This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Cardio, is properly cited. The complete bibliographic information, a link to the original publication on https://cardio.jmir.org, as well as this copyright and license information must be included.
Drug-induced prolongation of the corrected QT interval (QTc) increases the risk for Torsades de Pointes (TdP) and sudden cardiac death. Medication effects on the QTc have been studied in controlled settings but may not be well evaluated in real-world settings where medication effects may be modulated by patient demographics and comorbidities as well as the usage of other concomitant medications.
We demonstrate a new, high-throughput method leveraging electronic health records (EHRs) and the Surescripts pharmacy database to monitor real-world QTc-prolonging medication and potential interacting effects from demographics and comorbidities.
We included all outpatient electrocardiograms (ECGs) from September 2008 to December 2019 at a large academic medical system, which were in sinus rhythm with a heart rate of 40-100 beats per minute, QRS duration of <120 milliseconds, and QTc of 300-700 milliseconds, determined using the Bazett formula. We used prescription information from the Surescripts pharmacy database and EHR medication lists to classify whether a patient was on a medication during an ECG. Negative control ECGs were obtained from patients not currently on the medication but who had been or would be on that medication within 1 year. We calculated the difference in mean QTc between ECGs of patients who are on and those who are off a medication and made comparisons to known medication TdP risks per the CredibleMeds.org database. Using linear regression analysis, we studied the interaction of patient-level demographics or comorbidities on medication-related QTc prolongation.
We analyzed the effects of 272 medications on 310,335 ECGs from 159,397 individuals. Medications associated with the greatest QTc prolongation were dofetilide (mean QTc difference 21.52, 95% CI 10.58-32.70 milliseconds), mexiletine (mean QTc difference 18.56, 95% CI 7.70-29.27 milliseconds), amiodarone (mean QTc difference 14.96, 95% CI 13.52-16.33 milliseconds), rifaximin (mean QTc difference 14.50, 95% CI 12.12-17.13 milliseconds), and sotalol (mean QTc difference 10.73, 95% CI 7.09-14.37 milliseconds). Several top QT prolonging medications such as rifaximin, lactulose, cinacalcet, and lenalidomide were not previously known but have plausible mechanistic explanations. Significant interactions were observed between demographics or comorbidities and QTc prolongation with many medications, such as coronary disease and amiodarone.
We demonstrate a new, high-throughput technique for monitoring real-world effects of QTc-prolonging medications from readily accessible clinical data. Using this approach, we confirmed known medications for QTc prolongation and identified potential new associations and demographic or comorbidity interactions that could supplement findings in curated databases. Our single-center results would benefit from additional verification in future multisite studies that incorporate larger numbers of patients and ECGs along with more precise medication adherence and comorbidity data.
Prolongation of the corrected QT interval (QTc) increases the risk for malignant ventricular tachyarrhythmias such as Torsades de Pointes (TdP), which can degenerate into ventricular fibrillation and cause sudden cardiac death [
Given the high prevalence of drugs with QTc-prolonging risks, immense effort has been dedicated to maintaining databases of such medications [
To address these gaps, we used clinical electrocardiogram (ECG) and electronic health record (EHR) data from a large academic health care system to screen medications for their real-world effects on the QTc. Using this high-throughput approach comparing the same patients before and after initiation of drug therapy, we clarify the associations between medications and QTc prolongation, as well as potential interacting effects from demographics and comorbidities.
We analyzed QTc durations from all ECGs performed from September 30, 2008, to December 31, 2019, at Cedars-Sinai Medical Center, a large academic medical system in Los Angeles, California, which provides quaternary care and completes more than 800,000 outpatient visits a year. We excluded inpatient ECGs because of high uncertainty and variations in medication history while patients are hospitalized. Given the known limitations of QTc quantification at extremes of interval lengths and heart rate, we included only ECGs in sinus rhythm with a heart rate of 40-100 beats per minute (bpm), QRS duration of <120 milliseconds, and QTc duration between 300 and 700 milliseconds [
A patient was considered to be taking a particular medication if the date of the ECG obtained was within a filled medication prescription duration (
We assessed the difference in the mean QTc duration between ECGs when patients were likely to be taking the medication and those when patients were likely to not be taking the medication. We calculated 95% CIs for QTc differences by bootstrapping with 1000 replications. We visualized medications in accordance with their known risk of TdP per the publicly available database at CredibleMeds.org, which was founded in 2000 with the support of the Agency for Healthcare Research and Quality [
We additionally studied the interaction of patient-level demographics and clinical comorbidities with the QTc-prolonging effects of the top QTc-prolonging medications identified by at least 3 of the 5 QTc formulae. Demographics (age, sex, and race and ethnicity) and comorbidities (hypertension, coronary artery disease [CAD], heart failure, diabetes, chronic kidney disease, liver disease, and chronic obstructive pulmonary disease [COPD]) were determined from the EHR. Comorbidities were derived from International Classification of Diseases, Tenth Revision, codes associated with patient visits and problems lists per previously published methods [
Method for determining electrocardiograms (ECGs) of patients who are on and off a particular medication. EHR: electronic health record.
Significant interaction coefficients were displayed using a heat map. All hypothesis testing was 2-sided, and results were evaluated with a significance level of α=.05 after Bonferroni correction for multiple comparisons for each medication.
Data analysis and visualization was performed with R statistical software (version 3.4.1; R Project for Statistical Computing).
This study was approved by the institutional review board of Cedars-Sinai Medical Center (STUDY00001506).
A total of 310,335 ECGs from 159,397 individuals taking a total of 272 medications met our inclusion criteria (
We identified medications associated with the greatest change in mean QTc (
Baseline patient and electrocardiogram (N=310,335) characteristics.
Characteristics | Values | ||
Age (years), mean (SD) | 59.81 (18.67) | ||
|
|||
|
Female | 160,757 (53.5) | |
|
Male | 149,578 (46.5) | |
|
|||
|
American Indian | 540 (0.2) | |
|
Asian | 17,868 (5.8) | |
|
Black | 56,825 (18.3) | |
|
Hispanic | 33,146 (10.7) | |
|
Non-Hispanic White | 174,971 (56.4) | |
|
Pacific Islander | 611 (0.2) | |
|
Other | 26,374 (8.5) | |
|
|||
|
Hypertension | 71,715 (23.1) | |
|
Coronary artery disease | 33,900 (10.9) | |
|
Heart failure | 23,570 (7.6) | |
|
Diabetes mellitus | 32,005 (10.3) | |
|
Chronic kidney disease | 27,427 (8.8) | |
|
Liver disease | 4803 (1.5) | |
|
Chronic obstructive pulmonary disease | 8349 (2.7) | |
QRS duration (milliseconds), mean (SD) | 88.14 (11.16) | ||
QT interval (milliseconds), mean (SD) | 400.54 (36.05) | ||
Heart rate (beats per minute), mean (SD) | 73.20 (12.56) | ||
QTca Bazett (milliseconds), mean (SD) | 437.44 (29.41) | ||
QTc Fridericia (milliseconds), mean (SD) | 424.97 (26.78) | ||
QTc Framingham (milliseconds), mean (SD) | 400.56 (36.03) | ||
QTc Hodges (milliseconds), mean (SD) | 400.56 (36.03) | ||
QTc Rautaharju (milliseconds), mean (SD) | 428.13 (26.92) |
aQTc: corrected QT interval.
Medications with largest QTc-lengthening and -shortening effects. Mean change in QTc duration while on medication versus while off medication. Intervals represent 95% CIs. TdP risk class is in accordance with classifications by CredibleMeds.org. LQTS: long QT syndrome; TdP: Torsades de Pointes; #On: number of electrocardiograms of patients on medication; #Off: number of electrocardiograms of patients off medication; QTc: corrected QT interval.
There was also consistency in medications assessed to have the most QTc prolongation across different QTc assessment methods (
The QTc was lengthened by all major demographics and comorbidities including age (mean QTc difference 2.5, SD 0.03 milliseconds per 10 years), female sex (mean QTc difference 7.20, SD 0.11 milliseconds), non-White race (mean QTc difference 1.08, SD 0.11 milliseconds), hypertension (mean QTc difference 6.93, SD 0.13 milliseconds), CAD (mean QTc difference 4.70, SD 0.17 milliseconds), heart failure (mean QTc difference 17.1, SD 0.20 milliseconds), diabetes (mean QTc difference 8.57, SD 0.17 milliseconds), chronic kidney disease (mean QTc difference 14.7, SD 0.18 milliseconds), liver disease (mean QTc difference 16.4, SD 0.43 milliseconds), and COPD (mean QTc difference 6.89, SD 0.33 milliseconds). The QTc prolongation effects of several individual medications were significantly modified by demographics and comorbidities (
Given that QTc can be associated with heart rate, we performed a limited analysis of medications with the most QTc prolongation showing minimal changes in heart rate for most medications when comparing ECGs of patients who are on versus those who are off medication (
Comparison of medications with largest QT interval (QTc) effects across different QTc formulae.
Interaction of demographics and comorbidities on medication with QTc-prolonging effects. For each medication, we performed linear regressions across each demographic or comorbidity of the form: QTc change = a + b × medication + c × demographic/comorbidity + d × medication × demographic/comorbidity. Significant interaction coefficients are displayed using a heat map. All hypothesis testing was 2-sided, and results were evaluated with a significance level of α=.05 after Bonferroni correction for multiple comparisons. CAD: coronary artery disease; CKD: chronic kidney disease; COPD: chronic obstructive pulmonary disease; HF: heart failure; QTc: corrected QT interval.
Interaction between amiodarone and CAD on QTc duration. CAD: coronary artery disease; QTc: corrected QT interval.
We demonstrate a new, high-throughput technique for identifying and monitoring real-world medication-related QTc prolongation effects using readily accessible clinical data. Using this approach, we confirmed previously known medications associated with QTc prolongation as well as identified potential new medications not identified in curated databases. While considerable effort has been made to maintain well-researched lists of potentially dangerous medications using data from highly controlled studies, little has been published about the real-world effects of medications on QTc duration. It is well known that demographics, comorbidities, and drug-drug interactions may have profound effects on QTc but may not be represented in conventional drug study data [
Our analysis was able to distinguish drugs with known QTc-prolonging effects, including antiarrhythmics (dofetilide, amiodarone, sotalol, and disopyramide), antibiotics (clarithromycin), and antidepressants (fluoxetine, citalopram, and escitalopram). Interestingly, we also identified medications such as rifaximin, lactulose, cinacalcet, lenalidomide, mercaptopurine, and ritonavir, for which there are few data about QTc prolonging effects, but which may deserve additional study. Cinacalcet, for example, has been shown, in very small-scale studies, to be associated with QTc prolongation beyond what might be expected from chronic kidney disease, possibly through its effects on serum calcium levels [
We found that our technique was relatively uninfluenced by the QTc assessment method in identifying medications at the highest risk for QTc prolongation. While prior studies suggested that some methods may perform better at certain heart rate ranges or may better predict even mortality, we found that the effects of QTc-prolonging medications were still observable regardless of the QTc method used [
Demographics and comorbidities all increased the average QTc duration. Consistent with risk factors that are included in the frequently used Tisdale Risk Score for QTc prolongation, we found that age, female sex, heart failure, and CAD were associated with particularly significant increases in QTc [
Several limitations of this study warrant consideration. Although QTc prolongation correlates with TdP risk, prolongation in itself is not fully predictive of TdP. As such, identifying medication that prolong QTc is likely a first step in truly understanding a medication’s effects on TdP risk. While we believe that our method for identifying whether a patient was on or off a medication was quite rigorous, using both Surescripts pharmacy and EHR data, medication adherence remains challenging to assess, and patients could still receive prescriptions from external sources. Both of these cases (nonadherence and external prescription) would bias our results to the null. We pooled ECGs from patients both before as well as after they were on medication to create our off-medication cohorts. This increased the number of ECGs that could be used and did not significantly change whether a medication prolonged the QTc when compared to using only ECGs from patients before starting a medication (
In this study, we demonstrate a high-throughput method using accessible clinical data to identify and monitor the real-world QTc prolongation associations of all commonly prescribed medications. Such a technique might be easily deployable across multiple medical centers for identifying and confirming suspected medications with QTc-prolonging risks and the demographic and comorbidity factors that may enhance or mitigate such risks.
QTc changes (ms) across all medications studied.
Change in heart rate when comparing ECGs on versus off of medication.
Changes in QTc when on a medication compared to when off a medication before or after being on the medication.
beats per minute
coronary artery disease
chronic obstructive pulmonary disease
electrocardiogram
electronic health records
corrected QT interval
Torsades de Pointes
This work was supported by 2 following grants from the National Institutes of health (K23 HL153888 and K99 HL157421). Funding sources were not involved in the study design, data collection, or analysis.
Given the potential identifying nature of the intersection of a patient’s ECG, demographics, comorbidities, and medications, the entire data set will not be made publicly available. However, a limited subset of deidentified data may be made available upon reasonable request to the corresponding author.
None declared.