Demographics of the COVID-19 patients
81 consecutively hospitalized patients had been treated with HCQ and AZI combination from March 23rd to May 10th 2020 and were enrolled into the study (Table 1). The median age was 59 years (35 – 87), 58.0% (n=47) were female. The largest patient group according to age was the 60-69 years old group (24.7%). The median baseline Cumulative Illness Rating scale (CIRS) score (6) was 4 (0 – 15). The majority of the study population (82.7%) had comorbidities and half of the patients (50.6%) had cardiological diseases. 33 patients (40.8%) were taking 1-2 and 10 patients (12.3%) 3-4 concomitant drugs.
Clinical data and laboratory findings of the COVID-19 patients
The median time from symptom onset to hospitalization and treatment with HCQ-AZI were both 7 days (-1 – 42) (Table 2). 80 patients (98.8%) had radiological signs of pneumonia. The median baseline National Early Warning score (NEWS) was 2 (0 – 13). On admission, 34 patients (42.0%) required low-flow oxygen, 2 patients (2.5%) had to be on invasive ventilation and 1 patient (1.3%) was connected to an extracorporeal membrane oxygenation (ECMO) to sustain oxygen saturation above 92%. 3 patients (3.7%) were admitted directly to ICU. Two-thirds of the patients (67.9%) had electrolyte imbalance during the follow-up period.
Cardiotoxicity of HCQ-AZI treatment
More than half of the patients (51.9%) were prescribed at least one additional QT interval prolonging drug during the hospitalization, the majority of these drugs (87.7%) being in the “conditional risk of TdP” group according to “CredibleMeds®” (Table 2).
The median baseline QTcF was 416 ms (365 – 498). The median QTcF was rising daily and the peak of 436 ms (333 – 483) was observed on the 10th day of the HCQ-AZI treatment (Figure 1a). The highest median ΔQTcF was observed on the 8th day (Figure 1b).
Seven patients (8.6%) had QTcF prolongation of ≥500 ms during the 14-day period from the initiation of the treatment (Table 3). Four of these cases were observed during and three immediately after the administration of HCQ-AZI. HCQ-AZI was discontinued in 4 patients (4.9%): one and three in 480-499 ms and ≥500 ms groups, respectively. None of the patients developed ventricular tachycardia. The risk factors significantly associated with QTcF≥500 ms were hypokalemia (p = 0.032) and the use of diuretics during the treatment (p = 0.020), the odds ratios (95% CI) were 6.188 (1.168-32.774) and 7.778 (1.388-43.595), respectively. Multivariate logistic regression analysis was not performed due to strong dependence between hypokalemia and the use of diuretics (phi = 0.4).
14 patients (17.3%) experienced QTcF≥480 ms (Table 3) and 16 patients (19.8%) had a change of QTcF≥60 ms. Higher baseline NEWS score, presence of cardiological comorbidities, higher number of concomitant medications, hypokalemia, use of diuretics during the treatment and higher baseline QTcF were associated with QTcF prolongation ≥480 ms in the univariate logistic regression model (Table 4). On multivariate analysis, cardiological comorbidities (p = 0.034) and hypokalemia (p = 0.008) were found to be independent factors for QTcF≥480 ms interval prolongation (Table 4, Figure 2).
Outcomes of the COVID-19 patients
11 patients (13.6%) were transferred to ICU and 3 patients (3.7%) were connected to ECMO. Cytokine adsorbtion using CytoSorb® filters was applied in 7 cases (8.6%) and interleukin-6-receptor inhibitor Tocilizumab was administered in 4 patients (4.9%). 5 patients (6.2%) were still hospitalized and 73 patients (90.1%) were discharged from the hospital by the end of the follow-up on May 10th, 2020. 3 patients (3.7%) died, all cases were related to progression of multiple organ dysfunction syndrome and none were related to ventricular arrhythmias.