Antibiotics and cardiac risk: A look at recent data, a call for increased awareness
In recent years, reports of increased risk for cardiac events associated with use of some antibiotics have been under the spotlight. Widely used antibiotics such as azithromycin have been linked to risk for mortality and others have been linked to pronounced effects on QT interval prolongation. Growing attention on the link led the FDA to issue public health advisories cautioning physicians and patients about usage of certain antibiotics in patients who are already at risk for CV events.
Experts Cardiology Today interviewed said recent reports should not cause undue alarm, but rather underscore the need for careful consideration of antibiotic regimens in patients with known CV risk. Increased awareness and education on which antibiotics can raise the risk for CV events and in which patients higher risk is most likely to occur are key, they said.
“The prescribing of any antibiotic is a trade-off between the risks and the benefits. The adverse cardiac events are just like any other adverse effects, such as photosensitivity or superinfection with Clostridium difficile,” Wayne A. Ray, PhD, professor of public health policy at Vanderbilt University School of Medicine, said in an interview. “These effects need to be considered in the prescribing decision.”
Joseph A. Hill, MD, PhD, director of the Harry S. Moss Heart Center and chief of cardiology at University of Texas Southwestern Medical Center, Dallas, said the growing attention on the link between antibiotics and cardiac event risk is akin to patient fear about the link between statins and muscle injury.
Image: Marty Perlman
“We see patients all the time who are unable to … compare the very small risk that a statin imposes via muscle issues compared with the many orders of magnitude greater benefit that they derive, that benefit being invisible. While the CV risks that have been associated with some antibiotics is an issue, it would be wrong to alarm people that they cannot take [one of these antibiotics because] this is a dangerous medicine, because the benefits outweigh the risks,” Hill, who also is a professor of medicine and molecular biology, James T. Willerson MD Distinguished Chair in Cardiovascular Diseases and Frank M. Ryburn Jr. Chair in Heart Research at University of Texas Southwestern Medical Center, told Cardiology Today.
Mechanisms of risk
CV risks associated with certain antibiotics usually stem from prolongation of the QT interval. This is most commonly observed with two classes of antibiotics, fluoroquinolones and macrolides. One fluoroquinolone, moxifloxacin, is known to have such a pronounced effect on QT interval prolongation that the FDA uses it as a control when evaluating new drugs to see if it prolongs the QT interval.
“The majority of arrhythmic events that occur with the antibiotics, especially the quinolones, is because they prolong the QT interval,” Peter R. Kowey, MD, FACC, FHRS, professor of medicine and clinical pharmacology at Jefferson Medical College, Philadelphia, and chief of cardiology at Main Line Health, Wynnewood, Pa., said in an interview. “They have a direct effect on a particular potassium current that causes a delay in cardiac repolarization that is reflected on the surfaces of an ECG as a long QT; we know that kind of change in repolarization disposes patients to the development of torsades de pointes, a very characteristic polymorphic ventricular tachycardia that can be lethal.”
Fortunately, Hill said, an antibiotic by itself usually is not enough to cause torsades de pointes.
“To get torsades de pointes, you generally need two or three things to line up,” he said. “Being on an antibiotic like azithromycin is one of them, but it is usually not enough. There have been millions of doses of azithromycin given out, and people aren’t dropping dead. However, there is literature that in people who are already predisposed to a rhythm disturbance — for example, those with structural heart disease, prior MI, HF, hypertrophy from hypertension or other significant risk factors for developing torsades de pointes — exposure to azithromycin pushes the needle a little bit further toward developing that rhythm disturbance. It is a small signal, but it is statistically detectable. This is true for a number of drugs: For example, some of the tricyclic antidepressants do this. For whatever reason, a lot of drugs have some modest activity on one of those potassium channels.”
Recent reports of interest
Wayne A. Ray
The cardiology community and the FDA took special notice of CV risks associated with some antibiotics in July 2012, when Ray and colleagues published data in The New England Journal of Medicine demonstrating a small absolute increase in CV deaths associated with a 5-day regimen of azithromycin in a Tennessee Medicaid cohort. Risk was most pronounced in people in the highest decile of risk for CVD.
In that study, patients taking azithromycin for 5 days had a greater risk for CV mortality (HR=2.88; 95% CI, 1.79-4.63) and all-cause mortality (HR=1.85; 95% CI, 1.25-2.75) compared with no antibiotic use. Relative to amoxicillin, the researchers reported approximately 47 more deaths per million courses of therapy in patients taking azithromycin, a figure that jumped to 245 additional CV deaths per million in patients at the highest decile of risk for CVD. Compared with azithromycin, levofloxacin was associated with similar risk for CV mortality and ciprofloxacin was associated with lower risk for CV mortality.
As a result of that study, the FDA issued a drug safety communication in March 2013 warning that azithromycin can cause abnormal changes in the electrical activity of the heart and can potentially lead to CV death, a warning that was added to azithromycin’s label. “Patients at particular risk for developing this condition include those with known factors such as existing QT interval prolongation, low blood levels of potassium or magnesium, a slower than normal heart rate, or use of certain drugs used to treat … arrhythmias. The potential risk of QT prolongation with azithromycin should be placed in the appropriate context when choosing an antibacterial drug,” according to the FDA statement.
However, a cohort study in young and middle-aged Danish adults conducted by Henrik Svanström, MSc, from the department of epidemiology research at Statens Serum Institut in Copenhagen, Denmark, and colleagues indicated no heightened risk for CV mortality associated with azithromycin in that patient population. The data were published in The New England Journal of Medicine in 2013.
In a discussion of the data, Thomas M. File, Jr., MD, MSc, MACP, FIDSA, FCCP, chair of the infectious disease division at Summa Heath System in Akron, Ohio, said there are differences between the Tennessee study and the Danish study. “The Tennessee study included sicker participants with more comorbidities. So those patients more likely would be at risk for already having prolonged QT interval and more likely might be on other medications that would predispose them to arrhythmias,” he said. File also is a professor of internal medicine, master teacher and chair of the infectious disease section at Northeast Ohio Medical University.
In March, Gowtham A. Rao, MD, PhD, MPH, and colleagues published results of a retrospective cohort study in Annals of Family Medicine that found that azithromycin and levofloxacin raised the risk for mortality and arrhythmias when compared with amoxicillin in a population of US veterans. In the first 5 days of treatment, veterans taking azithromycin had a greater risk for mortality (HR=1.48; 95% CI, 1.05-2.09) and serious arrhythmias (HR=1.77; 95% CI, 1.2-2.62) compared with veterans taking amoxicillin; there was no difference in risk during treatment days 6 through 10. Moreover, compared with amoxicillin, veterans taking levofloxacin had a greater risk for mortality in the first 5 days (HR=2.49; 95% CI, 1.56-3.79) and the second 5 days (HR=1.95; 95% CI, 1.32-2.88). A similar trend was observed in risk for serious arrhythmias during the first 5 days (HR=2.43; 95% CI, 1.56-3.79) and the second 5 days (HR=1.75; 95% CI, 1.09-2.82).
“The key message from our finding is that use of antibiotics is not free of risk,” Rao, adjunct clinical assistant professor at the University of South Carolina School of Medicine, told Cardiology Today. “At the same time, we have to remember that, in the past century, antibiotics have saved millions of lives and been proven time and again that they cure bacterial infectious disease.”
In August, Svanström and colleagues published findings in the British Medical Journal indicating that clarithromycin use was associated with increased risk for cardiac mortality (RR=2.07; 95% CI, 1.28-3.35), and that the effect appeared to be greater in women than in men. They found the adjusted absolute risk difference compared with penicillin V use was 37 cardiac deaths per 1 million treatment courses.
Most recently, Michael Fralick, MD, from the department of internal medicine at the University of Toronto, and colleagues published in the British Medical Journal in March results of a population-based nested case-control study of adults aged 66 years and older taking ACE inhibitors or angiotensin receptor blockers. The data demonstrated that older adults taking co-trimoxazole (OR=1.38; 95% CI, 1.09-1.76) and ciprofloxacin (OR=1.29; 95% CI, 1.03-1.62) had an increased risk for sudden death compared with older adults taking amoxicillin.
With an array of data on cardiac risks with differing antibiotics, awareness and education are encouraged.
“The [potential] for cardiac effects is most important for persons with high baseline risks, such as prior MI or HF. In those patients, it would make sense to avoid the antibiotics with cardiac effects,” Ray said.
The risk appears to be evident only when the antibiotic regimen is active and not afterward, experts said. Thus, patients taking an antibiotic for a prolonged indication are exposed to risk longer than taking a shorter-term regimen.
In the studies of a 5-day azithromycin regimen, “the risk went up through 5 days, and on day number 6, it went right back down,” Hill said. “It was only when you were actively taking the medication. So if you’re taking it for 2 or 3 months, the period of time that that risk is elevated would likely be throughout that exposure period.”
Albert L. Waldo
Besides physician education, patient education is crucial for ensuring a prescription is not written that puts someone at undue risk, said Albert L. Waldo, MD, PhD (Hon), Walter H. Pritchard professor of cardiology, and professor of medicine and biomedical engineering at Case Western Reserve University, Cleveland.
For example, patients on antiarrhythmic drugs that prolong the QT should know of that effect, so they can “tell the doctor when they are getting a new medication, saying ‘I’m on a drug that prolongs the QT,’” Waldo, a member of the Cardiology Today Editorial Board, said in an interview. “Patients can be their best advocate.”
Nonetheless, he said, all doctors should double-check whether a patient is already taking a drug that prolongs the QT or has other risk factors for long QT syndrome.
Peter R. Kowey
After the 2012 NEJM study from Ray and colleagues was published, File said his institution changed practice to make sure alternative agents are considered for patients who might be at elevated cardiac risk. “For example, when patients come to the ED, we are requesting that [ED physicians] assess for known cardiac conditions, particularly medicines that are known to prolong the QT,” he said. “Then we ask them to do an ECG to be safe, and if they already have prolongation, then we ask them to use an alternative agent.” Also, he said, electronic medical records at his institution are configured so that if a physician orders an antibiotic known to prolong the QT interval for a patient on another drug also known to prolong the QT interval, the order will be flagged for a potential drug-drug interaction.
The challenge is getting primary care doctors to integrate these precautions into their practices, Kowey, the Cardiology Today Arrhythmia Disorders Section Editor, said.
“The first thing is to make sure that patients don’t have a history of sudden death in their family or any suggestion of long QT syndrome,” he said. “Just a couple of questions about whether there has been a sudden unexpected death in his or her family can be helpful.
“When you are using a drug that has a known effect on the QT interval, getting a baseline ECG is also helpful. But what is the likelihood that that will actually happen? You have to rely on the reliability of whoever is giving the antibiotic out to read the [ECG] or make sure they are getting an accurate reading from somebody else,” Kowey said. “It really imposes a very large burden on practicing physicians. What the FDA would prefer is to have drugs out there that don’t have this liability … so that we don’t have to depend on this behavior.”
Physicians, patients and technology can all collaborate to ensure more appropriate use of and less risk from antibiotics, according to Rao.
“Doctors should remember, ‘first, do no harm’ and should ask themselves if they will cause harm to a patient by dispensing an antibiotic,” he said. “Science can help, too, by creating decision-support tools that can compute the risk of benefit vs. harm. Finally, and most important, doctors and patients are now becoming more engaged with each other and making shared decisions on their care.”
Any decision “is going to depend on the reason the antibiotic is prescribed,” Ray said. “If the patient has chronic sinusitis, treatment with a drug with known cardiac effects doesn’t make sense. On the other hand, if the patient has pneumonia that is susceptible to the particular antibiotic, then it may make sense to continue the antibiotic. There simply has to be a discussion of benefits vs. risks.”
Waldo agreed. “If the indication for the drug is clear and compelling, you can give the drug safely but have to take precautions,” he said. “Otherwise, there could be trouble.”
An example of such precautions, he said, is a set of exclusion criteria he helped develop for a trial in which azithromycin was studied for prevention of exacerbations of chronic obstructive pulmonary disorder to ensure that participants would not be exposed to excess cardiac risks. Of the 1,577 study participants, none developed torsades de pointes, and a benefit of azithromycin for that indication was observed.
“We insisted that they couldn’t have a resting heart rate greater than 100 bpm; when the heart rate gets much faster than that, the corrected QT interval becomes unreliable and the consequences are not good,” Waldo said. Participants with a corrected QT interval exceeding 450 ms on two occasions separated by at least 1 week also were excluded from this trial. “Then we had to determine the corrected QT interval. We measured the QT interval from the beginning of the QRS complex to the end of the T wave for a minimum of three cardiac cycles, in leads II, V5 and V6. This QT interval was then corrected for heart rate using Fridericia’s formula, because it’s a little better at rapid rates. If there was a left or right bundle branch block, we measured the JT interval,” and those with a JT interval >420 ms, corrected for heart rate, also were excluded. Participants with a history indicating increased risk for torsades de pointes and those taking medications that prolong the QT interval also were excluded, he said.
In the long run, there are two factors that could reduce the chances of antibiotics causing CV harm, according to Kowey.
One factor is smarter use of antibiotics. “There is horrible overuse of antibiotics; people give antibiotics out for all the wrong reasons in clinical practice,” he said, adding that if doctors limit their prescriptions of antibiotics to cases of confirmed bacterial infections, that will mitigate a lot of unnecessary risk. File agreed, noting that many participants in the 2012 NEJM study appear to have received antibiotics for inappropriate reasons. “A good thing about this increased awareness is that it may have reduced some of the misappropriate use of azithromycin because a lot of physicians would think twice if there is a potential downside, causing an arrhythmia,” File said.
The second factor is that “the drug industry … is looking for ways to screen multiple compounds so as to come up with drugs that don’t have a cardiac liability before they get into clinical development,” Kowey said. “There are many ways to alter the chemical structure of drugs to … avoid some of the toxicities. Rapid throughput and preclinical screening of compounds in order to come up with drugs that are safer before you even get to the clinic is very important.” – by Erik Swain
Thomas M. File Jr., MD, MSc, MACP, FIDSA, FCCP, can be reached at 75 Arch St., Suite 506, Akron, OH 44304; email: email@example.com.
Joseph A. Hill, MD, PhD, can be reached at University of Texas Southwestern Medical Center, 6000 Harry Hines Blvd., Dallas, TX 75390; email: firstname.lastname@example.org.
Peter R. Kowey, MD, FACC, FHRS, can be reached at Lankenau Medical Office Building East, Suite 558, 100 E. Lancaster Ave., Wynnewood, PA 19096; email: email@example.com.
Gowtham A. Rao, MD, PhD, MPH, can be reached at: firstname.lastname@example.org.
Wayne A. Ray, PhD, can be reached at Village at Vanderbilt, Suite 2600, 1501 21st Ave. South, Nashville, TN 37212; email: email@example.com.
Albert L. Waldo, MD, PhD (Hon), can be reached at University Hospitals Case Medical Center, 11000 Euclid Ave., Cleveland, OH 44106; email: firstname.lastname@example.org.
Disclosures: File reports receiving research funding from Cempra and Pfizer and serving as a consultant or scientific advisory board member for Cempra, Cubist, Durata, Forest, Merck, Nabriva, Pfizer, Tetraphase and Venatorx. Kowey reports consulting for a number of pharmaceutical companies for cardiac safety. Waldo reports consulting for Sanofi. Hill, Rao and Ray report no relevant financial disclosures.