Beta-lactam-associated hypokalemia

Abstract

The aim of this narrative review was to discuss the literature on β-lactam antibiotic-associated hypokalemia, a potentially life-threatening electrolyte disorder. The PubMed, Web of Science, Cochrane Library, and Scopus databases were searched for articles published between 1965 and 2023, using the following terms: ‘hypokalemia’ OR ‘potassium loss’ OR ‘potassium deficiency’ AND ‘beta-lactams’ OR ‘penicillin’ OR ‘penicillin G’ OR ‘cephalosporins’ OR ‘ceftazidime’ OR ‘ceftriaxone’ OR ‘flucloxacillin’ OR ‘carbapenems’ OR ‘meropenem’ OR ‘imipenem’ OR ‘cefiderocol’ OR ‘azlocillin’ OR ‘ticarcillin’. Additional search terms were ‘hypokalemia’ AND ‘epidemiology’ AND ‘ICU’ OR ‘intensive care unit’ OR ‘ER’ OR ‘emergency department’ OR ‘ambulatory’ OR ‘old’ OR ‘ageing population’, and experimental (animal-based) studies were excluded. A total of eight studies were selected and discussed, in addition to nine case reports and case series. Both older and currently used β-lactam antibiotics (e.g., ticarcillin and flucloxacillin, respectively) have been associated with therapy-related hypokalemia. The incidence of β-lactam antibiotic-associated hypokalemia may be as high as 40%, thus, the issue of β-lactam-associated hypokalemia remains clinically relevant. Although other causes of hypokalemia are likely to be diagnosed more frequently (e.g., due to diuretic therapy or diarrhea), the possibility of β-lactam-induced renal potassium loss should always be considered in individuals with so-called ‘unexplained hypokalemia’.

Introduction

Hypokalemia is a potentially life-threatening electrolyte disorder, with an estimation that up to 11% of all hospitalized patients in the USA develop hypokalemia during their treatment. ¹ Potassium is mainly stored in the intracellular compartment and plays a role in stabilizing the membrane potential. A decrease in the extracellular potassium concentration (hypokalemia) causes cellular hyperpolarization, leading to decreased cellular excitability, ² which puts affected patients at a significantly higher risk of cardiac arrhythmia, particularly tachyarrhythmia. Other complications may include impaired intestinal and bladder function, with severe cases potentially resulting in paralytic ileus or even muscular paralysis. ³

A typical western diet provides a daily potassium intake range of 2.8–3 g/day. ⁴ Most potassium (>90%) is eliminated through the kidneys, but in individuals with impaired kidney function, intestinal elimination of potassium may increase to over 10%. ⁵ The uptake of potassium into cells is stimulated by β2-adrenergic agonists and insulin. Additionally, a lower extracellular hydrogen ion concentration (alkalosis) is associated with increased movement of potassium into the intracellular space. ⁶ Hypokalemia, or low potassium levels, may be caused by imbalances between intake and elimination, as well as disorders affecting the distribution of potassium within cells.

Diuretics, specifically thiazides and loop diuretics, are the most common causes of drug-induced hypokalemia. However, other medications may also cause this condition. The literature suggests that β-lactam antibiotics, which are commonly used in hospitals and ambulatory medicine worldwide, have been associated with hypokalemia.^7,8 These medications may need to be administered for up to 6 weeks or even longer in certain infectious diseases, for example, in cases of endocarditis. ⁹

The purpose of the present narrative review was to discuss the epidemiology and pathogenesis of hypokalemia associated with β-lactam antibiotics, in order to increase physicians' awareness of the risk of hypokalemia linked to β-lactam-antibiotic treatment, with the aim of improving the prognosis of patients who need a β-lactam-antibiotic-based regimen.

Pathophysiological considerations

Daily potassium intake ranges from 2.8 to 3 g/day in Central Europe and from 2.4 to 2.6 g/day in the USA. ⁴ The two main causes of hypokalemia relate to external or internal imbalances, ⁶ with external balance disturbances characterized either by reduced potassium intake or increased losses. Reduced potassium intake may result from extreme deficiency or malnutrition, for example, in patients with anorexia nervosa. ¹⁰ Additionally, any disruption to the absorptive function of the intestine can decrease potassium absorption (malabsorption syndrome). Increased potassium losses may occur either through the gastrointestinal tract or the kidneys. In the first case, diarrheal diseases are usually responsible, while renal losses often occur concomitantly with the use of loop diuretics or thiazides. ¹¹ Another underestimated cause is the recovery phase of acute kidney injury. Although the polyuria that occurs during this phase is generally a favorable prognostic sign, ¹² the reabsorptive capacity of the renal tubules is still inadequate, and glomerular-filtered and tubular-secreted potassium cannot be reabsorbed back into the blood. Less common causes of increased renal potassium loss are partial dysfunctions of the renal tubes, specifically renal tubular acidosis types 1 and 2. ¹³ Disturbances in tubular potassium reabsorption may also occur as a result of acute or chronic interstitial nephropathies. Conversely, prolonged hypokalemia may also lead to interstitial nephritis. ¹⁴ External balance disturbances that are associated with hypokalemia are usually accompanied by a reduction in total body potassium stores. In the context of internal balance disorders, this is not necessarily the case. The crucial factor is rather a distribution disorder between the intra- and extracellular space. Increased cellular potassium uptake is manifested, for example, by the use or overdose of insulin or β2-adreceptor agonists. ¹⁵ Additionally, metabolic alkalosis often perpetuates hypokalemia. ¹⁶ Other possible causes of distributive hypokalemia include treatment of megaloblastic anemia with vitamin B12 or folic acid. The stimulation of erythropoiesis is associated with consumption of intracellular substrates, such as potassium, ¹⁷ and the same phenomenon may occasionally be observed in rapidly proliferating leukemias or lymphomas. ⁶ Finally, rare, genetically determined diseases may be associated with an internal potassium imbalance, such as familial hypokalemic periodic paralysis. ¹⁸ Primary and secondary hyperaldosteronism cause a combined external/internal balance disorder, in that aldosterone reduces renal potassium reabsorption, but most likely also stimulates the uptake of potassium intracellularly. ¹⁹

Epidemiology of hypokalemia

The reported prevalence of hypokalemia varies significantly. The PolSenior study, first published in 2011, ²⁰ investigated medical, socio-economic and psychological aspects of the aging Polish population, and examination of a cohort of 4 654 participants from this study (mean age, 76.5 years) found hypokalemia in only 39 individuals (0.84%). ²¹ The Stockholm CREAtinine Measurements (SCREAM) project was initiated in 2010 with the aim of continuously monitoring the excretory kidney function of individuals in the greater Stockholm area. ²² Analysis of data extracted from the SCREAM cohort, including nearly 370 000 individuals with an observation period of 3 years, revealed that hypokalemia was detected in 13.7% of all participants. ²³ Furthermore, 33% of these cases experienced recurrent episodes of hypokalemia. An investigation of nearly 60 000 women and men who participated in the US National Health and Nutrition Examination Survey (NHANES) from 1999 to 2016, a longitudinal assessment of the health status and nutritional condition of adults and children in the USA through a combination of patient interviews and physical examinations, found an increase in the prevalence of hypokalemia from an average of 3.78% in 1999 to 11% in 2016. ¹

The incidence of de novo hypokalemia also depends on the medical situation, as evidenced by the observation that nearly 50% of emergency patients have hypokalemia.^24,25 In a Japanese study from 2022, ²⁶ in which 21 616 patients who attended the emergency department between 2012 and 2019 were retrospectively analyzed, severe hypokalemia (plasma potassium below 2.5 mmol/L) was found in 0.4% of all cases. Exact numerical data on moderate or mild hypokalemia were not provided in the study. Surprisingly, few systematic studies on the prevalence of hypokalemia in critically ill patients have been published. A Turkish study from 2003 reported a de novo hypokalemia rate of 40% among 440 patients in a surgical intensive care unit. ²⁷

The frequency of this potentially life-threatening electrolyte disorder is clearly shown to vary significantly depending on the age of the individuals being examined, the medical circumstances (outpatient environment versus hospital), and ultimately the invasiveness of the medical therapy itself.

Search strategy and results

With the aim of discussing the epidemiology and pathogenesis of hypokalemia associated with β-lactam antibiotics, the following databases were searched for studies involving β-lactam-associated hypokalemia published between 1965 and 2023: PubMed, Web of Science, Cochrane Library, and Scopus. The following search terms were utilized: ‘hypokalemia’ OR ‘potassium loss’ OR ‘potassium deficiency’ AND ‘beta-lactams’ OR ‘penicillin’ OR ‘penicillin G’ OR ‘cephalosporins’ OR ‘ceftazidime’ OR ‘ceftriaxone’ OR ‘flucloxacillin’ OR ‘carbapenems’ OR ‘meropenem’ OR ‘imipenem’ OR ‘cefiderocol’ OR ‘azlocillin’ OR ‘ticarcillin’. Additional search terms were ‘hypokalemia’ AND ‘epidemiology’ AND ‘ICU’ OR ‘intensive care unit’ OR ‘ER’ OR ‘emergency department’ OR ‘ambulatory’ OR ‘old’ OR ‘ageing population’. Experimental (animal-based) studies were excluded. A total of eight studies in addition to nine case reports and case series were selected and discussed (summarized in Table 1).^28–44

Table 1.

Published studies and case reports/case series on β-lactam-associated hypokalemia.

Clinical studies (chronological order)
Reference	Design	Outcome
Brunner and Frick, 1968 28	Penicillin G therapy in 2 patients with endocarditis	Increased urinary potassium excretion with hypokalemia
Nanji and Lindsay, 1982 29	Follow-up of 16 individuals with ticarcillin therapy, variable doses (<10 – >18 g daily)	Higher incidence of hypokalemia in patients receiving doses above 18 g daily, volume depletion promotes hypokalemia
Parry et al., 1983 30	Azlocillin therapy in 631 patients with urogenital or other systemic infections	Hypokalemia in 0.5%
Gibson et al., 1989 31	Prospective, randomized study in 102 individuals with neutropenia-associated systemic infection; ceftazidime or azlocillin, both combined with flucloxacillin	Lower rate of moderate to severe hypokalemia with the use of ceftazidime
Kaye at al., 2019 32	Comparison between fosfomycin and piperacillin/tazobactam in the therapy of complicated urinary tract infections in 465 patients	Hypokalemia in 15 fosfomycin- and 3 piperacillin/tazobactam-treated patients
Van der Heijden et al., 2019 33	Retrospective comparison between 77 flucloxacillin-treated and 84 ceftriaxone-treated patients	De novo-hypokalemia in 42% of the flucloxacillin and 14% of the ceftriaxone group (P < 0.001)
Wunderink et al., 2021 34	Prospective randomized study, comparison between cefiderocol and meropenem, both combined with linezolid; treatment of gram-negative pneumonia, either ventilator-associated or nosocomial (non-ventilator-associated) or healthcare-associated	Incidence of hypokalemia higher in the meropenem group (23 out of 150 patients [15%])
Leegwater et al., 2022 35	Retrospective, monocentric study including analysis of 835 patients for flucloxacillin-associated hypokalemia	De novo-hypokalemia in 23.7%
Case reports/series (alphabetical order)
Reference	Key finding
Cabizuca and Desser, 1976 36	Hypokalemic metabolic alkalosis upon carbenicillin therapy of pulmonary pseudomonas infection
De Backer and Hannon, 2018 41	Severe hypokalemia in high dose intravenous flucloxacillin treatment
Gill et al., 1977 37	Hypokalemic metabolic alkalosis in high-dose intravenous ampicillin therapy
Hussain et al., 2010 40	Severe hypokalemia secondary to piperacillin/tazobactam
Kumar et al., 2020 43	Cephalexin-induced acute interstitial nephritis with hypokalemic periodic paralysis
Mohr et al., 1979 38	Nafcillin-induced potassium loss with subsequent hypokalemia
Schlaeffer 1988 39	Oxacillin-associated hypokalemia in staphylococcus bacteremia
Tai et al., 2020 44	Severe acute hypokalemia, associated with piperacillin/tazobactam
Toro et al., 2019 42	Hypokalemia in ceftozolane/tazobactam therapy (Spanish)

β-lactam-associated hypokalemia

β-lactam-associated hypokalemia is believed to be a class-related effect. As non-reabsorbable anions, β-lactams create a negative gradient on the inner side of the cortical collecting duct, which leads to increased excretion of potassium by the kidneys, resulting in hypokalemia.^28,45 Studies relating to β-lactam-associated hypokalemia are reviewed below in chronological order of publication, from the oldest to the most recent.

One of the first reports of β-lactam-associated hypokalemia was published in 1968. ²⁸ The article describes the manifestation of a hypokalemic alkalosis in two patients with subacute bacterial endocarditis. Both patients received 100 million units of penicillin G each, and the electrolyte imbalance manifested itself within a few days, accompanied by inadequate potassium excretion in the urine. In this study, the effect of penicillin G as a non-absorbable anion was already established.

In 1982, Nanji and Lindsay conducted a study among 16 patients who received intravenous ticarcillin treatment. ²⁹ Four patients were administered doses below 10 g/day, while nine received doses exceeding 18 g/day, and in the latter group, six patients experienced significant hypokalemia. Dehydration or hypovolemia were observed to lead to potassium loss, whereas three patients who received the same doses, but who had a simultaneous syndrome of abnormal vasopressin secretion, did not develop hypokalemia. Consequently, volume administration was proposed as a preventive strategy to avoid ticarcillin-associated hypokalemia.

The tolerability of azlocillin was investigated in a study published in 1983. ³⁰ A total of 631 patients with urogenital or other systemic infections were included, with a daily azlocillin dose of 260 mg/kg and an average treatment duration of approximately 11 days. The tolerability of the medication was assessed as good, with hypokalemia occurring in only 0.5% of patients.

In a prospective, randomized study in 102 patients with neutropenia-associated systemic infection, where neutropenia was either a manifestation of hematologic malignancy or a result of chemotherapy, patients were treated with flucloxacillin combined with either ceftazidime or azlocillin. ³¹ An infection caused by gram-positive bacteria was initially suspected in 25% of all treated individuals. At 96 h from starting treatment, the response rates did not differ significantly between treatments (ceftazidime 59.6% versus azlocillin 44%). However, moderate to severe hypokalemia was significantly less frequently diagnosed with the use of ceftazidime.

The efficacy of intravenous fosfomycin therapy for complicated urinary tract infections, including pyelonephritis, was evaluated by Kaye et al. ³² Affected patients were treated with 6 g of fosfomycin every 8 h, and compared with those who received a treatment regimen of piperacillin/tazobactam (4.5 g every 8 h). The total therapy duration was 7 days, with the option to extend to 14 days in case of manifest bacteremia. A total of 465 patients were randomized, with 233 in the fosfomycin group and 231 in the piperacillin/tazobactam group. The clinical cure rates at 19–21 days were comparable between the groups (90.8% for fosfomycin and 91.6% for piperacillin/tazobactam). Severe adverse events occurred at similar frequencies in both groups, with hypokalemia observed in 15 cases in the fosfomycin group and 3 cases in the other group. All cases of hypokalemia were mild to moderate.

In a retrospective cohort study published in 2019, ³³ both flucloxacillin-treated (n = 77) and ceftriaxone-treated patients (n = 84) were evaluated. While hypokalemia occurred in both treatment groups, it manifested significantly more often under flucloxacillin therapy (42% versus 14%, P < 0.001). Additionally, the electrolyte disorder occurred more frequently in women than in men.

A multicenter, prospective randomized study, published in 2021, ³⁴ compared the outcome and tolerability of therapy with cefiderocol or meropenem. Cefiderocol has been developed with stability against all β-lactamases and activity against carbapenem-resistant multidrug-resistant gram-negative bacteria. ⁴⁶ The study included patients with gram-negative pneumonia, either ventilator-associated or nosocomial (non-ventilator-associated) or healthcare-associated. ³⁴ Cefiderocol was administered once daily, while meropenem was given every 8 h, with a treatment duration of 7–14 days. Additionally, all patients received intravenous linezolid for 5 days. The primary endpoint was mortality on day 14 after initiation of treatment. The cefiderocol group comprised 148 patients and 152 patients were included in the meropenem group. There was no statistically significant between-group difference in primary endpoint (mortality on day 14: cefiderocol, 12.4% and meropenem, 11.6%), however, a significantly higher incidence of hypokalemia was diagnosed in the meropenem group (23 out of 150 patients [15%]).

In a retrospective, monocentric study, Leegwater et al. ³⁵ analyzed the incidence of flucloxacillin-associated hypokalemia in patients, recruited between January 2017 and October 2020, who received flucloxacillin intravenously for more than 24 h. Patients with pre-existing hypokalemia were not included. Among the 835 patients treated with flucloxacillin, 23.7% developed hypokalemia, with a more than 4-fold higher risk in women whose daily dose exceeded 8 g. Lower dose ranges were not associated with gender differences. Other identified risk factors included older age, lower body weight, low-normal baseline potassium, longer treatment duration, and combination antibiotic therapies.

In addition to the abovementioned studies, several case studies and case series have reported on hypokalemia in patients treated with β-lactam antibiotics (summarized in Table 1).^36–44

Three guidelines for the treatment of important clinical infections should be considered at this point: the 2018 updated S3 guideline ‘Epidemiology, Diagnosis, and Treatment of Nosocomial Pneumonia in Adults’ (German Society for Anaesthesiology and Intensive Care Medicine, the German Society for Infectious Diseases, the German Society for Hygiene and Microbiology, the German Respiratory Society and the Paul-Ehrlich-Society for Chemotherapy, the German Radiological Society and the Society for Virology); ⁴⁷ the 2017 updated ‘German Clinical Guideline on Epidemiology, Diagnostics, Therapy, Prevention, and Management of Uncomplicated Urinary Tract Infections in Adult Patients’;^48,49 and finally, the 2023 renewed guideline on endocarditis management from the European Society of Cardiology. ⁹ The recommendations for pneumonia therapy mention at least 10 distinct β-lactam antibiotics. The selection of drugs depends on the site of pneumonia development, the severity of inflammation, and the cumulative morbidity of the patients. The urological guideline lists 12 different β-lactam antibiotics alone that can be used to treat uncomplicated pyelonephritis in premenopausal women, depending on the response to therapy and the disease severity. The endocarditis guideline also includes numerous β-lactams, with treatment durations of up to 6 weeks in some cases (e.g. prosthetic valve endocarditis, penicillin G, or amoxicillin, or ceftriaxone). These guidelines emphasize that β-lactams will continue to be a fundamental component of antibiotic regimens for the control of severe, sometimes life-threatening infections. Although the side-effect profile of β-lactams is generally considered favorable, different preparations within this group, whether penicillin derivatives, cephalosporins, or carbapenems, may cause hypokalemia, particularly with prolonged treatment duration. However, there appear to be significant substance-related differences in the risk of β-lactam antibiotic-associated hypokalemia. Parry et al. ³⁰ reported an incidence of 0.5% (with azlocillin), while van der Heijden et al. ³³ found an almost 80-fold higher incidence (42%) with use of flucloxacillin. In current clinical practice, azlocillin is rarely used, at least not in Central Europe, however, flucloxacillin is a widely used medication in clinical practice, specifically for the systemic treatment of sensitive Staphylococcus infections. ⁵⁰

Conclusion

The issue of β-lactam-associated hypokalemia remains clinically relevant. Although other causes of hypokalemia are likely to be diagnosed more frequently (e.g. diuretic therapy or diarrhea), the possibility of β-lactam-induced renal potassium loss should always be considered in individuals with so-called ‘unexplained hypokalemia’.