Effect of ciprofloxacin and levofloxacin on the QT interval: is this a significant "clinical" event?

Background: The widespread use of the fluoroquinolones has raised the question of the cardiac safety of these medications. This widespread use of this class of antibiotics has displayed their safety profile, which is actually more favorable than many other drug classes. The cardiac toxicity issue at the center of this discussion is the prolongation of the QT interval leading to torsade de pointes. Ciprofloxacin and levofloxacin, two of the more commonly used fluoroquinolones, are considered less likely than other fluoroquinolones to prolong the QT interval. The authors set out to evaluate the effect on the QT interval of patients after administration of ciprofloxacin and levofloxacin.

Methods: A prospective evaluation of 38 consecutive patients evaluated by the infectious disease service and receiving either ciprofloxacin or levofloxacin was undertaken. Twelve-lead electrocardiograms were obtained at baseline and at least 48 hours after the first dose of the antibiotic was administered. Both the longest QT interval and the mean QT interval were evaluated. To account for variations in heart rate, the corrected QT interval was calculated by using Bazett's formula (QTc = QT/[square root of (R - R)]). Statistical analysis was undertaken to assess for the presence of a change after the administration of the antibiotic.

Results: Thirty-eight patients (mean age, 65 [+ or -] 19 years), 23 women and 15 men, were studied. There was a small but significant increase in the longest QTc intervals over baseline in patients receiving levofloxacin; there was no significant change in the mean QTc interval. However, one patient who received levofloxacin was, statistically, an outlier and, on retrospective analysis, had demonstrated severe electrolyte disturbances at the time of the study. When this patient was excluded, the increase in the longest QTc interval was not significant. Patients receiving ciprofloxacin did not demonstrate any significant change in the longest QTc interval or mean QTc interval.

Conclusions: Neither levofloxacin nor ciprofloxacin significantly prolonged the mean QTc interval over baseline. When electrolyte deficiencies in one of the patients evaluated were taken into account, this also held true for the longest QTc interval. There is, therefore, evidence that taking ciprofloxacin or levofloxacin, assuming that there are not any concurrent risk factors, will not cause a significant prolongation in the QT interval.

Key Words: QT interval, torsade de pointes, fluoroquinolones, ciprofloxacin, levofloxacin

**********

The fluoroquinolones are one of the most widely used classes of antibiotics across both patient groups and ages. This widespread use has raised the question of the cardiac safety of these medications. Use of this class of antibiotics has displayed their safety profile, which is actually more favorable than many of the other classes. The cardiac toxicity issue at the center of this discussion is the prolongation of the QT interval leading to torsade de pointes.

Prolongation of the repolarization period of the heart represented on the ECG by the QT interval has, in some instances, been associated with noncardiac medications. (1,2) This effect is manifested on the electrocardiogram as a lengthening of the corrected QT interval (QTc interval), indicating an increased delay in electrolyte flow within cardiac tissue. Specifically, pharmacologic agents that prolong the QT interval have been implicated in blocking the rapid component of the delayed rectifier potassium channel ([I.sub.Kr]). (3,4) Torsade de pointes, which is a polymorphic ventricular tachycardia characterized by a continuous "twisting" in QRS axis around an imaginary baseline is preceded by an elongation of the QTc interval. Torsade de pointes is clinically significant as it may progress to ventricular fibrillation and death. (5)

The fluoroquinolone class of antibacterial agents is among the drugs associated with this effect on the QT interval. Grepafloxacin, sparfloxacin, moxifloxacin, and gatifloxacin are among the drugs commonly noted to cause QTc elongation. (5) Grepafloxacin and sparfloxacin, two quinolones most associated with long QTc, are no longer on the market. (5) More recently developed quinolones, those belonging to the second and third generations of these drugs, are considered safer in this respect. Ciprofloxacin, a second-generation fluoroquinolone, and levofloxacin, a third-generation fluoroquinolone, are considered less likely to prolong the QTc. However, the prescribing information for levofloxacin mentions an "[association] with prolongation of the QT interval," "infrequent cases of arrhythmia," and "rare cases of torsade de pointes" (Table 1). (6)

We prospectively examined a cohort of patients receiving levofloxacin and ciprofloxacin to see whether the described effects in the literature are reproduced in clinical practice. We further examined the literature with regard to the fluoroquinolones and specifically ciprofloxacin and levofloxacin to explore the clinical significance of this reported effect of QTc prolongation.

Materials and Methods

Patient Identification

Thirty-eight consecutive patients prospectively identified and treated with either levofloxacin or ciprofloxacin were included in this study. Verbal informed consent was obtained from each patient, and institutional review board approval was obtained for this study.

Antibiotics Used and Conditions Treated

Twenty-seven patients received levofloxacin and 11 received ciprofloxacin. Patient demographic data are listed in Table 2. Each patient was prospectively evaluated by standard 12-lead ECG before and between at least 2 and 48 hours after the first dose of the antibiotic was administered. Patients were treated with standard doses of levofloxacin and ciprofloxacin for the condition being treated according to the Sanford Guide to Antimicrobial Therapy. (7) Table 2 lists demographic data as well as the conditions treated and doses of each antibiotic used for the condition being treated. When necessary, doses were adjusted for renal impairment according to creatinine clearance calculated, using the Cockroft-Gault formula. (8) Patient renal function was assessed at baseline and at the time of repeat assessment of QTc interval and was ascertained to ensure stability over the study period.

Electrocardiographic Analysis

Standard resting 12-lead ECGs were obtained on each of the patients before initiation of the antibiotic. From these ECGs, QT intervals were manually measured from the beginning of the QRS complex to the end of the T wave (when it returned to the T/P baseline), by a single observer blinded to the treatment data (that is, the observer was not aware which ECG was the pre- or post-treatment ECG). Two methods of defining the QT interval for comparison were used. The first method defined the measured QT interval as the longest of the 12 observed leads; the second defined the measured QT interval as the average of all leads that yielded observable results. To take into account variations in heart rate, the corrected QT interval was calculated by using Bazett's formula (QTc = QT/[square root of (R - R)]). (9)

Exclusions

Excluded patients consisted of those with artifacts on the ECG that precluded evaluation of the QT interval, the presence of atrial fibrillation, or any other arrhythmia that prevented accurate QT interval measurements. Patients receiving other medications that have been described as prolonging the QT interval were also excluded from this study.

Analysis

Statistical analysis was performed in two stages for each medication by using SPSS 11.0 statistical analysis software (Chicago, IL). First, the QTc intervals for each patient before and after receiving medication were paired and compared, generating a P value that represented the significance of the change in QTc interval length. Second, 95% confidence intervals were calculated for the patients' QTc interval before and after receiving medication to determine if observed variations fell within the margins of error. A P value of less than 0.05 was considered to be significant.

Results

Among the 27 patients receiving levofloxacin, there was, on average, an increase in the QTc interval after the antibiotic had been administered. However, when considering a 95% confidence interval to account for the relatively small size of this sample, the observed change is within the margins of error. The average change in the longest QTc interval and mean QTc interval for each individual patient was 0.01 seconds and 0.00 seconds, respectively. In each case, the margin of error demonstrates that the QTc interval may have just as easily remained the same or even decreased (see Table 3).

In Table 3, the P values for the increase in the QTc interval for the longest and mean QTc are 0.04 and 0.2, respectively. The large disparity between these values highlights the variability inherent to the 12-lead ECGs. However, the value of 0.04 suggests a significant increase. Closer inspection of the patient data revealed that one of the 27 patients observed was, statistically, an outlier, demonstrating a change in QTc interval length of over 0.1 second. Analysis of the patient's laboratory studies before and after the antibiotic had been administered revealed hypokalemia as well as other electrolyte deficiencies at the time that the second ECG was taken. These concurrent conditions allow for the exclusion of this data from the study, and results in notable changes in the overall data. The P values for an increase in QTc for the longest QTc and mean QTc jump to 0.09 and 0.39, respectively, indicating a far less significant increase in interval length after exclusion of the outlier.

The results observed among the 11 patients taking ciprofloxacin indicate that on average, the patients involved actually demonstrated a decrease in QTc interval. Both the longest QTc and mean QTc readings yielded an average change of -0.01 seconds. However, the small number of patients results in a significant margin of error, 0.02 seconds. Therefore, as with levofloxacin, there is no evidence that the antibiotic will tend to increase the QTc interval.

The likelihood that this observed change is significant is not great--both the longest QTc data and mean QTc data yielded P values of approximately 0.10. No outliers among the 11 patients taking ciprofloxacin were noted. As with levofloxacin, the relatively high P values indicate that there was no significant change in QTc interval length.

Discussion

Prolongation of the QTc interval is a clinically significant physiologic effect, and, as such, it is important to investigate possibilities that commonly used drugs may have such a consequence. (9) Acquired long QT syndrome is most commonly caused by drug administration, typically that of antiarrhythmic drugs such as quinidine. Other antiarrhythmics, specifically class IA or class III agents, may also potentiate this risk. (10) Other conditions, including bradycardia, electrolyte abnormalities such as hypokalemia or hypomagnesemia, very low-energy diets, and central autonomic nervous system disorders may place patients at risk of acquiring prolonged QT. (11)

In addition to the acquired forms of long QT, there are two types of congenital genetic conditions that result in a predisposition to prolonged QT. The most common form of inherited long QT syndrome is the Romano-Ward syndrome, which is transmitted as an autosomal dominant trait. The other less commonly inherited form of long QT syndrome is the Jervel and Lange Nielsen syndrome, which is transmitted as an autosomal recessive trait. Although these patients typically have an identical clinical presentation to those suffering from Romano-Wade syndrome, they also have associated sensorineural deafness and usually have a longer QT interval. (3,12)

Long QT syndrome typically presents itself clinically as an occurrence of syncope or cardiac arrest. The danger of severely prolonged QT lies in its tendency to induce torsade de pointes. This polymorphic ventricular tachycardia often degenerates into ventricular fibrillation, which precipitates syncope and cardiac arrest. Syncope induced by long QT syndrome will sometimes present itself in children as a seizure. In rare cases, cardiac arrest due to long QT may manifest itself as sudden cardiac death during sleep or rest. (12, 13)

These significant risks associated with developing a delay in electrolyte transport within cardiac tissue highlight the importance of taking measures to prevent its onset. Fluoroquinolones, as a class, have been implicated in having an effect on QTc length. (14) This effect is felt to be due to the blocking of the cardiac voltage-gated potassium channels, particularly the rapid component (IKr) of the delayed rectifier potassium current (IK). On the molecular level, IKr is coded for by HERG (human ether-a-go-go related gene). The degree of the effect on IKr, however, is not the same for all of the fluoroquinolones. The radical in position 5 of the fluoroquinolone ring has recently been discovered as the cause of QT prolongation. A methyl group in position 5 as in sparfloxacin prolongs the QT by 14 milliseconds; an amino group in this position, as in grepafloxacin, prolongs the QT by 11 milliseconds; whereas a hydrogen molecule in this position is associated with a QTc prolongation of less than 2 milliseconds for ciprofloxacin, 3 milliseconds for gatifloxacin, and 5 to 6 milliseconds for gemifloxacin, moxifloxacin, and levofloxacin. (15-18)

Two quinolones in particular, sparfloxacin and grepafloxacin, have demonstrated notable effects on the QTc as noted above. A clinically significant increase in QTc interval, considered to be 0.5 seconds or greater, was observed in 1 to 3% of patients receiving sparfloxacin in clinical studies. (2) This drug, however, has since been withdrawn from the market by its manufacturer due to phototoxicity concerns. Grepafloxacin was voluntarily withdrawn from the market by its manufacturer as well in October of 1999 due to reports of its causing at least seven serious cardiac events, including torsade de pointes. These two antimicrobials were, however, members of the earlier generation of quinolones, and newer generation fluoroquinolones promise less risk.

Ciprofloxacin has been used in more than 250 million patients, with a reporting rate of possibly serious cardiac dysrhythmias of only one case per million treatments. (14,19) Levofloxacin has enjoyed similar widespread use as ciprofloxacin. Between March 1997 and March of 2000, approximately 15 million prescriptions for levofloxacin were dispensed in the United States. During that period, less than one case of QT prolongation or torsade de pointes per million prescriptions dispensed was reported during postmarket safety data surveillance. (20) Both of these cited figures for these two antibiotics that the risk of significant QTc prolongation or torsade de pointes is very small. This is echoed by our data in which none of the 27 patients taking levofloxacin or the 11 patients taking ciprofloxacin demonstrated a change in QT length that could be attributed to either antibiotic.

In a recent review, Katritsis et al (18) echo the currently growing widespread feeling that although QT interval prolongation with the fluoroquinolones is a class effect, it is minimal with respect to certain fluoroquinolones (ie, ciprofloxacin and levofloxacin) and therefore presents a very low risk of drug-induced torsade de pointes. With a frequency of torsade de pointes related to fluoroquinolones estimated at about 0.2 to 2.7 per million prescriptions, Katritsis et al (18) recommend that ECG monitoring during the initiation of a quinolone only be undertaken if there are underlying conditions that predispose the patient to torsade de pointes, or in those patients receiving concomitant medications that might prolong the QT interval.

A recent survey of healthcare practitioners by Al-Khatib et al (21) found that of approximately 517 respondents to the survey, only 224 (43%) measured the QT interval correctly. They also found that a majority of healthcare practitioners could not correctly identify factors and medications that can prolong the QT interval. Interestingly, physicians in training (residents, fellows) and academicians were more likely to measure the QT interval correctly. (21) With this realization, it is important to keep in mind the method for measuring the QT interval. The QT interval represents the time from ventricular depolarization to repolarization. On the ECG, this corresponds to the time from the start of the QRS complex to the point where the T wave returns to baseline. A general rule of thumb states that the QT interval should generally be less than half the preceding R-R interval. This rule of thumb, however, only really holds true for heart rates in the 60 to 90/min range. A more precise method to take variations in heart rate into account obtains the corrected QT interval by using Bazett's formula (QTc = QT/[square root of (R - R)]). A normal QTc interval for women is generally thought to be about 0.41 seconds and 0.39 seconds in men. (9,22)

Indiscriminately avoiding the use of noncardiac medications, particularly quinolones, due to the perceived risk of arrhythmia may be more harmful than beneficial. Sufficient evidence as to the safety of current quinolones must be gathered to prevent such an occurrence. Efforts have been made to perform objective studies assessing the risk presented by quinolones for long QT-related disorders such as torsade de pointes through compiling nationally reported incidents. (23) Inherent deficiencies in this approach, however, preclude its use as definitive data. Failure to account for electrolyte abnormalities, concurrent medical conditions, and concomitant medications, for example, has led some to propose that aggressive action should be taken to strengthen postmarketing surveillance and increase accuracy and depth of reported cases. (23) This study represents a step toward providing a clearer picture of the possibility of a "clinically" significant risk. If patients are screened properly, eliminating the possibility that concurrent conditions may confound the data, definitive results will be obtained.

Conclusion and Present Recommendations

Despite past concerns regarding the safety of prescribing quinolones with respect to QTc prolongation, these data demonstrate that two of the most commonly used drugs of this class, levofloxacin and ciprofloxacin, did not have an appreciable effect on the QT interval in this sample of patients. Larger studies may be used to obtain definitive data regarding the severity of the cardiac-related risks presented by levofloxacin and ciprofloxacin. For now, however, both this study and other published reports support the fact that although it is important to monitor and anticipate those patients that will be at greater risk for developing prolonged QT and torsade de pointes, this effect is not a very "clinically" significant one with the use of ciprofloxacin and levofloxacin when an accurate objective assessment of risk is undertaken. Therefore, the unmonitored use of quinolones is most likely safe and the risk of proarrhythmia is low in patients who do not have other factors increasing their risk for QT interval prolongation.

References

1. Yap YG, Camm J. Risk of torsade de pointes with non-cardiac drugs. BMJ 2000;320:1158-1159.

2. Fish DN. Fluoroquinolone adverse effects and drug interactions. Pharmacotherapy 2001;21:253S-272S.

3. Khan IA. Long QT syndrome: diagnosis and management. Am Heart J 2002;143:7-14.

4. Sedgewick ML, Lip G, Rae AP, et al. Chemical cardioversion of atrial fibrillation with intravenous dofetilide. Int J Cardiol 1995;49:159-166.

5. Napolitano C, Priori SG, Schwartz PJ. Torsade de pointes: mechanisms and management. Drugs 1994;47:51-65.

6. Ortho-McNeil Pharmaceutical. Levaquin (levofloxacin) package insert. Raritan, NJ; 2000.

7. Gilbert DN, Moellering RC Jr, Sande MA, Ed. Sanford Guide to Antimicrobial Therapy 2002. Thirty-second Ed. Jeb C. Sanford Publishers, Hyde Park, VT. 2002.

8. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976;16:31-41.

9. Sorawicz B, Knoebel SB. Long QT: good, bad, or indifferent? J Am Coll Cardiol 1984;4:398-413.

10. Fuster V, Alexander RW, O'Rourke RA ed. Hurst's The Heart. 10th edition. 2001: McGraw-Hill, New York.

11. Moss AJ. Prolonged QT-interval syndromes. JAMA 1986;256:2985-2987.

12. Towbin JA, Vatta M. Molecular biology and the prolonged QT syndromes. Am J Med 2001;110:385-398.

13. Chiang C, Roden DM. The long QT syndromes: Genetic basis and clinical implications. J Am Coll Cardiol 2000;36:1-12.

14. Ball P. Quinolone-induced QT interval prolongation: a not-so-unexpected class effect. J Antimicrob Chemother 2000;45:557-559.

15. Frothingham R. Rates of torsade de pointes associated with ciprofloxacin, ofloxacin, levofloxacin, gatifloxacin, and moxifloxacin. Pharmacotherapy 2001;21:1468-1472.

16. Rubenstein E, Camm J. Cardiotoxicity of fluoroquinolones. J Antimicrob Chemother 2002;49:593-596.

17. Iannini PB, Tillotson GS. Evaluating the risk of cardiac toxicity. Pharmacotherapy 2001;21:261-262.

18. Katritsis D, Camm J. Quinolones: cardioprotective or cardiotoxic? PACE 2003;26:2317-2320.

19. FDC Report. FDA/PhRMA Task Force to assess QT risk by clinical markers. The Pink Sheet-Prescription Pharmaceutical and Biotechnology 1999;61:15-16.

20. Kahn JB. Quinolone-induced QT interval prolongation: a not-so-unexpected class effect; Correspondence. J Antimicrob Chemother 2000;46:847-848.

21. Al-Khatib A, Lapointe NM, Kramer JM, et al. A survey of health care practitioners' knowledge of the QT interval. J Gen Int Med 2005;20:392-396.

22. Al-Khatib SM, LaPointe NM, Kramer JM, et al. What clinicians should know about the QT interval. JAMA 2003;289:2120-2127.

23. Owens Jr, RC Ambrose PG. Torsade de pointes associated with fluoroquinolones. Pharmacotherapy 2002;22:663-668.

24. Bayer Corporation. Cipro (ciprofloxacin) package insert. West Haven, CT; 2000.

Amgad N. Makaryus, MD, Kory Byrns, Mary N. Makaryus, Usha Natarajan, MD, Carol Singer, MD, and Bruce Goldner, MD, FACC 

From North Shore-Long Island Jewish Health System, Division of Cardiology, Electrophysiology Section, and the Division of Infectious Diseases, Long Island Jewish Medical Center, New Hyde Park, NY.

Reprint requests to Dr. Bruce Goldner, Division of Cardiology, Long Island Jewish Medical Center, 270-05 76th Avenue, New Hyde Park, NY 11040. E-mail: [e-mail]

Accepted August 26, 2005.

RELATED ARTICLE: Key Points

* The widespread use of the fluoroquinolones has raised the question of the cardiac safety of these medications.

* The cardiac toxicity issue at the center of the discussion is the prolongation of the QT interval leading to torsade de pointes.

* Ciprofloxacin and levofloxacin, two of the more commonly used fluoroquinolones, are considered less likely than other fluorquinolones to prolong the QT interval.

* We found that neither levofloxacin nor ciprofloxacin significantly prolonged the mean QTc interval over baseline. There is therefore, evidence that taking levofloxacin or ciprofloxacin, assuming that there are no concurrent risk factors, will not cause a significant prolongation in the QT interval.

Table 1. Drug information

                            Levofloxacin                   Ciprofloxacin

Trade name                  Levaquin[R]                    Cipro[R]
Nationally reported cases   14 (15)                        9 (15)
  of torsade de pointes
  while taking antibiotic
Mention of QT prolongation  "Some quinolones, including    None (24)
  in prescribing              levofloxacin, have been
  information                 associated with
                              prolongation of the QT
                              interval on the
                              electrocardiogram and
                              infrequent cases of
                              arrhythmia. During post-
                              marketing surveillance,
                              rare cases of torsade de
                              pointes have been reported
                              in patients taking
                              levofloxacin. These reports
                              generally involved patients
                              with concurrent medical
                              conditions or concomitant
                              medications that may have
                              been class Ia or class III
                              antiarrhythmic agents; in
                              addition, use of
                              levofloxacin in the
                              presence of risk factors
                              for torsade de pointes such
                              as hypokalemia, significant
                              bradycardia, and
                              cardiomyopathy should be
                              avoided." (6)

Table 2. Patient demographics and conditions treated

                           Levofloxacin          Ciprofloxacin

Total No. of patients        27                   11
Male                         11                    4
Female                       16                    7
Mean age (yr [+ or -] SD)    65 [+ or -] 20       67 [+ or -] 16
Bronchitis/sinusitis        500 mg PO/IV q 24 h   --
Pneumonia                   500 mg PO/IV q 24 h   --
Prostatitis                 500 mg PO/IV q 24 h   --
Urinary tract infection     250 mg PO/IV q 24 h  250 mg PO q 12 h

Table 3. Change in QTc postadministration of the antibiotic

                                   Levofloxacin
                      Longest QT (sec)      Mean QT (sec)

Average QTc at        0.46 [+ or -] 0.01    0.43 [+ or -] 0.01
  baseline
Average QTc after     0.48 [+ or -] 0.02    0.44 [+ or -] 0.02
  antibiotic
Average change in     0.01 [+ or -] 0.02    0.00 [+ or -] 0.01
  QTc interval
P value for change    0.04 (0.09 excluding  0.20 (0.39 excluding
  in QTc interval       outlier)               outlier)
Instances of torsade                     None
  de pointes

                                   Ciprofloxacin
                      Longest QT (sec)      Mean QT (sec)

Average QTc at         0.48 [+ or -] 0.03    0.45 [+ or -] 0.03
  baseline
Average QTc after      0.47 [+ or -] 0.03    0.44 [+ or -] 0.03
  antibiotic
Average change in     -0.01 [+ or -] 0.02   -0.01 [+ or -] 0.02
  QTc interval
P value for change     0.09                  0.10
  in QTc interval
Instances of torsade                     None
  de pointes