaciduria, bicarbonate wasting) were not present. Serum phosphate remained stable, there was no glycosuria or proteinuria, and bicarbonate was elevated rather than low, arguing against Fanconi syndrome. Urine chloride was appropriately elevated, supporting a chloride resistant metabolic alkalosis consistent with a Bartter-like pattern.

Across published cases, electrolyte abnormalities typically resolve within days of discontinuing polymyxin therapy [⁷,⁸]. Our patient demonstrated rapid correction of potassium and magnesium levels and resolution of polyuria within 48 hours of drug withdrawal, mirroring prior reports and reinforcing the diagnosis.

Inherited Bartter syndromes typically present in infancy or childhood and may include hypercalciuria and prostaglandin mediated symptoms [⁹,¹⁰]. In contrast, acquired Bartter-like syndromes emerge abruptly after a clear exposure and resolve with removal of the offending agent, as seen here. Given the rarity of inherited Bartter syndrome, a drug-induced etiology should remain high on the differential in adults with new-onset salt wasting.

Conclusion

Bartter-like syndrome with electrolyte wasting can develop rapidly and resolve with timely discontinuation of the drug. Early monitoring of electrolytes, including magnesium, from day 1 of therapy is essential. Development of persistent hypokalemia, hypomagnesemia, metabolic alkalosis, or polyuria should prompt nephrology involvement and consideration of alternative antimicrobials when feasible. Proactive surveillance can prevent complications and support safer antibiotic use.