The use of hydroxychloroquine in coronavirus disease (COVID-19) pandemic raised significant concerns as regards safety and efficacy in hospitalized patients. The objective was to examine the effect of hydroxychloroquine on clinical improvement and mortality among hospitalized patients with COVID-19.
A prospective cohort study was conducted at four general hospitals in the Western region, Saudi Arabia. Patients who had absolute or relative contraindication for using hydroxychloroquine were excluded. Patients concomitantly receiving other medications including azithromycin, antivirals, and supportive treatment were not excluded.
A total 267 patients were included in the current analysis; 185 (69.3%) on hydroxychloroquine and 82 (30.7%) on non-hydroxychloroquine treatments. The average age was 46.0±13.3 years and 78.3% of the patients were males. Approximately 95.9% of the patients were symptomatic with mild (50.6%), moderate (32.6%), severe (8.2%), or ARDS symptoms (4.5%). Compared with no hydroxychloroquine, those on hydroxychloroquine had significantly longer length of stay (11.5±7.1 versus 7.8±4.3 days, p<0.001), more ICU admission (22.7% versus 9.8%, p=0.012), and more intubation (12.4% versus 3.7%, p=0.026). Improvement of symptoms (84.3% versus 81.7%, p=0.595) and hospitalization death (7.0% versus 1.2%, p=0.071) were not significantly different between groups. With exception of length of stay, the association of hydroxychloroquine with the above negative outcomes disappeared after adjustment for several factors including disease severity and concomitant use of azithromycin.
Hydroxychloroquine is not associated with better improvement of symptoms compared with other treatments. Moreover, it is associated with longer length of stay but not mortality or ICU admission in adjusted analysis.
Academic Editor: Jose Luis Turabian, University of Madrid, Toledo, Spain.
Checked for plagiarism: Yes
Review by: Single-blind
Copyright © 2020 Mohammed Alghamdi, et al.
The authors have declared that no competing interests exist.
Since its first appearance in Wuhan (China) in late 2019, more than 30 million patients globally were infected with severe acute respiratory syndrome coronavirus number 2 (SARS-CoV-2) with more than 950 thousands related deaths by mid-September 2020 1, 2. Although the mortality observed in the current coronavirus disease (COVID-19) pandemic (4.1% of closed cases) is much lower than other coronaviruses such as SARS-CoV (10%) and the Middle East Respiratory Syndrome-CoV (MERS-CoV, 35%)3, the rapid and universal spread of COVID-19 caused unprecedented global public health emergency and major healthcare crises 4, 5.
With the lack of recognized therapeutic medications or effective vaccine, the management of COVID-19 was largely dependent on off-label use of available medications 6, 7. Several medications have been tried during the pandemic including anti-viral drugs, antimalarial drugs, and immunomodulatory agents (such as tocilizumab and interferons) 6, 7.
Chloroquine and hydroxychloroquine have been used for decades in the prevention and treatment of malaria and then the treatment of some autoimmune diseases 8. Their earlier use in the COVID pandemic was based on pre-pandemic reports that showed their ability to inhibit viral replication of several viruses including SARS and human immunodeficiency virus (HIV) 8, 9. Additionally, in vitro reports published early in the pandemic showing the ability of hydroxychloroquine to inhibit SARS-CoV-2 replication 10. The use of hydroxychloroquine in the COVID pandemic raised significant concerns as regards safety and efficacy, specially among cardiac patients 11. Accumulating evidence and scientific debates forced several international organizations such as the World Health Organization (WHO) and the US Food and Drug Administration (FDA) to limit or halt the use of hydroxychloroquine in the management of patients with COVID-19 11, 12. The objective of the current study was to examine the effect of hydroxychloroquine on clinical improvement and mortality among patients with COVID-19 admitted earlier in the pandemic to general hospitals in Saudi Arabia.
The current study was conducted at four general hospitals at Western region, Saudi Arabia. The hospitals included 700-bed King Fahad General Hospital, 500-bed Alnoor Specialist Hospital, 300-bed East Jeddah Hospital, and 300-bed King Abdullah Medical Complex. All were located in Jeddah with exception of Alnoor Specialist Hospital which is located at Mecca. The hospitals were allowed to provide healthcare services for patients with COVID-19 in addition to other types of patients. The hospitals were following the guidelines of Saudi Ministry of Health (MOH) as regards testing, diagnosis, admission, isolation, management, and discharge 13, 14.
It was a prospective cohort study conducted between March 1, 2020 and May 30, 2020. The study design obtained all required ethical approvals from the Institutional Review Board (IRB) of the Saudi MOH. Informed consent was obtained from patients or their immediate family members after explaining the objectives of the study.
Sample Size Calculation
Assuming an improvement rate of 50% and assuming that hydroxychloroquine is used in the majority of patients, 240 patients (160 use hydroxychloroquine and 80 do not use hydroxychloroquine) would be required to detect 20% difference (60% versus 40%) in improvement using 80% power and 95% level of significance.
The study targeted adult patients (age >18 years) with polymerase chain reaction (RT-PCR)-confirmed COVID-19 diagnosis admitted to any of the included four hospitals during the study duration. The patients were divided into two cohorts based on the status of hydroxychloroquine treatment, as per the Saudi MOH guidelines. Patients who had absolute or relative contraindication for using hydroxychloroquine were excluded from the study. These included known hypersensitivity to hydroxychloroquine or similar compounds, Glucose-6-Phosphate Dehydrogenase (G6PD) deficiency, decompensated heart failure, prolonged QTc interval, and preexisting retinopathy. Patients concomitantly receiving other medications including azithromycin, antivirals, and supportive treatment were not excluded.
Recruitment and Data Collection
Patients were conveniently recruited after obtaining informed consent. Structured study data collection sheet was used to collect the patient information. These included socio-demographic characteristics, exposure history, symptoms, comorbidity, chest imaging findings, relevant laboratory examinations, and outcomes.
Clinical improvement was based on comparing the assessments of symptoms, disease severity, and chest imaging before and after the use of treatments. Disease severity was categorized into five groups; asymptomatic, mild (symptomatic without evidence of pneumonia or hypoxia), moderate (clinical signs of pneumonia but no hypoxia), severe (severe pneumonia or hypoxia), and acute respiratory distress syndrome (ARDS) 14, 15. Other outcome measures included length of stay, need for intensive care unit (ICU) admission, need for intubation, need for inotropic support, and hospitalization death.
Categorical variables were presented as frequencies and percentage while continuous variables were presented as means and standard deviations (SD) or median and inter-quartile range (IQR), as appropriate. Demographic, clinical, and outcome variables were compared between those who received and those who did not receive hydroxychloroquine. Chi-square or Fisher’s exact test, as appropriate, were used to examine differences in categorical variables while student t-test or Mann–Whitney test, as appropriate, were used to examine differences in continuous variables. To detect independent differences in outcome variables between those who received and those who did not receive hydroxychloroquine, multivariate logistic regression analysis models (for categorical outcomes) and general linear models (for length of stay) were run after adjusting for the variables that were significantly associated with hydroxychloroquine in univariate analysis. All P-values were two-tailed. P-value <0.05 was considered as significant. SPSS software (release 25.0, Armonk, NY: IBM Corp) was used for all statistical analyses.
A total 267 patients have been included in the current analysis; 185 (69.3%) on hydroxychloroquine and 82 (30.7%) on non-hydroxychloroquine treatments. As shown in Table 1, the average age was 46.0±13.3 years and 78.3% of the patients were males. More than half (53.6%) of the patients were from Asian countries while Saudi patients represented less than a quarter (23.2%). One-fifth (19.5%) of the patients were currently employed and only 4.5% were healthcare workers (HCWs). Approximately 16.9% of the patients were smokers and only 3.4% of the females were pregnant. Contact with patients with confirmed COVID-19 (37.1%) was much more common than recent travel within 14 days of symptoms (7.9%). Compared with no hydroxychloroquine, those who received hydroxychloroquine were more likely to be between 35 and 54 years and with unknown employment status.Table 1. Socio-demographic characteristics and exposure history of patients with confirmed COVID-19 by the treatment status
|<35||29 (35.8%)||33 (18.6%)||62 (24.0%)||0.017|
|35-44||15 (18.5%)||41 (23.2%)||56 (21.7%)|
|45-54||15 (18.5%)||53 (29.9%)||68 (26.4%)|
|≥55||22 (27.2%)||50 (28.2%)||72 (27.9%)|
|Male||60 (73.2%)||149 (80.5%)||209 (78.3%)||0.178|
|Female||22 (26.8%)||36 (19.5%)||58 (21.7%)|
|Saudi||16 (19.5%)||46 (24.9%)||62 (23.2%)||0.084|
|Arab||24 (29.3%)||32 (17.3%)||56 (21.0%)|
|Asia||39 (47.6%)||104 (56.2%)||143 (53.6%)|
|Others||3 (3.7%)||3 (1.6%)||6 (2.2%)|
|Currently employed||20 (24.4%)||32 (17.3%)||52 (19.5%)||0.002|
|Not working||21 (25.6%)||31 (16.8%)||52 (19.5%)|
|Retired||0 (0.0%)||2 (1.1%)||2 (0.7%)|
|Umrah||6 (7.3%)||2 (1.1%)||8 (3.0%)|
|Unknown||35 (42.7%)||118 (63.8%)||153 (57.3%)|
|Recent travel within 14 days of symptoms||10 (12.2%)||11 (5.9%)||21 (7.9%)||0.080|
|Contact with patients with confirmed COVID-19||25 (30.5%)||74 (40.0%)||99 (37.1%)||0.138|
|Healthcare workers (HCWs)||4 (4.9%)||8 (4.3%)||12 (4.5%)||>0.99|
|Smoking||9 (11.0%)||36 (19.5%)||45 (16.9%)||0.088|
|Pregnancy||2 (9.1%)||0 (0.0%)||2 (3.4%)||0.140|
(Table 2) shows the clinical characteristics of the patients at admission. Approximately 42.7% of the patients had one or more comorbid disease, specially diabetes (30.7%) and hypertension (23.6%). The majority (95.9%) of the patients were symptomatic; with mainly cough (70.0%), fever (65.2%), and shortness of breath (33.3%). The severity ranged between mild (50.6%), moderate (32.6%), severe (8.2%), or ARDS symptoms (4.5%). Approximately two-thirds (67.7%) of chest imaging were abnormal (mainly opacities and consolidations). The average temperature was 37.4±0.7 C°, oxygen saturation was 94.9%±6.0% at room air, and QTc duration was 414.8±40.6. Median C-reactive protein (CRP) was 6.7 (1.5-11.8) mg/dL and ferritin was 517 (254-1031) ug/L. Compared with no hydroxychloroquine, those who received hydroxychloroquine were more likely to be symptomatic and have cough, shortness of breath, severer form of disease, diabetes, abnormal chest imaging, higher temperature, lower oxygen saturation at room air, lower hemoglobin, higher CRP, and higher ferritin at admission.Table 2. Clinical characteristics at admission among patients with confirmed COVID-19 by the treatment status
|None||54 (65.9%)||99 (53.5%)||153 (57.3%)||0.081|
|One||17 (20.7%)||40 (21.6%)||57 (21.3%)|
|Two or more||11 (13.4%)||46 (24.9%)||57 (21.3%)|
|Type of comorbidity|
|Hypertension||14 (17.1%)||49 (26.5%)||63 (23.6%)||0.095|
|Diabetes||16 (19.5%)||66 (35.7%)||82 (30.7%)||0.008|
|Chronic lung disease||0 (0.0%)||6 (3.2%)||6 (2.2%)||0.182|
|Heart disease||4 (4.9%)||9 (4.9%)||13 (4.9%)||>0.99|
|Asymptomatic||9 (11.0%)||2 (1.1%)||11 (4.1%)||0.001|
|Symptomatic||73 (89.0%)||183 (98.9%)||256 (95.9%)|
|Type of symptoms|
|Fever||51 (62.2%)||123 (66.5%)||174 (65.2%)||0.497|
|Cough||43 (52.4%)||144 (77.8%)||187 (70.0%)||<0.001|
|Shortness of breath||13 (15.9%)||76 (41.1%)||89 (33.3%)||<0.001|
|Chest pain||2 (2.4%)||2 (1.1%)||4 (1.5%)||0.589|
|Nausea or vomiting||8 (9.8%)||11 (5.9%)||19 (7.1%)||0.264|
|Diarrhea||6 (7.3%)||9 (4.9%)||15 (5.6%)||0.404|
|Severity of symptoms|
|Asymptomatic||9 (11.0%)||2 (1.1%)||11 (4.1%)||<0.001|
|Mild||50 (61.0%)||85 (45.9%)||135 (50.6%)|
|Moderate||21 (25.6%)||66 (35.7%)||87 (32.6%)|
|Severe||2 (2.4%)||20 (10.8%)||22 (8.2%)|
|ARDS||0 (0.0%)||12 (6.5%)||12 (4.5%)|
|Chest imaging findings|
|Unremarkable||49 (59.8%)||37 (20.1%)||86 (32.3%)||<0.001|
|Findings detected||33 (40.2%)||147 (79.9%)||180 (67.7%)|
|Other related examinations|
|Systolic BP (mm Hg)||124.6±13.3||125.6±13.2||125.3±13.2||0.546|
|Diastolic BP (mm Hg)||77.0±8.3||76.5±9.6||76.7±9.2||0.728|
|Oxygen saturation at room air||97.0±3.2||94.0±6.7||94.9±6.0||<0.001|
|White blood cell count (x10 ^9/L)||6.7±3.2||7.3±3.4||7.1±3.3||0.163|
|Absolute lymphocyte count (x10 ^9/L)||2.0±1.3||1.7±1.4||1.8±1.4||0.186|
|Platelet count (x10 ^9/L)||248.7±124.3||251.0±112.3||250.3±115.8||0.884|
|C-reactive protein (mg/dL)||1.6 (0.4-9.1)||7.5 (3.0-13.2)||6.7 (1.5-11.8)||0.001|
|Creatinine (mg/dL)||0.9 (0.7-1.1)||0.9 (0.8-1.1)||0.9 (0.8-1.1)||0.053|
|D.dimer (mg/L)||0.6 (0.3-1.3)||0.9 (0.5-1.5)||0.8 (0.5-1.4)||0.076|
|QTc duration (ms)||406.5±27.7||417.7±44.0||414.8±40.6||0.121|
Irrespective of hydroxychloroquine, patients included in the study were receiving azithromycin (73.0%), supportive treatment (48.7%), antivirals (17.6%), and zinc (9.0%). Table 3 shows the clinical characteristics of the patients after receiving hydroxychloroquine or non-hydroxychloroquine treatments. Approximately 58.8% of the patients were still symptomatic. Approximately 10.9% of the patients still had severe disease or ARDS. Approximately 60.0% of chest imaging was still abnormal. The average temperature slightly decreased to 37.1±0.4 C° and oxygen saturation slightly increased to 96.5%±3.3% at room air. Compared with no hydroxychloroquine, those who received hydroxychloroquine were more likely to use azithromycin, still symptomatic with more fever, cough, and shortness of breath, have severer form of disease, abnormal chest imaging, lower levels of oxygen saturation at room air and hemoglobin, and higher levels of white blood cell (WBC) count, platelet count, creatinine, D-dimer, and ferritin after receiving hydroxychloroquine.Table 3. Clinical characteristics after using hydroxychloroquine among patients with confirmed COVID-19 by the treatment status
|Azithromycin||45 (54.9%)||150 (81.1%)||195 (73.0%)||<0.001|
|Lopinavir/ritonavir (Kaletra)||16 (19.5%)||27 (14.6%)||43 (16.1%)||0.313|
|Oseltamivir (Tamiflu)||2 (2.4%)||2 (1.1%)||4 (1.5%)||0.589|
|Zinc||8 (9.8%)||16 (8.6%)||24 (9.0%)||0.770|
|Supportive treatment||45 (54.9%)||85 (45.9%)||130 (48.7%)||0.178|
|Asymptomatic||43 (52.4%)||67 (36.2%)||110 (41.2%)||0.013|
|Symptomatic||39 (47.6%)||118 (63.8%)||157 (58.8%)|
|Type of symptoms|
|Fever||1 (1.2%)||17 (9.2%)||18 (6.7%)||0.017|
|Cough||3 (3.7%)||32 (17.3%)||35 (13.1%)||0.002|
|Shortness of breath||4 (4.9%)||32 (17.3%)||36 (13.5%)||0.006|
|Nausea or vomiting||0 (0.0%)||1 (0.5%)||1 (0.4%)||>0.99|
|Diarrhea||0 (0.0%)||2 (1.1%)||2 (0.7%)||>0.99|
|Severity of symptoms|
|Asymptomatic||43 (52.4%)||67 (36.2%)||110 (41.2%)||0.033|
|Mild||29 (35.4%)||69 (37.3%)||98 (36.7%)|
|Moderate||7 (8.5%)||23 (12.4%)||30 (11.2%)|
|Severe||2 (2.4%)||10 (5.4%)||12 (4.5%)|
|ARDS||1 (1.2%)||16 (8.6%)||17 (6.4%)|
|Chest imaging findings|
|Unremarkable||39 (79.6%)||33 (25.2%)||72 (40.0%)||<0.001|
|Findings detected||10 (20.3%)||98 (74.8%)||108 (60.0%)|
|Other related examinations|
|Systolic BP (mm Hg)||121.9±7.6||121.2±8.8||121.4±8.5||0.595|
|Diastolic BP (mm Hg)||75.1±7.6||74.9±8.8||74.9±8.5||0.863|
|Oxygen saturation at room air (%)||97.5±2.4||96.1±3.5||96.5±3.3||0.010|
|White blood cell count (x10 ^9/L)||6.7±3.4||8.7±5.6||8.1±5.1||0.011|
|Absolute lymphocyte count (x10 ^9/L)||2.4±1.7||2.1±1.8||2.2±1.8||0.478|
|Platelet count (x10 ^9/L)||288.1±150.8||354.4±149.9||335.0±152.8||0.005|
|C-reactive protein (CRP, mg/dL)||2.5 (0.9-11.9)||6.2 (2.4-11.8)||5.3 (1.5-11.6)||0.118|
|Creatinine (mg/dL)||0.8 (0.7-1.0)||0.9 (0.8-1.1)||0.9 (0.7-1.1)||0.018|
|D.dimer (mg/L)||0.8 (0.2-1.2)||1.2 (0.7-4.3)||1.0 (0.6-3.1)||0.018|
As shown in Table 4, patients experienced clinical improvement mainly in symptoms (83.5%), disease severity (47.9%), and to less extent chest imaging (10.6%) after receiving hydroxychloroquine or non-hydroxychloroquine treatments. The average length of stay was 10.4±6.6 days. Approximately 18.7% of the patients required ICU admission, 9.7% required intubation, 2.6% required inotropic support, and 5.2% died during admission. Compared with no hydroxychloroquine, those on hydroxychloroquine had significantly longer length of stay (11.5±7.1 versus 7.8±4.3 days, p<0.001), more ICU admission (22.7% versus 9.8%, p=0.012), and more intubation (12.4% versus 3.7%, p=0.026). The difference in ICU admission and intubation but not length of stay disappeared after stratification by disease severity and to less extent by the presence of comorbidity (data not shown).Table 4. Study outcomes among patients with confirmed COVID-19 by treatment status
|Improved symptoms||67 (81.7%)||156 (84.3%)||223 (83.5%)||0.595|
|Reduced severity||42 (51.2%)||86 (46.5%)||128 (47.9%)||0.475|
|Improved chest imaging||8 (16.3%)||11 (8.5%)||19 (10.6%)||0.128|
|Length of stay (days)||7.8±4.3||11.5±7.1||10.4±6.6||<0.001|
|Length of stay (days)||7 (5-9)||10 (7-14)||9 (6-13)||<0.001|
|ICU admission||8 (9.8%)||42 (22.7%)||50 (18.7%)||0.012|
|Intubation||3 (3.7%)||23 (12.4%)||26 (9.7%)||0.026|
|Inotropic support||0 (0.0%)||7 (3.8%)||7 (2.6%)||0.104|
|Death||1 (1.2%)||13 (7.0%)||14 (5.2%)||0.071|
(Table 5) shows the results of multivariate analysis of study outcomes. With two exceptions, patients who received hydroxychloroquine had generally similar outcomes compared with those who received non-hydroxychloroquine treatments after adjusting for several factors that were significantly different between the two groups in univariate analysis (Table 1 through Table 4. Nevertheless, the length of stay was significantly longer (9.8±2.4 versus 7.6±2.0, p=0.006) and the improvement in chest imaging was significantly lower (odds ratio was 0.02, 95% CI 0.001-0.20, p=0.001) among patients who received hydroxychloroquine compared with those who received non-hydroxychloroquine treatments.Table 5. Adjusted outcomes* (multivariate analysis) among patients with confirmed COVID-19 who received hydroxychloroquine compared with those who received non-hydroxychloroquine treatments
|Odds ratio (OR)||95% confidence interval of OR||P-value||Adjusted R-square|
|Improved chest imaging||0.02||0.001||0.20||0.001||0.499|
|Length of stay (days)||7.6±2.0||9.8±2.4||8.6±2.5||0.006|
The current study examined the efficacy and safety of hydroxychloroquine among admitted patients with COVID-19 of different disease severity. The findings showed that hydroxychloroquine was the second most commonly used single medication (69.3%) after azithromycin (73.0%). It was used approximately four-folds higher than all antiviral medications. The heavy use of hydroxychloroquine is probably reflecting the MOH guidelines at the time of study (March through May) which placed hydroxychloroquine as a first-line drug for all patients (mild to ARDS) while adding antivirals for severe and critical patients 14. The high use of hydroxychloroquine underscores the importance of reviewing local efficacy and safety data.
The patients in the current study experienced a significant clinical improvement specially in symptoms and disease severity after treatment. However, hydroxychloroquine was not associated with better clinical improvement (including symptoms and severity) than non-hydroxychloroquine treatments in both univariate and multivariate analysis. On the other hand, improvement of chest imaging was significantly slower among those receiving hydroxychloroquine. Consistent with current findings, meta-analysis studies could not detect significant clinical improvement (mainly symptoms) in patients receiving hydroxychloroquine [16-18]. However, the results of these studies were conflicting as regards the ability of hydroxychloroquine to limit the radiologic progression [17-19]. Variability may be related to the difference in disease severity and time of re-assessment of chest imaging in different study designs.
Hydroxychloroquine in the current study was associated with longer length of stay in both univariate and multivariate analysis. Similarly, hydroxychloroquine with or without azithromycin was associated with longer length of stay in both univariate and multivariate analysis in a retrospective cohort design 20. However, most of the previous studies either did not focus on the length of stay as an outcome or could not find benefit of hydroxychloroquine on the length of stay 21, 22. It should be mentioned that the length of stay may be easily affected by the hospital capacity and discharge polices 20.
The patients in the current study who were receiving hydroxychloroquine experienced negative outcomes including ICU admission and intubation, which largely disappeared in multivariate analysis adjusted for several factors including disease severity and concomitant use of azithromycin. This finding may indicate that the above negative outcomes were not caused by hydroxychloroquine itself but rather by the difference between treatment groups as regards clinical picture, severity, and other received treatments. Similarly, previous studies could not detect any significant increase in ICU admission or intubation in patients receiving hydroxychloroquine 19, 21, 23.
The current study showed that hydroxychloroquine was associated with an insignificant increase in mortality that completely disappeared in multivariate analysis. Previous studies examining the impact of hydroxychloroquine on all-cause mortality were conflicting and probably changing overtime. For example, earlier meta-analysis reports that included few studies published before the end of April 2020 showed a significant increase in mortality among patients receiving hydroxychloroquine, with risk ratios above two 19, 24. However, the majority of include studies were observation with high level of heterogeneity due to variable doses, disease severity, and concomitant treatments. Later meta-analysis reports that included more recent studies published up to June or July 2020 found no significant increase in mortality among patients receiving hydroxychloroquine 16, 17, 25, 26. A recent local prospective cohort study (pre-review publication) showed that hydroxychloroquine was associated with lower hospital admission and mortality among outpatients with mild-moderate COVID-19 symptoms 27. The apparently conflicting finding may underscore the variability of the outcome by the type of patient and severity of the disease. Interestingly, a two recent report showed a significant increase in mortality only among patients who were receiving both hydroxychloroquine and azithromycin 25, 26. Consistently, a sub-analysis of the current data showed that this group had the worst outcome in univariate analysis, probably because physicians were reserving hydroxychloroquine/azithromycin combination for severe/critical patients (data not shown). Additionally, adjustment for concomitant azithromycin use was a major factor that pushed the adjusted odds ratio for mortality towards null.
The current study is considered the first study in Saudi Arabia to comprehensively examine the efficacy and safety of hydroxychloroquine in admitted patients. The study used a prospective multi-hospital design, included patients with different severity, and reported both univariate and multivariate analysis. Nevertheless, we acknowledge a number of limitations. The lack of randomization may have introduced selection bias. However, the observational design allowed for evaluation of actual treatment practices while adjusting for group differences at admission in multivariate analysis. The concomitant use of other types of treatments represented an important confounding factor for the current outcomes. However, it is probably unethical to deprive patients from other potential treatments at the time of pandemic with limited therapeutic information. Additionally, this has been adjusted for in multivariate analysis.
In conclusion, hydroxychloroquine have been heavily used to treat patients with COVID-19 admitted to general hospitals in Saudi Arabia during the time of the study. Hydroxychloroquine was not associated with better clinical improvement and it was associated with longer length of stay. The observed univariate associations between hydroxychloroquine and negative outcomes such as ICU admission, intubation, and may be mortality can be can be largely explained by the differences between treatment groups in clinical severity and other received treatments. The current findings represent a significant addition to the debate about hydroxychloroquine use in COVID-19 pandemic.
Thanks for all staff at four hospitals who assisted in data collection and other study logistics
The authors report no conflicts of interest in this work
All authors have been acknowledged as contributors of submitted work and fulfill the standard criteria for authorship. All authors have read and approved the submission of the current version of the manuscript. The material included in this manuscript is original and it has been neither published elsewhere nor submitted for publication simultaneously.