key: cord-0869834-s928u6g3 authors: Isaksen, J. L.; Holst, A. G.; Pietersen, A.; Nielsen, J. B.; Kanters, J. K. title: Chloroquine, but not hydroxychlorquine, prolongs the QT interval in a primary care population date: 2020-06-20 journal: nan DOI: 10.1101/2020.06.19.20135475 sha: 952a7e8021a1ffa14ed62e8176141fb5321385e5 doc_id: 869834 cord_uid: s928u6g3 Background: Chloroquine (CQ) and Hydroxychloroquine (HCQ) have recently been suggested as treatment for the current Corona Virus Disease 2019 (COVID-19) pandemic. However, despite their long-term use and only few case reports on adverse effects, CQ and HCQ are listed as a known risk of the lethal ventricular arrhythmia Torsade de Pointes and their cardiac safety profile is being questioned. Thus, we aimed to investigate the electrocardiographic and mortality effects of CQ and HCQ in a primary care population. Methods: We used Danish health care registers and electrocardiograms (ECGs) from primary care to define three studies. 1) A paired study of subjects with ECGs before and during use of CQ/HCQ, 2) a matched ECG study of subjects taking CQ/HCQ compared to controls, and 3) a mortality study on people taking HCQ matched to control. In both matched studies, we adjusted for connective tissue diseases, use of QT-prolonging drugs, and cardiac disease. We used the QTc interval as the marker for electrocardiographic safety. In the mortality study, cases were followed from first claimed prescription until 300 days after estimated completion of the last prescription. 95% confidence intervals follow estimates in parenthesis. Results: Use of CQ was associated with a 5.5 (0.7;10) ms increase in QTc in the paired study (n=10). In the matched study (n=28, controls=280), QTc was insignificantly increased in subjects taking CQ by 4.7 (-3.4;13) ms. With a {Delta}QTc of 1.0 (-5.6;7.5), use of HCQ was not associated with an increased QTc in the paired study (n=32). In the matched study (n=172, controls=1,720), QTc also was not different between groups (p=0.5). In the mortality study (n=3,368), use of HCQ was associated with a hazard ratio of 0.67 (0.43;1.05). Conclusions: In subjects free of COVID-19, we found a small increase in QTc associated with use of chloroquine, but not hydroxychloroquine. We found no increased mortality associated with use of hydroxychloroquine. The Corona Virus Disease 2019 pandemic has led to investigation and use of chloroquine (CQ) and hydroxychloroquine (HCQ) in the treatment of COVID-19. CQ and HCQ are 4-aminoquinilines approved for medical use in 1949 and 1955, respectively, and both drugs interfere with important physiological mechanisms that may improve the clinical course of patients with COVID-19. CQ and HCQ are weak bases and accumulate in lysosomes increasing pH which interferes with virus entry and fusion(1). CQ and HCQ also inhibit viral gene expression and post translational modification necessary for viral replication(1). Furthermore, CQ and HCQ have direct anti-inflammatory actions by inhibition of B-and T-cell receptors and especially by decreasing TNF-α and cytokine production (interleukin-1 and interleukin-6)(2). The drawback is that CQ and HCQ have known cardiotoxic effects(3) including vasodilatation, hypotension, hypokalemia, negative inotropy, and arrhythmias. The risk of arrhythmias may partly be mediated through the other cardiotoxic effect and partly due to ion channel blockade. Both CQ and HCQ are known hERG-blockers, but also block a variety of other ion channels including sodium, calcium, I K1 potassium and pacemaker funny channels(4). Since CQ and HCQ are old drugs, the approvals were granted long before thorough QT studies were required and only limited ECG safety data exists. Despite the known hERG block, there is a paucity of documented Torsades de Pointes ventricular tachycardia (TdP) during treatment,(5) which may be attributed to the concomitant calcium block protecting against TdP.(6) Only three cases of TdP have been published(7-9), all in patients with concomitant morbidities as cirrhosis, heart failure, or preexisting QT-prolongation. Using the 12SL algorithm version 23 (GE Healthcare, Wauwatosa, WI), we obtained heart rate, QT interval, QRS duration, JT interval, PR interval, and presence of Right Bundle Branch Block (RBBB) or Left Bundle Branch Block (LBBB). The QT interval was corrected with Bazett (QTcB) and Framingham (QTcFrh) formulae. Based on the JT interval, JTc was obtained by linear correction for heart rate (JTc = JT -119.4108*(RR-1), whereby RR is the R-to-R interval in seconds and JT is measured in ms). From the primary study population, we defined three subsets, each with a CQ and a HCQ population: 1) Paired study, 2) Matched study, and 3) Mortality study. As an additional analysis, we conducted a fourth study on connective tissue diseases with three populations independent of CQ/HCQ prescriptions. The Paired study is comprised of people with an ECG taken during a period of CQ or HCQ treatment, respectively, and with a prior baseline ECG without HCQ/CQ treatment. Thus, in this study, the subjects served as their own controls. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020 . . https://doi.org/10.1101 /2020 Classification of Diseases, 10 th revision (ICD-10). Diabetes mellitus, hypertension, and congestive heart failure were all defined using a combination of ICD-10 and ATC codes at baseline as done previously. (11) For subjects in the paired population, subjects served as their own controls and the results were thus inherently adjusted for comorbidity, age, and sex. We adjusted for a change in use of QT-prolonging drugs where appropriate. We used a random-effects regression model for continuous variables. For subjects in the matched population, we compared continuous electrocardiographic variables using linear regression crudely as well as with adjustment for RA, SLE, SS, use of QT-prolonging drugs, hypertension, ischemic heart disease congestive heart failure, and diabetes mellitus, as appropriate. Categorical variables were tested using a two-sample test of proportions. For subjects in the mortality population, we used Cox regression with full adjustment as with the matched population, as appropriate. Subjects were censored upon emigration . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 20, 2020 . . https://doi.org/10.1101 /2020 individually or 300 days after the last claimed prescription for each case/control pair. We used time-on-study as underlying time variable in the Cox model. The Danish Data Protection Agency approved the use of de-identified data (2007-58-0015) on the conditions that the exact number of subjects in groups of < 3 subjects not be disclosed and that no calculations (including p-values) be reported on groups of < 5 subjects. All estimates were reported along with a 95 % confidence interval indicated as (lower to upper). Data management and statistical computations were performed in Stata (version 16), and a p-value < 0.05 was considered significant. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) We identified ten subjects who had an ECG taken during a treatment of CQ and who had a prior ECG without treatment of CQ ( Compared to their baseline ECG, subjects taking CQ had a significantly longer QTcFrh by 5.5 ms (0.7 to 10, p=0.03). We found no evidence of an increased QTcFrh after use of HCQ (1.0 ms, -5.6 to 7.5 ms, p=0.8). We identified 28 and 172 subjects with an ECG taken during a period of treatment with CQ or HCQ, respectively, and matched them 10:1 with controls. Subjects on CQ (Table 2) , compared to controls, were more likely to suffer from ischemic heart disease, but not hypertension. With full adjustment for RA, SLE, SS, QT-prolonging drugs, hypertension, diabetes, ischemic heart disease, and congestive heart failure, QTcFrh was 4.7 ms longer in the CQ group compared to control, but with a wide confidence interval (-3.4 to 13 ms, p=0.3). Subjects on HCQ (Table 3 ) more often than controls had RA, SLE or SS, as well as hypertension, ischemic heart disease, congestive heart failure, and diabetes. In the unadjusted analysis, we found an increased QTcFrh interval of 5 ms (p=0.002). However, . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 20, 2020. . after adjustment for RA, SLE, SS and QT-prolonging drugs, there was no significant difference in QTcFrh between cases and controls (3.1 ms, -1.1 to 7.3 ms, p=0.15) and after additional adjustment for ischemic heart disease, hypertension, diabetes, and congestive heart failure, the difference in QTcFrh was reduced further to 1.3 ms (-2.9 to 5.6, p=0.5). QRS duration followed a similar pattern in the adjusted and unadjusted analyses, whereas the JT interval also in the unadjusted analysis was comparable between the groups. We identified 3,368 subjects with a prescription for HCQ irrespective of an ECG (Table 4) . Subjects on HCQ were more likely to have RA, SLE, SS, hypertension, ischemic heart disease, and congestive heart failure. In the most crude analysis, adjusted only for sex and age, mortality was not different between subjects on HCQ and controls (HR=1.20, 0.85 to 1.71, p=0.3). With adjustment for RA, SLE, SS, and use of other QT-prolonging drugs, we still found no significant difference between Hydroxychloroquine and matched controls (HR=0.80, 0.51 to 1.26, p=0.3). Further adjustment for ischemic heart disease, diabetes, hypertension, and congestive heart failure did not change the association materially. Kaplan-Meier survival curves are shown in Figure 1 . . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020. . In the RA/SLE/SS study we defined three populations of cases and controls with and without RA, SLE, and SS respectively. We found no difference in crude or adjusted QTcFrh compared to control in any of the populations (Supplementary Tables 1-3) . Heart rate was significantly higher in subjects with RA and SLE, but lower in subjects with SS. Crude PR interval was increased in subjects with SS, but not with adjustment for diabetes, ischemic heart disease, hypertension, and congestive heart failure. In the fully adjusted adjusted for sex, age, SS, SLE, QT-prolonging drugs, ischemic heart disease, diabetes, hypertension, and congestive heart failure, patients with RA had a shorter QRS duration compared to controls. In this real-world study, we found a small, but significant, heart rate-corrected QT-prolongation of 5 ms with CQ, but no QT prolongation with HCQ. We did not find increased mortality in subjects taking CQ or HCQ. CQ and HCQ are both listed as known causes of TdP on CredibleMeds.org,(12) and CQ is recognized by the World Health Organization (WHO) as prolonging the QT/QTc interval. (5) However, WHO states that CQ is associated with a low risk of cardiotoxicity based on PK/PD modelling.(5) In the present study, we found an increase in QTcFrh of 5.5 ms in the . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020. . https://doi.org/10. 1101 /2020 paired CQ population, and no difference the paired HCQ population. Another study (10) showed that a 1,500 mg dose of CQ yielded an increase in QTc of 15-30 ms. Matched on sex and age, subjects receiving HCQ showed a 5 ms increase in QTcFrh. However, the increased QTc disappeared upon adjustment for use of QT-prolonging medication and connective tissue diseases. Hence, the increased QTcFrh was likely caused by rheumatic disease and not HCQ. Thus, in both the matched and the paired analyses, we were unable to demonstrate any statistically significant increase in QTc associated with use of HCQ. We did not find that use of HCQ was associated with an increased mortality. This is not surprising, since the toxicology profile of HCQ is better than that of CQ.(13) WHO reports that the few deaths associated with use of HCQ were caused by overdosing and chronic indications for HCQ. (5) Only three case studies have demonstrated arrhythmic adverse events in association with use of CQ/HCQ. One study(7) involved a syncope in a patient with SLE and end stage renal disease and QT prolongation before HCQ initiation. Another study(8) documents TdP in a patient with SLE, with a history of cirrhosis, HBV-related hepatoma, prior myocardial infarction, and ventricular septal defect. The third study(9) involved a patient with SLE, who also suffered from Congestive Heart Failure, Chronic Kidney Disease stage 5, and Hypertension. All three case studies thus features patients with severe QT-prolonging comorbidity and chronic rather than episodic use of Hydroxychloroquine. Collectively, . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020. . https://doi.org/10.1101/2020.06.19.20135475 doi: medRxiv preprint evidence from the present and other studies found little risk of death or arrhythmic events associated with episodic use of HCQ. Malaria was and still remains an indication for HCQ in Denmark, although it is no longer recommended. (14) HCQ is commonly prescribed for connective tissue diseases such as Rheumatoid Arthritis, Systemic Lupus Erythematous, and Sjögren's Syndrome. We were only able to identify the reason for the HCQ treatment in forty to sixty percent of the cases, likely because milder connective tissue cases were treated in primary care and therefore never obtained a hospital ICD-10 diagnosis detectable in our registries. Furthermore, some subjects may have been prescribed HCQ as malaria prophylaxis. CQ/HCQ has emerged as a possible treatment for COVID-19.(15) Naturally, given the known risk of increased QTc, concerns about the safety of the drugs have been raised.(16, 17) We have found different electrocardiographic safety profiles for CQ and HCQ. With CQ, we have demonstrated an increased QTcFrh of 5.5 ms. With HCQ, we found no increase in QTcFrh. For use of CQ and HCQ, respectively, we found no increased mortality. CQ and HCQ have been used as malaria prophylaxis for decades without raising red flags and the use of HCQ in higher doses to treat auto-immune diseases also has not been . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020. . https://doi.org/10.1101/2020.06.19.20135475 doi: medRxiv preprint associated with an increased mortality. The few case studies that exist were associated with chronic use of HCQ beyond the duration corresponding to a COVID-19 treatment and with severe QT-prolonging comorbidity. Arrhythmias are commonly seen in patients with COVID-19 (one study (18) . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020. . https://doi.org/10. 1101 /2020 Using matched and paired studies of subjects free of COVID-19 receiving chloroquine or hydroxychloroquine, we found that use of chloroquine but not hydroxychloroquine was associated with a small increase in QTc. We were unable to show increased mortality associated with hydroxychloroquine. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020. . https://doi.org/10.1101/2020.06.19.20135475 doi: medRxiv preprint H y d r o x y c h l o r o q u i n e i n t h e T r e a t m e n t o f C O V I D -1 9 I n f e c t i o n -A S y s t e m a t i c L i t e r a t u r e R e v i e w [ P r e p r i n t s e r v e r ] . . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020. . a l e h M , G a b r i e l s J , C h a n g D , K i m B S , M a n s o o r A , M a h m o o d E , e t a l . T h e E f f e c t o f C h l o r o q u i n e , H y d r o x y c h l o r o q u i n e a n d A z i t h r o m y c i n o n t h e C o r r e c t e d Q T I n t e r v a l i n P a t i e n t s w i t h S A R S -C o V -2 I n f e c t i o n . C i r c A r r h y t h m E l e c t r o p h y s i o l . 2 0 2 0 . . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020. . . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020 . . https://doi.org/10.1101 /2020 is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020. . https://doi.org/10.1101/2020.06.19.20135475 doi: medRxiv preprint DP: Not disclosed due to data protection restrictions (groups of <5). NA: Not applicable. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020. . . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020. , Diabetes, Ischemic Heart Disease, Hypertension, and Congestive Heart Failure. + Excluding Chloroquine and Hydroxychloroquine DP: Not disclosed due to data protection restrictions (groups of <5). . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020. . . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020 . . https://doi.org/10.1101 /2020 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020 . . https://doi.org/10.1101 /2020 Supplemental material for Chloroquine, but not hydroxychlorquine prolongs the QT interval in a primary care population. Jonas L. Isaksen, Anders G. Holst, Adrian Pietersen, Jonas B. Nielsen, Claus Graff, Jørgen K. Kanters Use of Chloroquine was defined when a prescription for Chloroquine (ATC code: P01BA01) was claimed. Use of Hydroxycloroquine was defined in the same way using ATC code P01BA02. Systemic Lupus Erythematous was defined as any of the ICD-10 codes DM32, DG058A, DG737C, DI328B, DI398C, DJ991C, or DN085A, or either ICD-8 code 695.4 or 734.1. Rheumatoid Arthritis was defined as any of the ICD-10 codes DM05 or DM06, or any of the ICD-8 codes 712.0, 712.1, or 712.3. Ischemic heart disease was defined as any of the ICD-10 codes I20, I21, I23, I24, I25 or the ICD-8 code 410. Diabetes mellitus was defined as any of the ICD-10 codes E10, E11, E12, E13 or E14, or use of insulin (ATC: A10A) or oral antidiabetics (ATC: A10B), or the ICD-8 code 761.1. Hypertension was present if any of the ICD-10 codes I10 or I15, or the ICD-8 code 401 was found, or if we found concurrent use of two of these classes of medication: alpha blockers (ATC: C02A, C02B, C02C), non-loop diuretics (ATC: C02L, C03A, C03B, C03D, C03E, C03X, C07X, C07C, C07D, C08G, C02DA, C09BA, C09DA, C09XA52), vasodilators (ATC: C02DB, C02DD, C02DG, C04, C05), beta blockers (ATC: C07), calcium blockers (ATC: C08, C09BB, C09DB), or angiotensin converting enzyme inhibitors (ATC: C09). Heart failure was present if the ICD-10 codes I42, I50, or J81, or the ICD-8 code 427.0 was found, or if a prescription of loop diuretics (ATC: C03C) was found. The following drugs and ATC codes were considered QT-prolonging in the present study. The list was taken from CredibleMeds.org on April 30, 2020 (revision March 19, 2020 , and is comprised of those drugs categorized as having a known risk of Torsade de Pointes (TdP), with the omissions of Chloroquine and Hydroxychloroquine. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 20, 2020. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 20, 2020. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 20, 2020. . https://doi.org/10.1101/2020.06.19.20135475 doi: medRxiv preprint RBBB 4 (0.7 %) 14 (0.9 %) Too few cases Too few cases LBBB <3 9 (0.6 %) Too few cases Too few cases # Except Chloroquine and Hydroxychloroquine. *Adjusted for sex, age, Rheumatoid Arthritis, Sjögren's Syndrome, and QT-prolonging drugs # . § Adjusted for sex, age, Rheumatoid Arthritis, Sjögren's Syndrome, QT-prolonging drugs # , Diabetes, Ischemic Heart Disease, Hypertension, and Congestive Heart Failure. DP: Not disclosed due to data protection restrictions (groups of <5). . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 20, 2020. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 20, 2020. . https://doi.org/10. 1101 /2020 436 (38 %) 1,816 (29 %) Ischemic Heart Disease