Fatal Hypoglycemia Associated With Levofloxacin

Lawrence V. Friedrich, Pharm.D.; Richard Dougherty, M.D


Pharmacotherapy. 2004;24(12) 

In This Article


Hypoglycemia is an infrequent adverse event that has occurred in patients receiving fluoro-quinolones. Published reports are available for ciprofloxacin, gatifloxacin, and clinafloxacin,[1,2,3,4,5,6,15] but there are no published reports, to our knowledge, of hypoglycemia with levofloxacin. However, symptomatic hypoglycemia has been reported in patients receiving levofloxacin, usually in those with diabetes receiving concomitant treatment with antidiabetic agents or insulin (P. F. Helleis, R.N., personal communication, Ortho-McNeil Medical Information, Clinical Communications, February 13, 2004). Measured blood glucose levels as low as 35 mg/dl were reported in levofloxacin's new drug application.[8]

Hypoglycemia as a serious adverse event in adults receiving treatment for community-acquired pneumonia was recently reported.[16] In addition, the Food and Drug Administration, through its MedWatch adverse event reporting program, has received reports of hypoglycemia with all fluoroquinolones, including levofloxacin.[21]

Similar to patients described in other case reports of fluoroquinolone-associated hypoglycemia,[1,2,3,4,5,6] our patient was elderly, had renal dysfunction, was receiving an oral sulfonylurea, and had a history of type 2 diabetes and several other medical conditions. In those case reports, the average patient age was 80 years (range 70-94 yrs). All patients were receiving a concurrent oral hypoglycemic agent; 75% were taking glyburide.

Consistent with most published case reports, the hypoglycemia in our patient was documented within 24 hours of levofloxacin administration, and intravenous dextrose was required. However, the degree of hypoglycemia in our patient was greater, as was the amount of dextrose necessary to correct the hypoglycemia and maintain normal glucose levels. In both case reports involving ciprofloxacin,[1,2] the patients apparently received dosing regimens appropriate for their renal function. However, of the six patients described in the case reports with gatifloxacin,[3,4,5,6] dosing was adjusted for renal dysfunction in only one, although one of the other five had a calculated creatinine clearance just below the cutoff for dosage adjustment. In contrast, our patient received a levofloxacin dosage appropriately adjusted for his renal function.

Other factors that could have contributed to the profound and prolonged hypoglycemia in our patient must be considered. Renal failure itself may predispose both patients with and without diabetes to hypoglycemia.[22,23,24,25] This can result from a number of factors, such as poor caloric intake due to chronic malnutrition, anorexia, or vomiting; drug effects magnified and prolonged due to decreased clearance of the parent compound or metabolites; decreased endogenous insulin clearance; or diminished renal gluconeogenesis.[22,26]

Our patient's appetite was poor throughout his hospital stay; he eventually required tube feedings to meet anticipated caloric requirements. His recorded daily consumption ranged from 5-50% of his meals, and his serum albumin levels decreased from 2.4 g/dl on admission to 1.6 g/dl on day 27.

Decreased elimination of glyburide and its metabolites secondary to chronic renal failure may have contributed to his hypoglycemia. Renal dysfunction has not been found to affect the pharmacokinetics of glyburide significantly, and the estimated creatinine clearance of our patient was above 5 ml/minute, the point at which metabolite elimination may be affected.[25,26,27,28] Furthermore, the patient received glyburide early in his hospital stay without incident, and he had received only four doses (on days 28-31, administered at 7:00 A.M.) immediately before the hypoglycemic event occurred. Thus, accumulation of glyburide or its metabolites would be an unlikely cause of the observed acute decrease in blood glucose concentration in our patient. Renal function was constant throughout his hospital stay, and no abrupt decrease in renal function occurred before the hypoglycemic event. Since his renal function was constant and not different from preadmission levels, it seems unlikely that decreased endogenous insulin clearance was contributory; however, insulin levels were not measured.

Concurrently administered drugs should be considered for their potential to cause hypoglycemia alone or as a result of a drug-drug interaction. All sulfonylurea drugs can produce hypoglycemia. Among these agents, glyburide has been associated with a higher risk of hypoglycemia than the others.[25] High doses of aspirin 4-6 g/day have decreased blood glucose levels in patients with and without diabetes.[29,30] The mechanism of this effect is unknown, and the effect is not normally seen with usual analgesic dosages of aspirin.[30]

Adverse effects of β-adrenergic blockade on glucose metabolism and hypoglycemic symptoms have been reported.[31] However, in contrast to nonselective β-blockers (e.g., propranolol), β1-selective agents have not affected glucose metabolism, nor have they increased the risk of hypoglycemia in patients with diabetes.[32,33,34] Our patient was receiving atenolol, a β1-receptor blocker, a cardioselective agent that has not altered blood glucose levels at therapeutic doses.

Potential drug-drug interactions with glyburide may have occurred, precipitating or exacerbating the hypoglycemic episode in our patient. Glyburide is metabolized in the liver by way of cytochrome P450 (CYP) 2C9 isoenzymes.[35,36] Of concomitantly administered drugs, only fluconazole has been reported to inhibit the CYP2C9 enzyme system.[35] Although fluconazole is primarily renally eliminated, coadministration of this agent and other CYP2C9 substrates (e.g., warfarin, phenytoin, and tolbutamide) has resulted in clinically significant drug interactions; therefore, it is possible that an effect may also be seen with glyburide.[35,37]

Fluconazole 50 mg/day for 14 days did not significantly alter blood glucose control or precipitate hypoglycemia in women who were also taking glibenclamide (glyburide) or gliclazide.[38] However, in a study of healthy male volunteers, fluconazole 100 mg/day for 7 days increased the glyburide area under the curve and maximum concentration, which required oral glucose administration in some subjects.[39] In addition, a fatality was reported in that study from hypoglycemia in association with fluconazole and glyburide administration.[39]

Thus, fluconazole may have contributed to the hypoglycemia in our patient, although he received 4 days of concomitant fluconazole and glyburide therapy before the hypoglycemic event without incident. Since glyburide is highly protein bound, displacement from its binding sites by other highly protein-bound drugs may lead to enhanced hypoglycemic action. High doses of aspirin 4 g/day has increased the free fraction of glyburide in healthy subjects.[40] However, since our patient was receiving a low dose of aspirin 81 mg/day, the likelihood of displacement of glyburide at the resulting low salicylate concentrations was probably minimal.

Levofloxacin, like some of the other fluoroquinolones, is primarily renally eliminated and has not interacted with other drugs metabolized by the CYP enzyme system. However, symptomatic hypoglycemia has been reported, usually in patients with diabetes receiving concomitant treatment with an oral hypoglycemic agent or with insulin.[10] Thus, a pharmacodynamic drug-drug interaction may exist between levofloxacin and oral hypoglycemic agents. For example, our patient received levofloxacin and glyburide at different times during his hospital stay without incident; however, when these agents were administered concurrently, his blood glucose level decreased dramatically after a single dose of levofloxacin.

The fluoroquinolones have been linked to alterations in blood glucose levels, although the exact mechanism is unknown. Fluoroquinolones have inhibited specific potassium channels in myocardial cells, resulting in QTc prolongation. Thus, it has been postulated that glucose alterations may be due to blockage of the adenosine 5'-triphosphate (ATP)-sensitive potassium channels in pancreatic β-cells that help regulate calcium influx and thus augment insulin release. One study in a rat pancreas model demonstrated that older fluoroquinolones (lomefloxacin, enoxacin, sparfloxacin, tosufloxacin) close the ATP-sensitive potassium channel.[41] This leads to depolarization of the β-cell membrane and opening of voltage-dependent calcium channels, allowing calcium movement into the cells with subsequent insulin release. Furthermore, the authors demonstrated that all the quinolones tested augmented insulin release in a concentration-dependent manner.

Other investigators have also demonstrated the effect of lomefloxacin and norfloxacin on ATP-sensitive potassium channels in pancreatic β-cells.[42] Of interest, there was a marked difference in potency to inhibit potassium channel activity, much like that seen in cardiac cells. To our knowledge, no published data have assessed the effects of available fluoroquinolones in pancreatic β-cells.

The exact frequency of hypoglycemia associated with the fluoroquinolones is not known. Evaluation of the respective new drug applications revealed rates of 0.65-2%.[7,8,9] Similar percentages were reported for comparator drugs (e.g., ciprofloxacin and lomefloxacin) in these submissions. However, the criteria used to define hypoglycemia were different in each new drug application, so one must be cognizant of this when evaluating these data. Furthermore, these data were not stratified for evaluation based on presence of diabetes.

A large phase IV postmarketing trial with gatifloxacin (the Tequin Clinical Experience Study) reported an overall frequency of 0.08% (12 of 15,625 patients), although the frequency was higher in patients with (0.55%) than without diabetes (0.04%).[43] Whether the higher frequency in patients with diabetes was due to gatifloxacin administration or the presence of diabetes is not known. The package inserts for levofloxacin[10] and moxifloxacin[11] list decreased glucose levels of 2.2%[10] and 2% and greater,[11] respectively; however, whether these changes were caused by the drug or the underlying condition is not known.

A recent phase IV study in adult patients receiving levofloxacin for treatment of community-acquired pneumonia reported that hypoglycemia as a serious adverse event occurred in 0.1% of patients.[16] However, the relationship to the study drug was not specified. In a study that reported an overall frequency of 0.02% with ciprofloxacin, the frequency was 0.5% in patients with diabetes.[14] Hypoglycemia with clinafloxacin was 4% in phase II and III clinical trials.[15] Of interest, 72% of the patients with hypoglycemia did not have a history of diabetes.

According to the Naranjo probability scale,[44] a possible relationship existed between levofloxacin and hypoglycemia in our patient. However, three parameters evaluated by the scale could not be assessed in our patient and affected the computed score, most notably rechallenge with levofloxacin. Although the patient had other potential risk factors and causes, the temporal relationship between levofloxacin administration and hypoglycemia and the fact that no other drugs were administered between the levofloxacin and the hypoglycemic episode strongly support levofloxacin as the primary cause in this case.

Furthermore, the contribution of other agents as a cause of hypoglycemia appears to be minimal. Decreased blood glucose levels have been a markedly abnormal laboratory value in patients receiving levofloxacin, and cases of symptomatic hypoglycemia have been reported in patients receiving levofloxacin and oral hypoglycemics concurrently.[10] Thus, it is possible that levofloxacin does cause an unexplained insulin-secreting effect, or it may augment that of oral sulfonylureas in patients with diabetes, as reported with other fluoroquinolones.