Biological and Pharmaceutical Bulletin
Online ISSN : 1347-5215
Print ISSN : 0918-6158
ISSN-L : 0918-6158
Regular Articles
A Retrospective Analysis of Risk Factors for Linezolid-Associated Hyponatremia in Japanese Patients
Ryota TanakaYosuke SuzukiYukie TakumiMotoshi IwaoYuhki SatoKazuhiko HashinagaKazufumi HiramatsuJun-ichi KadotaHiroki Itoh
Author information
JOURNAL FREE ACCESS FULL-TEXT HTML

2016 Volume 39 Issue 12 Pages 1968-1973

Details
Abstract

Linezolid is an oxazolidinone antibiotic against Gram-positive bacteria. Although thrombocytopenia is a major adverse effect of linezolid, hyponatremia also often develops after linezolid administration. This study examined the frequency of hyponatremia that developed during linezolid treatment and identified its risk factors. In this retrospective, single-center, observational cohort study, 61 hospitalized patients treated with linezolid between January 2013 and January 2015 were analyzed. Hyponatremia was defined as a sodium level of ≤134 mEq/L for the duration of linezolid treatment. Its risk factors were identified via a logistic regression analysis. Hyponatremia occurred in 11 (18.0%) patients, and it was severe in a case (a sodium level of ≤128 mEq/L). Univariate and multiple logistic regression analyses identified the plasma C-reactive protein (CRP) level before the initial administration of linezolid and the concomitant use of a potassium-sparing diuretic as the independent variables associated with the development of hyponatremia. The odds ratios were 1.081 (95% confidence interval [CI]; 1.008–1.158) (p=0.028) and 11.017 (95% CI; 1.869–64.939) (p=0.008), respectively. Before linezolid treatment, the CRP levels of the hyponatremia group were significantly higher than those of the no-hyponatremia group (p<0.001). The frequency of hyponatremia development was significantly higher in the patients who received both the potassium-sparing diuretic and linezolid (p=0.016). These results suggest that the plasma sodium levels of patients with severe inflammation who are treated with linezolid and those of linezolid-treated patients co-administered a potassium-sparing diuretic should be continuously monitored.

Linezolid is an oxazolidinone antibiotic with potent broad-spectrum activities against clinically important Gram-positive organisms, including methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant Enterococci. It binds to the bacterial ribosome 50S subunit and inhibits the initiation phase of protein synthesis by preventing the fusion of 30S and 50S ribosomal subunits.1) Owing to high tissue penetration and good bioavailability (almost 100%), it can be used as both intravenous and oral formulations.1,2) The dosage of linezolid does not generally have to be adjusted in patients with impaired renal function, even if the patient is receiving hemodialysis.3) In addition, no adjustment of dosage by body weight is needed.1) Twice daily doses of 600 mg has an appropriate therapeutic impact.1)

Thrombocytopenia is a common adverse effect of linezolid treatment in adult patients, with a reported prevalence rate of about 15–50%.4,5) Thrombocytopenia often limits the use of linezolid in clinical practice. Many retrospective studies have described various risk factors for linezolid-induced thrombocytopenia, with the main potential risk factors including renal impairment,411) prolonged treatment duration,4,5,10) hypoalbuminemia,10) and lower body weight.10,11)

Hyponatremia was observed in 7% of patients treated with linezolid in a phase III clinical trial in Japan.12) Hyponatremia is the most common impairment of body fluid and electrolyte balance encountered in the clinical setting. It can lead to a broad spectrum of clinical symptoms, from subtle to severe or even life threatening.13) Moreover, hyponatremia is associated with increased mortality and morbidity and longer hospital stays in patients presenting with a range of conditions.13) A few cases of severe hyponatremia after linezolid administration have been reported.14,15) However, there have been no studies of the frequency of linezolid-induced hyponatremia and its risk factors.

The overall aim of this study was to examine the frequency of hyponatremia that developed during linezolid treatment and to identify associated risk factors.

PATIENTS AND METHODS

Subjects

This study was conducted at the Oita University Hospital and investigated 119 hospitalized Japanese patients treated with 1200 mg (600 mg q12 h) linezolid for at least 3 d between January 2013 and January 2015. Patients who were admitted to the intensive care unit or emergency center for the duration of linezolid treatment were excluded. Patients whose serum sodium levels were not routinely determined at least twice before and during linezolid treatment, or were less than 134 mEq/L before the initial administration of linezolid, were also excluded. The final study enrolment was 61 patients. This study was approved by the Oita University Faculty of Medicine Ethics Committee (Oita, Japan). Judgement reference number: 845.

Data Collection

The patients’ medical records were reviewed retrospectively. The following clinical data, recorded before linezolid treatment, were extracted: gender, age, height, body weight, and body mass index (BMI). The recorded laboratory data included white blood cell, neutrophil, and red blood cell counts, as well as hemoglobin, platelets, C-reactive protein (CRP), serum albumin, alanine aminotransferase, alkaline phosphatase, total bilirubin, serum creatinine, blood urea nitrogen (BUN), serum potassium, serum chlorine, and serum sodium. Creatinine clearance was calculated according to the Cockcroft–Gault equation.16) The level of serum sodium was monitored until the termination of linezolid therapy.

Evaluation of Hyponatremia

Previous retrospective studies defined a sodium level ≤134 and ≤128 mEq/L after the initiation of drug therapy as hyponatremia and severe hyponatremia, respectively.1719) These same criteria of hyponatremia and severe hyponatremia were also applied in the present study.

Analysis of Risk Factors Associated with Hyponatremia

The risk factors related to hyponatremia were analyzed by univariate and multiple logistic regression, using patient background, formulation, administration period, laboratory data, and co-administered drugs as independent variables. Drugs combined with linezolid for more than 3 d were classified as co-administered drugs, and these drugs must have been used in more than six patients to be included.

Statistical Analysis

Statistical analyses were performed using Predictive Analysis Software (PASW) Statistics version 21 (SPSS Inc., Chicago, IL, U.S.A.). Data are expressed as the mean±standard deviation (S.D.). Differences between the hyponatremia and no-hyponatremia groups were analyzed by a Mann–Whitney U test or Fisher’s exact test. A value of p<0.05 was considered statistically significant.

RESULTS

In total, 119 patients were identified who satisfied the selection criteria, while 58 patients who met the exclusion criteria were excluded. The excluded patients included 30 who were in the intensive care unit or emergency center for the duration of linezolid treatment, 18 whose serum sodium levels were less than 134 mEq/L prior to linezolid treatment, and 10 whose serum sodium levels were not routinely determined. Table 1 (left column) shows the baseline characteristics of the 61 patients on the first day of linezolid treatment. The mean age of these patients was 58.4 years old, and 67% were males. Four patients were children under the age of fifteen. The mean body weight and BMI were 58.9 kg and 22.2 kg/m2, respectively. Linezolid was delivered by injection in 33 patients, orally in 19 patients, and by both injection and orally in 9 patients. The mean administration period for linezolid was 11.6 d. Linezolid was mainly administered for pneumonia; incision infections (including surgical site infection or cellulitis); sepsis; febrile neutropenia; and abscessus. MRSA was detected in 24 (39%) patients. Table 1 (right column) presents the co-administered drugs that were combined with linezolid for at least 3 d. In the study, 28 (46%) patients received a combination of linezolid and carbapenems as empirical therapy for critical illness.

Table 1. Demographic and Clinical Data of the Patients Treated with Linezolid
CharacteristicValueCo-administered drugs during linezolid therapyNumber (n)
No. of subjects (n)61Carbapenems28
No. of hyponatremia (n)11 (18%)β-Lactamase inhibitors/penicillin compounding agents6
Male/Female41/20Cephems6
Age (year)58.4±22.1Fluoroquinolones7
Height (cm)160.5±10.5Trimethoprim-sulfamethoxazole14
Weight (kg)58.9±21.1Antifungal agents12
Body mass index (kg/m2)22.5±6.2Loop diuretics22
Formulation (n)Potassium-sparing diuretics7
Injection33β-Blockers13
Oral19Angiotensin-converting enzyme inhibitors, angiotensin receptor blockers11
Both injection and oral9Calcium blockers13
Dose/Weight22.0±5.9Steroid16
Administration period (d)11.6±6.9Nonsteroidal anti-inflammatory agents22
Infection (n)Acetaminophen8
Sepsis4Proton pump inhibitors28
Febrile neutropenia3H2 Blockers11
Pneumonia19Gastric coating agents15
Incision infection18Intestinal regulators17
Abscessus3Platelet aggregation inhibitors15
Others14Warfarin8
White blood cells (cells/mm3)9.35±6.64HMG-CoA reductase inhibitors11
Neutrophils (cells/mm3)7.27±6.26Diabetes drugs7
Red blood cells (cells/mm3)3.7±1.0Expectorants12
Hemoglobin (g/dL)11.0±2.8Antihistamines11
Platelets (/µL)211.0±117.6Benzodiazepines23
C-Reactive protein (mg/dL)8.42±9.01Hyperuricemic agents14
Albumin (g/dL)2.92±0.65Potassium preparations7
Alanine aminotransferase (U/L)32.2±43.6Albumin6
Alkaline phosphatase (U/L)390.1±248.5Hepatoprotective agents7
Total bilirubin (mg/dL)0.85±0.84
Serum creatinine (mg/dL)1.04±0.94
Creatinine clearance (mL/min)91.9±69.7
Blood urea nitrogen (mg/dL)23.4±18.5
Serum potassium (mmol/L)4.05±0.56
Serum chlorine (mmol/L)102.6±4.9
Serum sodium (mmol/L)139.8±3.6
Dialysis patient (n)3
Detection of MRSA (n)24

MRSA, methicillin-resistant Staphylococcus aureus. HMG-CoA, hydroxymethylglutaryl-coenzyme A. Data are expressed as the numbers or the mean±S.D.

Of the 61 patients, hyponatremia was observed in 11 (18%), and severe hyponatremia was observed in 1 (2%). The detailed clinical data for each hyponatremia patient are shown in Table 2. The time to hyponatremia development from the initial administration of linezolid was not characteristic in 11 patients. Five of the 11 patients recovered from hyponatremia after discontinuation of linezolid. In one severe case, hyponatremia was resolved by sodium chloride treatment. A univariate analysis performed to identify the risk factors for hyponatremia revealed that the following variables reached statistical significance: age, CRP, alkaline phosphatase, BUN, platelet, and concomitant use of an antifungal agent and a potassium-sparing diuretic (Table 3). Multiple logistic regression analysis by stepwise selection of the factors extracted by the univariate analysis identified plasma CRP levels before the initial administration of linezolid and the concomitant use of a potassium-sparing diuretic as the independent variables associated with the development of hyponatremia (Table 4). The odds ratios (95% confidence interval [CI]) for CRP levels and concomitant use of a potassium-sparing diuretic were 1.081 (CI, 1.008–1.158) (p=0.028) and 11.017 (CI, 1.869–64.939) (p=0.008), respectively.

Table 2. Clinical Data of 11 Patients with Hyponatremia
Plasma sodium levels (mmol/L)The time to hyponatremia development from the initial administration of linezolid (d)The convalescent period from hyponatremia
Before linezolid treatmentMinimum value during linezolid treatmentThe rate of decrease (%)
136.8133.12.75During linezolid treatment
134.3129.63.59No determination after discontinuation of linezolid treatment
145.3126.413.07By sodium chloride treatment
137.2132.23.632 d after discontinuation of linezolid treatment
140.7133.94.84No determination after discontinuation of linezolid treatment
137.5132.53.633 d after discontinuation of linezolid treatment
141.7131.67.19During linezolid treatment
135.4132.72.0714 d after discontinuation of linezolid treatment
138.3131.84.759 d after discontinuation of linezolid treatment
134.9129.04.433 d after discontinuation of linezolid treatment
137.4132.03.914During linezolid treatment
Table 3. Univariate Analysis of Demographic and Clinical Data in the No-hyponatremia and Hyponatremia Groups
Covariate (univariate analysis)Hyponatremia groupNo-hyponatremia groupp Value
Age (year)69.755.90.058
C-Reactive protein (mg/dL)13.47.30.031
Alkaline phosphatase (U/L)539.0354.90.058
Blood urea nitrogen (mg/dL)34.121.10.033
Platelets (/µL)149.9224.40.055
Antifungal agent, n (%)5 (45.5)7 (14.0)0.018
Potassium-sparing diuretic, n (%)4 (36.4)3 (6.0)0.004
Table 4. Multivariate Analysis of Demographic and Clinical Data in the No-hyponatremia and Hyponatremia Groups
Covariate (multivariate analysis)CoefficientS.E.p ValueOdds ratio
Estimate95% CI
C-Reactive protein0.0780.0350.0281.0811.008–1.158
Potassium-sparing diuretic2.3990.9050.00811.0171.869–64.939

S.E.: standard error, 95% CI, 95% confidence interval.

As shown in Fig. 1, the plasma CRP levels were significantly higher in the hyponatremia group than in the no-hyponatremia group (p<0.001). The percentage of hyponatremia patients who were co-administered a potassium-sparing diuretic was 57.1% (4/7) (Table 5), and the percentage without co-administration of a potassium-diuretic was 13.0% (7/54). The Fisher’s exact test revealed that the frequency of the development of hyponatremia was significantly higher in the patients treated with a combination of the potassium-sparing diuretic and linezolid (p=0.016).

Fig. 1. Mann–Whitney U Test of the Plasma C-Reactive Protein (mg/mL) Levels in the No-hyponatremia and Hyponatremia Groups

The bar indicates the mean value.

Table 5. Fisher’s Exact Test of Patients Treated with the Combination of Potassium-Sparing Diuretics and Linezolid in the No-hyponatremia and Hyponatremia Groups
Hyponatremia groupNo-hyponatremia groupRatio of hyponatremia (%)p Value
Potassium-sparing diuretic (+)4357.10.016
Potassium-sparing diuretic (−)74713.0

DISCUSSION

Hyponatremia is the most common electrolyte abnormality in hospitalized patients and can be encountered in a variety of conditions.13) Specific drugs, such as diuretics, antidepressants, and antiepileptics, have been implicated in either asymptomatic or symptomatic hyponatremia.20) The aim of this study was to detect the frequency of linezolid-induced hyponatremia and to identify the risk factors in a retrospective analysis of our hospital’s medical records.

A clinical phase III trial in Japan reported a frequency for hyponatremia during linezolid treatment of 7%.12) By contrast, in the present study, the frequency was 18% (Table 1). The higher frequency in the current study may be due to the use of a sodium level of ≤134 mEq/L to denote hyponatremia. We did not check for symptoms caused by reduced sodium, such as lethargy, headache, dizziness, nausea, vomiting, restlessness, and disorientation. This is one of the limitations of this retrospective analysis.

In the current study, the plasma CRP level before the initial administration of linezolid was identified as an independent variable associated with linezolid-induced hyponatremia (Table 4). Hyponatremia is a recognized complication of several inflammatory diseases, with or without infection.21) In the presence of inflammation characterized by a high level of CRP, large amounts of inflammatory cytokines, including tumor necrosis factor-α, interleukin (IL)-1β, and IL-6, are secreted into the systemic circulation, triggered by endotoxins.22) A previous review suggested that these cytokines, especially IL-6, induced antidiuretic hormones (ADHs), which regulate the secretion of ADH under physiological conditions.23) Impaired renal water excretion, either due to low extracellular fluid volume or inappropriate secretion of ADH (SIADH), is a common cause of hyponatremia. Thus, SIADH, caused by the secretion of IL-6 in the inflammatory reaction, may have been associated with the development of hyponatremia in the patients treated with linezolid. As shown in Fig. 1, the CRP levels in the present study were significantly higher in the hyponatremia patients than in the patients without hyponatremia.

Baik et al. reported that linezolid was a probable cause of SIADH.15) In their study, SIADH-induced hyponatremia occurred following linezolid treatment for more than 3 weeks, and the hyponatremia resolved after the cessation of linezolid. However, recent in vitro24,25) and in vivo26) studies demonstrated that linezolid had anti-inflammatory effects, in addition to its antimicrobial effects. In those studies, linezolid reduced the production of cytokines from peripheral blood cells stimulated with lipopolysaccharide.24,25) Matsumoto et al. demonstrated that linezolid treatment resulted in a significant and concentration-dependent suppression of paw edema induced by carrageenan.26) Hence, it is unlikely that linezolid-related hyponatremia is associated with SIADH induced by inflammatory cytokines. As shown in Table 4, the logistic regression analysis in the present study gave an odds ratio for CRP levels that was not large (1.081). Taking the findings in the literature and the odds ratios in the current study into consideration, the increased plasma CRP levels in linezolid-related hyponatremia may be due to pre-existing disease and the severity of infection, rather than the drug itself.

Table 4 shows that the concomitant use of a potassium-sparing diuretic was an independent variable associated with linezolid-induced hyponatremia. Additionally, the frequency of the development of hyponatremia was significantly higher in the patients who received a combination of a potassium-sparing diuretic and linezolid (Table 5). Five patients were treated with spironolactone as a potassium-sparing diuretic and two patients were treated with potassium canrenoate. Two of the five patients treated with spironolactone and both patients treated with potassium canrenoate had an onset of hyponatremia (data not shown). Potassium-sparing diuretic agents exert anti-diuretic activities by suppressing the reuptake of water by the distal renal tubules.27) Triamterene inhibits local transport, which allows the influx of sodium ions and its exchange against potassium or hydrogen ions. Aldosterone antagonists inhibit the activities of aldosterone at the receptor level and impair the reabsorption of sodium ions and water and their exchange with potassium ions. Due to the potencies of potassium-sparing diuretic agents, both hyponatremia and hyperkalemia are side effects of these drugs.28) In common with potassium-sparing diuretic agents, loop diuretic agents exert antidiuretic effects by the inward transport of sodium, potassium, and chloride ions, as well as water, into tubular cells.28) Thus, hyponatremia is also an adverse side effect of these agents.28) Despite these findings, loop diuretics, unlike the potassium-sparing diuretics, were not identified as an independent variable in the present study (hyponatremia group vs. no hyponatremia group (54.5% vs. 32.0%), univariate analysis: p=0.159, data not shown). The site of action of potassium-sparing diuretics (distal renal tubules) differs from that of loop diuretics (Henle’s loop). Thus, linezolid may exert some influence on the distal renal tubules.

Suzuki et al. described a case of a male patient with acute myeloid leukemia who developed hyponatremia after linezolid administration.14) The patient’s hyponatremia was attributed to renal salt-wasting syndrome (RSWS), based on a calculation of the amount of sodium intake and amount of sodium excreted from the kidneys. The authors speculated that linezolid exacerbated a microscopic renal tubular disorder, which developed after chemotherapy. Hence, by affecting the distal renal tubules, RSWS seems to be another possible cause of linezolid-related hyponatremia. In the present study, in most cases, the patients’ urine sodium levels were not measured, which precluded a determination of whether the patients had RSWS. This is one of the limitations of this study. Additionally, since potassium-sparing diuretics on their own can have the side effect of hyponatremia, no clear linkage can be established between the concomitant use of these drugs and linezolid-induced hyponatremia. However, potassium-sparing diuretics were administered before the start of linezolid infusion in all cases in the hyponatremia group that received the combination of linezolid and a potassium-sparing diuretic. All of these patients recovered from hyponatremia after discontinuation of linezolid (one patient recovered after treatment with sodium chloride), and 3 of the 4 patients were also continuously administered a potassium-sparing diuretic after their recovery from hyponatremia (Table 2). Thus, hyponatremia development was not thought to be due to the potassium-sparing diuretics themselves, but to their combination with linezolid treatment.

In conclusion, this retrospective study suggested that patients with severe inflammation or those with concomitant use of a potassium-sparing diuretic tended to develop hyponatremia during linezolid treatment. This investigation is the first to report the risk factors for linezolid-related hyponatremia. Although the mechanism of linezolid-related hyponatremia has not been elucidated, severe inflammation and the concomitant use of potassium-sparing diuretics seem to be risk factors. In the presence of these risk factors, the plasma sodium levels of patients treated with linezolid should be continuously monitored.

Acknowledgment

The authors received no financial support for the research, authorship and publication of this article.

Conflict of Interest

The authors declare no conflict of interest.

REFERENCES
 
© 2016 The Pharmaceutical Society of Japan
feedback
Top