All-cause mortality among patients treated with repurposed antivirals and antibiotics for COVID-19 in Mexico City: A Real-World Observational Study

(250 words) Aim: To evaluate all-cause mortality risk in patients with laboratory-confirmed COVID- 19 in Mexico City treated with repurposed antivirals and antibiotics.

This real-world retrospective cohort study contemplated 395,343 patients evaluated for suspected COVID-19 between February 24 and September 14, 2020 in 688 primary-to-tertiary medical units in Mexico City. Patients were included with a positive RT-PCR for SARS-CoV-2; those receiving unspecified antivirals, excluded; and antivirals prescribed in <30 patients, eliminated. Survival and mortality risks were determined for patients receiving antivirals, antibiotics, both, or none.

136,855 patients were analyzed; mean age 44.2 (SD:16.8) years; 51.3% were men. 16.6% received antivirals (3%), antibiotics (10%), or both (3.6%). Antivirals studied were Oseltamivir (n=8414), Amantadine (n=319), Lopinavir-Ritonavir (n=100), Rimantadine (n=61), Zanamivir (n=39), and Acyclovir (n=36). Survival with antivirals (73.7%, p<0.0001) and antibiotics (85.8%, p<0.0001) was lower than no antiviral/antibiotic (93.6%). After multivariable adjustment, increased risk of death occurred with antivirals (HR=1.72, 95%CI:1.61-1.84) in ambulatory (HR=4.7, 95%CI:3.94-5.62) and non-critical (HR=2.03, 95%CI:1.86-2.21) patients. Oseltamivir increased mortality risk in the general population (HR=1.72, 95%CI:1.61-1.84), ambulatory (HR=4.79, 95%CI:4.01-5.75), non-critical (HR=2.05, 95%CI:1.88-2.23), and pregnancy (HR=8.35, 95%CI:1.77-39.30); as well as hospitalized (HR=1.13, 95%CI:1.01-1.26) and critical patients (HR:1.22, 95%CI:1.05-1.43) after propensity score-matching. Antibiotics were a risk factor in general population (HR=1.13, 95%CI:1.08-1.19) and pediatrics (HR=4.22, 95%CI:2.01-8.86), but a protective factor in . CC-BY-NC-ND 4.0 International licenseIt 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 (which was not certified by peer review) The copyright holder for this preprint this version posted January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 7 hospitalized (HR=0.81, 95%CI:0.77-0.86) and critical patients (HR=0.67, 95%CI:0.63- 0.72).

No significant benefit for repurposed antivirals was observed; oseltamivir was associated with increased mortality. Antibiotics increased mortality risk in the general population but may increase survival in hospitalized and critical patients. . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 8

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the etiologic agent of the coronavirus disease (COVID-19) pandemic, one of the most devastating infectious diseases of this century. Non-pharmacological interventions are the most effective means of limiting the impact of COVID-19. 1 However, several countries have not been able to contain the disease.2 One of the main strategies for finding ways to combat COVID-19 is drug repurposing since developing novel antivirals against SARS-CoV-2 may be protracted. 3 Repurposing existing antivirals is attractive due to their relative safeness and potential anti-SARS- CoV-2 mechanisms. 4 Neuraminidase inhibitors (i.e. oseltamivir, zanamivir) and HIV protease inhibitors (i.e. lopinavir-ritonavir) have been hypothesized to inhibit SARS- CoV-2 proteases involved in the degradation of polyproteins that control viral replication. 5 Adamantanes (i.e. amantadine, rimantadine) are thought to disrupt lysosomal trafficking, thereby impeding the release of SARS-CoV-2 ribonucleic acid (RNA) into the cell, 6 and by inhibiting conductance of the envelope (E) protein. 7 Acyclovir, a nucleotide analog antiviral, was found as a candidate drug for COVID-19 by potentially counteracting gene expression changes observed after SARS-CoV-2 infection.8 Evaluating repurposed drugs during a pandemic comprises numerous challenges. 9 Up to December 8, 2020 there were 413 registered studies to test antivirals for COVID-19, of which 400 were still active. 10 Most trials in the World Health Organization (WHO) platform were for lopinavir/ritonavir (156), favipiravir (62), remdesivir (52), oseltamivir . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 9 (19), and ribavirin (16). 11 Other common antivirals are not being tested for COVID-19 but could be having widespread use in the community and hospitals since practice guidelines do not discourage/recommend most antivirals due to a lack of evidence, 12,13 others advise against most, 14,15 or recommend oseltamivir empirically during the influenza season16 and when coinfection exists.17 The WHO recommends early antibiotic therapy in patients with severe COVID-19, but not for mild-to-moderate disease. 14 However, questions have arisen on the effectiveness of antibiotics due to its viral origin, whilst indiscriminate use of antibiotics could aggravate antimicrobial resistance.18 In this population-based study, we hypothesized that repurposed antivirals and antibiotics could be having widespread use in real-world settings. To determine their impact on mortality, we studied survival and all-cause mortality risk in patients with laboratory-confirmed COVID-19 in Mexico City receiving these drugs in both ambulatory and in-hospital settings.

Study Design We conducted a real-world multicenter retrospective cohort study in patients who received medical attention for suspected COVID-19 in any of the 688 registered and accredited COVID-19 medical units in Mexico City, to evaluate all-cause mortality (main outcome) in those receiving antivirals, antibiotics, both, or none (exposition groups). . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 10 We considered 395,343 patients for eligibility evaluated for COVID-19 in 688 medical units (primary-to-tertiary care) between February 24, 2020 and September 14, 2020. All patients with a positive RT-PCR for SARS-CoV-2 were included to maximize the power and generalizability. Patients treated with unspecified antivirals were excluded. To perform reliable analyses, a cut-off value of 30 patients receiving the same antiviral was set and groups of antivirals with <30 patients were eliminated. Source of Data We used the COVID-19 open dataset available in Mexico City Government’s Open Data platform, 19 collected and updated daily by the Secretariat of Health of Mexico City. Patients meeting criteria of suspected COVID-19 case have been included in this dataset starting on February 24, 2020 when the first suspected cases arrived in Mexico. Criteria for suspected COVID-19 case in Mexico included having at least two of three signs/symptoms (cough, fever, or headache) plus at least one other (dyspnea, arthralgias, myalgias, sore throat, rhinorrhea, conjunctivitis, or chest pain) in the last 7 days. This operational definition was changed on August 24, 2020 to increase sensitivity: 20 at least one of four signs/symptoms (cough, fever, dyspnea, or headache), plus at least one other (myalgias, arthralgias, sore throat, chills, chest pain, rhinorrhea, anosmia, dysgeusia, or conjunctivitis) in the last 10 days. For epidemiologic purposes, two strategies are outlined in the National COVID-19 Epidemiologic Surveillance Plan: 20 1. testing of 10% of ambulatory patients with mild symptoms of respiratory disease and 100% of patients with respiratory distress at evaluation in monitoring units of viral respiratory disease (USMER, for its acronym in Spanish), and 2. testing 100% of patients who meet diagnostic criteria of Severe Acute . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 11 Respiratory Infection (defined as shortness of breath, temperature ≥ 38 °C, cough, and ≥ 1 of the following: chest pain, tachypnea, or acute respiratory distress syndrome) in non-USMER units. Upon evaluating a patient suspected of having COVID-19, healthcare professionals are required to fill out a format (Supplementary Appendix) containing demographic, clinical, epidemiological, and treatment variables, later complemented with follow-up by accredited hospital epidemiologists (inpatients) and healthcare professionals in primary care units (ambulatory patients). For ambulatory patients, follow-up is performed daily for a minimum 7 days and patients are considered recovered 14 days after the onset of symptoms if alive and not hospitalized. For hospitalized patients, follow-up is done daily until death or discharge; follow-up time for patients discharged from hospital is highly variable since no consensus or requirements by authorities exist but may extend from 14 days to 3-6 months after discharge. Duration of follow-up for each patient is not provided in the dataset and cannot be calculated. For every medical unit there is only one responsible authority who ultimately uploads data into the Respiratory Diseases Epidemiologic Surveillance System and is accountable for accuracy. Results of diagnostic RT-PCR for SARS-CoV-2 are directly uploaded by the diagnostic facility; accreditation of diagnostic procedures by the Mexican Institute of Diagnostics and Epidemiological Reference is required to upload results. Reporting of all deaths of COVID-19 suspected or confirmed cases is obligatory and must be done in the first 48 hours after occurrence; in cases of deaths occurring in patients who had completed follow-up, registries are matched to death certificates and updated. There have been concerns that patients tested more than once may be . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 12 duplicated. Since no variables that could lead to identification of patients are released, we searched for patients with identical demographic variables and only one registry was kept. To determine whether prescription of oseltamivir occurred for cases of COVID-19 or influenza during the pandemic period, dates of prescription of oseltamivir and dates of admission of patients with laboratory-confirmed COVID-19 were obtained from the previously mentioned COVID-19 dataset and grouped according to epidemiological weeks. Data of patients who had a positive test for influenza in Mexico City for every epidemiological week were obtained from the Directorate General of Epidemiology’s Weekly Surveillance Reports of Influenza. 21 Management of Variables All categorical variables were classified as dummy variables (present/absent). Polytomous variables were created from frequencies of use of antivirals and antibiotics (no antiviral/antibiotic, antiviral only, antibiotic only, and antiviral plus antibiotic), type of antiviral with >30 patients, and the combination of every individual antiviral with antibiotics. These were considered as the exposition groups. Special populations for subgroup analyses were defined as: children and adolescents (<18 years), pregnancy, puerperium, and non-pregnant/puerperal adults ( ≥ 18 years). Further subgroups included ambulatory and hospitalized patients, as well as patients admitted to intensive care unit (ICU) and those requiring invasive mechanical ventilation (IMV). A variable of critical patients was built by grouping patients admitted to ICU and/or requiring IMV, whereas non-critical patients did not meet any of both. . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 13 Since it has been hypothesized that early use of antivirals for COVID-19 could diminish hospitalization rate22 and detain disease progression,23 thereby decreasing mortality, we distinguished early (≤ 2 days from symptom onset to initiation of antivirals) from late (>2 days) use of antivirals, and studied their relation to hospitalization rates and mortality; only patients who received antivirals before being evaluated in an accredited COVID-19 unit were included for this analysis since dates of initiation of antivirals are only collected for such patients. Therefore, we studied these patients as a different sub- cohort. Occupations were grouped as follows: technical services (laborers), education (students and teachers), healthcare (dentists, nurses, diagnostic laboratorian, physicians, and other healthcare workers), agricultural activities (peasants), commerce (drivers, informal commerce, employees, and businesspeople), unemployed, stay-at-home (stay-at-home parents and retired/pensioners), and other occupations (others, and other professions). Variables for adjustment of models were: sex, age, indigenous self-identification, diabetes, COPD, immunosuppression, hypertension, HIV/AIDS, cardiovascular disease, obesity, CKD, smokers, unemployed, time from symptom onset to medical attention, fever, cough, sore throat, shortness of breath, irritability, diarrhea, chest pain, chills, headache, myalgias, arthralgias, abrupt deterioration, rhinorrhea, polypnea, vomit, abdominal pain, conjunctivitis, cyanosis, and sudden onset of symptoms. Statistical Analysis Descriptive data were calculated and are provided as frequencies, percentages, mean with standard deviation (SD) or median with interquartile range (IQR). Qualitative . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 14 comparisons were made with χ2 or Fisher´s exact test. Independent-samples t-test and ANOVA were used for quantitative comparisons. Survival was calculated for all treatment groups (antiviral only, antibiotic only, antiviral plus antibiotic, and no antiviral/antibiotic) and specific antivirals (acyclovir, amantadine, lopinavir-ritonavir, oseltamivir, rimantadine, and zanamivir) alone or combined with antibiotics; survival curves were created for general population, ambulatory, hospitalized, non-critical, and critical patients. Survival between groups receiving distinct treatments were compared through the Log-Rank test against patients not receiving antivirals/antibiotics. Cox regression models were applied for general population, ambulatory, hospitalized, non- critical, and critical patients to determine mortality risk in patients receiving any treatment compared to no antivirals/antibiotics (reference). Resulting hazard ratios (HR) were adjusted for demographic and clinical variables considered as risk factors in the univariate analysis for every group; all variables with p<0.1 were included in the final model using the Enter method. To account for multicenter variability, adjusted risk was calculated through generalized estimating equations (GEE), setting the medical unit with the lowest CFR and the highest number of patients for every subgroup as the reference value. Further subgroup survival analyses and multivariable Cox regression models were applied for special populations (children and adolescents, pregnancy, puerperium, and non- pregnant/puerperal adults), invasive mechanical ventilation (IMV), and intensive care unit (ICU). To quantify the minimal association strength of an unmeasured confounding factor that could reduce the risk conferred by exposures in our study, E-values were calculated for the point estimate and lower limit of the confidence interval. . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 15 To reduce potential confounding and selection bias, we applied a propensity score analysis. Propensity scores were calculated with a logistic regression model adjusted for: sex, age, signs and symptoms (fever, dyspnea, arthralgias, myalgias), and comorbidities (hypertension, diabetes, chronic kidney disease, cardiovascular disease, obesity, and immunosuppression). A 1:1 pairing was performed through the nearest- neighbor algorithm matching. Density functions of treated patients against controls before and after matching were graphed to determine appropriate matching. Subsequently, multivariable Cox regression models were applied to determine mortality risk in patients treated with antivirals, antibiotics, and oseltamivir versus a control group. A two-sided p value <0.05 was used to define statistical significance. Analyses and figures were created with SPSS software v.21 and GraphPad Prism v.8.0.1.

No duplicated registries were found. After selection of eligible participants (Figure 1), 136,855 patients from all 688 medical units were analyzed. 97.83% (n=133,887) were residents of the Mexico City Metropolitan Area, conformed by 17 municipalities of Mexico City (83.29%, n=111,768), and 60 municipalities (16.71%, n=22,119) of the State of Mexico. The remaining 2.17% (n=2,968) sought medical attention from all other 30 states of the republic. Of all patients, 10.0% (n=13,743) received antibiotics only; 3.0% (n=4,044), antivirals only; 3.6% (n=4,925), antivirals plus antibiotics, and 83.4% (n=114,143), none (Table 1). More symptomatic ambulatory patients received antivirals and antibiotics more . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 16 frequently (Supplementary Table 1); hospitalized patients with more signs/symptoms had greater use of antivirals, but less antibiotics (Supplementary Table 2). Baseline and follow-up characteristics of survivors (91.47%, n=136,855) and non- survivors (8.53%, n=11,679) are shown in Supplementary Table 3. Case-fatality rates (CFR) in special populations were: 8.92% (95%CI:8.76-9.07%), for non- pregnant/puerperal adults; 1.72% (95%CI:0.66-2.77), pregnancy; 0.97% (95%CI:0- 2.90), puerperium; and 0.69% (95%CI:0.48-0.90), children and adolescents. Of all deaths, 92.7% (95%CI:92.2-93.2) and 99.6% (95%CI:99.5-99.7) occurred by day 28 and 56, respectively. Patients treated only with antivirals had lower survival than those not receiving antivirals/antibiotics in the general population (Figure 2a), ambulatory (Figure 2b), hospitalized (Figure 2c), non-critical (Figure 3a), critical (Figure 3c), IMV (Supplementary Table 4), ICU (Supplementary Table 5) and non-pregnant/puerperal adults (Supplementary Table 6); for children and adolescents (Supplementary Table 7) and pregnancy (Supplementary Table 8) differences in survival were not significant; there were not enough events for analysis in puerperal women. Increased survival with only antibiotics was observed in hospitalized, critical, and IMV, whereas decreased survival occurred in the general population, non-pregnant/puerperal adults, ambulatory, non-critical, ICU, and children and adolescents; there were no differences for pregnancy. Antivirals plus antibiotics resulted in decreased survival in the general population, ambulatory, non-critical, non-pregnant/puerperal adults, children and adolescents, pregnancy, and ICU; increased survival, in hospitalized; and no differences, in critical and IMV groups. . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 17 Decreased survival with oseltamivir was observed in the general population (Figure 2d), ambulatory (Figure 2e), non-critical (Figure 2d), ICU (Supplementary Table 5), non- pregnant/puerperal adults (Supplementary Table 6), children and adolescents (Supplementary Table 7), and pregnancy (Supplementary Table 8); no differences in survival occurred in hospitalized (Figure 2f), critical (Figure 3d), and IMV (Supplementary Table 4). Survival rates for amantadine, zanamivir, rimantadine, acyclovir, and lopinavir-ritonavir are shown in the same figures and tables as oseltamivir. Unadjusted (Supplementary Table 9) and adjusted (Table 2) risk of death for the general population, ambulatory, hospitalized, non-critical and critical patients, as well as for other subgroups (Supplementary Tables 10-14) were calculated. E-values for statistically significant risk groups are provided in Supplementary Tables 15-16. After adjusting for center through GEE, we found no statistically significant variability in mortality risk for antivirals, antibiotics, or both in all groups. Oseltamivir presented variability in hospitalized and critical patients, with the largest increases in risk occurring in public hospitals. However, center was not a modifying risk factor after logistic regression analysis in hospitalized and critical patients. After matching patients receiving antivirals (Supplementary Table 17), oseltamivir (Supplementary Table 18), or antibiotics (Supplementary Table 19) to controls, mortality risks were similar to those in the unmatched cohort (Table 3). However, contrary to the main analyses, antivirals in hospitalized patients, as well as oseltamivir in hospitalized and critical patients were a risk factor for death. Antibiotics were a protective factor in hospitalized and critical patients, but not in ambulatory and non-critical patients. . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 18 Density functions before and after matching are shown in (Supplementary Figures 1-3). Of all 8,969 patients receiving antivirals, 10% (n=903) received antivirals before evaluation in COVID-19 accredited units; their baseline and follow-up characteristics are available in Supplementary Table 20. 25.2% (n=227) were admitted to hospital. Median time from symptom onset to initiation of antivirals was 1 day (IQR:0-4) for both ambulatory and hospitalized patients; time from symptom onset to ambulatory care in accredited units was 5 days (IQR:3-8) and 6 days (IQR:4-9) for hospitalization. Time from initiation of antivirals to hospitalization was 3 days (IQR:0-6). Time-to-initiation of antivirals and time-to-hospitalization for specific antivirals are shown in Supplementary Figure 4. Early (≤2 days) and late (>2 days) initiation of antivirals occurred in 64.2% (n=580) and 35.8% (n=323) patients, respectively. Overall survival in early (91.3%) and late (88.9%) groups was not different (p=0.2). Survival for early/late use of antivirals is shown in Supplementary Table 21. Oseltamivir was associated with increased risk of death in both early (HR=3.00, 95%CI:2.14-4.20) and late (HR=2.99, 95%CI:1.83-4.89) groups, as well as late use of lopinavir-ritonavir (HR=9.9, 95%CI:2.49-39.83); all other early/late antivirals did not reach statistical significance. There were no differences in hospitalization rates between early and late groups for every antiviral (Supplementary Figure 5). Prescription of oseltamivir during the pandemic period was greater than confirmed influenza cases (Figure 4A) and followed a similar pattern to that of weekly new cases of patients with a positive SARS-CoV-2 test (Figure 4B). . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 19

To the best of our knowledge, this is the first observational study evaluating amantadine, rimantadine, zanamivir, and acyclovir for COVID-19; no registered studies to evaluate these drugs exist. 11 Only one study has evaluated risk of death for oseltamivir;24 lopinavir-ritonavir have been evaluated in clinical trials.25 We hypothesized that antivirals and antibiotics could be having widespread use in real- world settings as repurposed drugs for COVID-19. Therefore, we studied mortality in laboratory-confirmed COVID-19 patients treated with common antivirals and/or antibiotics in Mexico City. Most patients did not receive them (83.4%), although a substantial proportion received antivirals alone (3.0%) or combined with antibiotics (3.6%) despite national guidelines advising against antivirals out of clinical trials. 15 Patients receiving antivirals and antibiotics were overall more symptomatic, suggesting that florid clinical presentations and not evidence may be guiding decision to treat. Our cohort included mostly middle-aged adults (44.2 [SD:16.8] years), of which non- survivors were younger and with a higher burden of hypertension, diabetes and obesity compared with other countries, findings consistent with previous studies in Mexican patients. 26 Rates of patients with cardiovascular diseases, COPD, asthma, and smokers were low, suggesting that underdiagnosis could be occurring. Only 20-30% of patients had been vaccinated against influenza during the prior season. Of patients treated with antivirals, ~10% (n=903) started antivirals before seeking medical attention in accredited COVID-19 units. Hospitalization rates were not different when antivirals were used early ( ≤ 2 days) vs late (>2 days). In Mexico, antibiotics and . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 20 most antivirals are sold under prescription. However, private pharmacy-associated clinics are a rapidly growing sector not included in our study where physicians tend to have lower experience, qualifications, compliance with regulations, and higher prescription rates.27–29 Self-medication with amantadine could be occurring since it is a widely available over-the-counter medication. Importantly, the high prescription of oseltamivir in Mexico City during the pandemic period cannot be attributed to influenza outbreaks, since patients receiving oseltamivir largely exceed those with a positive test for influenza and follows a similar weekly pattern to that of newly diagnosed COVID-19 patients. In one single-center study, oseltamivir was associated with decreased mortality risk in COVID-19-hospitalized patients (HR=0.21; 95%CI:0.10-0.43). 24 Contrary to Liu et al., we found no benefit for oseltamivir in hospitalized patients (HR=1.07; 95%CI:0.99-1.15) which is consistent with studies of oseltamivir for SARS-CoV (HR=0.87; 95%CI:0.55- 1.38). 30 Combination of oseltamivir with antibiotics in hospitalized patients in our study resulted in decreased risk of death (HR=0.92; 95%:0.87-0.98), which could explain findings by Liu et al. since most patients in their cohort (87.7%) received antibiotics. Decreased mortality is likely driven by antibiotics since hospitalized patients in our study receiving only antibiotics had lower risk of dying (HR=0.81, 95%CI:0.77-0.86) than antibiotics plus oseltamivir. After multivariable adjustment, oseltamivir was associated with increased mortality in the general population (HR=1.72, 95%CI:1.61-1.84), ambulatory (HR=4.79, 95%CI:4.01-5.75), non-critical (HR=2.05, 95%CI:1.88-2.23), and pregnant (HR=8.35, 95%CI:1.77-39.30) patients. Mortality risk was also higher in the cohort of 903 patients . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 21 with both early (HR=3.00, 95%CI:2.14-4.20) and late (HR=2.99, 95%CI:1.83-4.89) use of oseltamivir. We performed propensity score matching analyses to further account for potential confounders and found that hospitalized (HR=1.13, 95%CI:1.01-1.26) and critical patients (HR:1.22, 95%CI:1.05-1.43) also had an increased risk of death when treated with oseltamivir. Further efforts to limit bias were adjustment for center through GEE and calculation of E-values which support a strong association between oseltamivir and increased mortality risk, since potential confounders not accounted for in our study should have substantially large hazard ratios and lower limit confidence intervals (i.e. 9.05 and 7.48, respectively, for ambulatory patients) to explain our findings. The potentially inhibitory activity of proteases by oseltamivir 5 was found to be weak through molecular modeling, while inhibition of SARS-CoV-2 in vitro and reduction of symptoms in hospitalized patients failed. 31 Furthermore, oseltamivir nearly suppresses the production of pro-inflammatory cytokines by dendritic cells, polymorphonuclear leukocytes, and CD8+ T cells through inhibition of endogenous neuraminidase (sialidase), 32 which could impair the immune response and limit the capacity to eliminate the infection. Use of oseltamivir for infections caused by viruses lacking a neuraminidase gene (i.e. respiratory syncytial virus) decrease viral clearance, 33 which could explain increased mortality in patients receiving oseltamivir. Additionally, sudden (cardiac arrest, respiratory suppression, hypothermia, and neuropsychiatric events) and delayed-onset (impaired renal function, hyperglycemic events, prolonged QTc interval, immunological impairment, and bleeding) adverse reactions could explain increased risk of death in patients with COVID-19 treated with oseltamivir. 32 Antiviral drug-related heart . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 22 damage is a concern since some antivirals may be cardiotoxic, aggravating myocardial damage caused by SARS-CoV-2. 34 Renal and psychiatric adverse events in patients receiving oseltamivir were higher compared to placebo in a systematic review.35 In the RECOVERY study, there were no differences in mortality risk between hospitalized patients receiving lopinavir-ritonavir vs placebo (HR=1.03, 95%CI:0.91- 1.17), 25 which is consistent with our findings in hospitalized patients. Notably, ambulatory and late (>2 days) use of lopinavir-ritonavir were risk factors for death in our study. Antibiotics were a risk factor for death in the general population after multivariable analysis but a protective factor in both ambulatory and hospitalized patients. Nonetheless, univariate models showed no overall effect of antibiotics in ambulatory patients; when adjusting only for demographic variables no effect persisted but were protective after adjusting only for clinical variables. This is explained by the fact that more symptomatic patients received antibiotics more often. Supporting this conclusion, no benefit was observed for antibiotics in non-critical patients in both matched and unmatched cohorts. We observed a potential benefit for use of antibiotics in hospitalized, IMV, and critical patients. Increased survival in these groups could be due to prevention or treatment of concomitant bacterial infections, thereby supporting current WHO recommendations. 14 However, categorization of antibiotics as a single category in this dataset limits our study. Randomized clinical trials testing antibiotics in severe COVID-19 adult patients to prevent secondary infections would be important to corroborate this hypothesis. . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 23 For children and adolescents, antibiotics were a risk factor for death (HR=4.22, 95%CI:2.01-8.86). However, we did not differentiate ambulatory from hospitalized pediatric patients and current recommendations include using antibiotics in hospitalized patients with multisystem inflammatory syndrome. 36 The lack of benefit from antivirals included in our study in pediatric patients supports current guidelines discouraging their use.37 The main limitation of our study is that we were not able to assess cointerventions being studied for COVID-19 since only data for antivirals and antibiotics were available. Steroids increase survival in patients requiring oxygen administration and decrease survival in patients without supplementary oxygen.38 However, modifications of mortality risks by steroids in our population should be large according to E-values, reducing the likelihood that these associations could be due to steroids. Another potential limitation is that Mexico has had a low diagnostic testing rate for SARS-CoV-2 (0.08 daily tests per 1,000 people). 39 However, health authorities require 100% of patients with severe disease to be tested. Since we only studied mortality, an outcome expected to occur in patients who progress to severe disease, our study feasibly included most events. Nonetheless, excess mortality rates suggest there could be an undercounting of deaths in Mexico City. 39 These patients could have refrained from seeking medical attention or received medical care in non-accredited COVID-19 units where mortality, quality of care, and use of antivirals/antibiotics could be different. Also, the number of ICU beds in Mexico City was relatively low in March 2020 (6.0 per 100,000 population) compared to most European countries (5 to 33.9 per 100,000) in the pre-pandemic period; this capacity was expanded to 29.5 ICU beds per 100,000 by . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 24 September 2020.39,40 Mortality rates, especially in patients younger than 60 years, are lower under high availability of ICU beds. 40 Thus, mortality rates could have varied throughout our study period. Although we were not able to determine duration of follow-up in our study, the mechanisms and resources used by epidemiologic authorities in Mexico are robust enough to guarantee adequate matching of patients who had completed follow-up with death certificates. Thus, our finding that 92.7% (95%CI:92.2-93.2) and 99.6% (95%CI:99.5-99.7) of deaths occurred by day 28 and 56, respectively, could be important for the interpretation and design of COVID-19 clinical trials assessing short- term mortality.

Repurposed antivirals (oseltamivir, zanamivir, amantadine, rimantadine, acyclovir, and lopinavir/ritonavir) did not provide additional benefit, whereas oseltamivir consistently increased mortality risk. Antibiotics were associated with increased risk of death in the general population but decreased mortality risk only in hospitalized and critical patients, thereby supporting current WHO recommendations.

We thank Dr. César Sebastián Salinas Nájera, Head of Epidemiology in “Hospital Materno Infantil Magdalena Contreras” of Mexico City’s Health Secretariat for his valuable clarifications of epidemiologic surveillance mechanisms in Mexico City. We . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 25 want to acknowledge the Government of Mexico City for their effort to increase transparency and aid research by providing open datasets in their Open Data Platform; this sort of actions is most needed in low and middle-income countries to understand and solve the problems of our people. JMG would like to thank “Dirección General de Calidad y Educación en Salud” for supporting his participation in “Programa Nacional de Servicio Social en Investigación en Salud”. CONFLICTS OF INTEREST STATEMENT The authors declare no competing interests. FUNDING This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Data availability statement: The data that support the findings of this study are openly available in the Open Data Platform of Mexico City’s Government at https://datos.cdmx.gob.mx/explore/dataset/base-covid-sinave/information/, reference number 19, and the Directorate General of Epidemiology of Mexico City’s Weekly Epidemiological Surveillance Reports of Influenza 2020 at https://www.gob.mx/salud/documentos/informes-semanales-para-la-vigilancia- epidemiologica-de-influenza-2020, reference number 21. . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 26 Author contributions – Conception and design: JMG and AKG; Analysis and interpretation: AKG, JMG, JMS, JOGM, and HIRG; Drafting the manuscript, making tables and figures: All authors; Revising the manuscript for important intellectual content: AKG, JOGM, and HIRG. All authors approved the final draft of the manuscript. . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 27

1. Islam N, Sharp SJ, Chowell G, et al. Physical distancing interventions and incidence of coronavirus disease 2019: natural experiment in 149 countries. BMJ. July 2020:m2743. doi:10.1136/bmj.m2743 2. Sachs JD, Abdool Karim S, Aknin L, et al. Lancet COVID-19 Commission Statement on the occasion of the 75th session of the UN General Assembly. Lancet. 2020;396(10257):1102-1124. doi:10.1016/S0140-6736(20)31927-9 3. Saul S, Einav S. Old Drugs for a New Virus: Repurposed Approaches for Combating COVID-19. ACS Infect Dis. 2020;6(9):2304-2318. doi:10.1021/acsinfecdis.0c00343 4. Senanayake SL. Drug repurposing strategies for COVID-19. Futur Drug Discov. 2020;2(2):fdd-2020-0010. doi:10.4155/fdd-2020-0010 5. Muralidharan N, Sakthivel R, Velmurugan D, Gromiha MM. Computational studies of drug repurposing and synergism of lopinavir, oseltamivir and ritonavir binding with SARS-CoV-2 protease against COVID-19. J Biomol Struct Dyn. April 2020:1- 6. doi:10.1080/07391102.2020.1752802 6. Araújo R, Aranda-Martínez JD, Aranda-Abreu GE. Amantadine Treatment for People with COVID-19. Arch Med Res. 2020;51(7):739-740. doi:10.1016/j.arcmed.2020.06.009 7. Aranda Abreu GE, Hernández Aguilar ME, Herrera Covarrubias D, Rojas Durán F. Amantadine as a drug to mitigate the effects of COVID-19. Med Hypotheses. 2020;140(January):109755. doi:10.1016/j.mehy.2020.109755 . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 28 8. Li Z, Yang L. Underlying Mechanisms and Candidate Drugs for COVID-19 Based on the Connectivity Map Database. Front Genet. 2020;11(October):558557. doi:10.3389/fgene.2020.558557 9. Shi J, Xiao Y, Zhang Y, et al. Challenges of drug development during the COVID‐ 19 pandemic: Key considerations for clinical trial designs. Br J Clin Pharmacol. December 2020:1-16. doi:10.1111/bcp.14629 10. National Institute for Health Research. COVID-19 Therapeutics. Innovation Observatory. http://www.io.nihr.ac.uk/report/covid-19-therapeutics/. Published 2020. Accessed December 8, 2020. 11. World Health Organization. International Clinical Trials Registry Platform (ICTRP). World Health Organization. https://www.who.int/ictrp/en/. Published December 2020. Accessed December 8, 2020. 12. Bhimraj A, Morgan RL, Shumaker AH, et al. Infectious Diseases Society of America Guidelines on the Treatment and Management of Patients with COVID- 19. IDSA Guidelines. https://www.idsociety.org/practice-guideline/covid-19- guideline-treatment-and-management/. Published December 2020. Accessed December 8, 2020. 13. COVID-19 Treatment Guidelines Panel. Coronavirus Disease 2019 (COVID-19) Treatment Guidelines. National Institutes of Health. https://www.covid19treatmentguidelines.nih.gov/. Published December 2020. Accessed October 7, 2020. 14. World Health Organization. Clinical Management of COVID-19 - Interim . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 29 Guidance. WHO. https://apps.who.int/iris/handle/10665/332196. Published May 2020. Accessed December 8, 2020. 15. Ministry of Health of Mexico. [Recommendations for treatment of SARS-CoV-2 infection, the etiologic agent of COVID-19] [Spanish]. Guidance documents for healthcare professionals. https://coronavirus.gob.mx/personal-de- salud/documentos-de-consulta/. Published July 2020. Accessed December 8, 2020. 16. Ministry of Health of Mexico. [Diagnostic algorithm for acute respiratory disease in the context of the COVID19-influenza syndemic] [Spanish]. Guidance documents for healthcare professionals. https://coronavirus.gob.mx/personal-de- salud/documentos-de-consulta/. Published September 2020. Accessed December 8, 2020. 17. Falavigna M, Colpani V, Stein C, et al. Guidelines for the pharmacological treatment of COVID-19. The task force/consensus guideline of the Brazilian Association of Intensive Care Medicine, the Brazilian Society of Infectious Diseases and the Brazilian Society of Pulmonology and Tisiology. Rev Bras Ter Intensiva. 2020;32(2):166-196. doi:10.5935/0103-507X.20200039 18. Coz Yataco AO, Simpson SQ. Coronavirus Disease 2019 Sepsis: A Nudge Toward Antibiotic Stewardship. Chest. 2020;158(5):1833-1834. doi:10.1016/j.chest.2020.07.023 19. [Dataset] Government of Mexico City; 2020; [Covid-19 National Epidemiologic Surveilance System (SINAVE) Mexico City] [Spanish]; Open Data Platform; . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 30 https://datos.cdmx.gob.mx/explore/dataset/base-covid-sinave/information/; Accessed September 14, 2020. 20. Directorate General of Epidemiology of Mexico. [Standardized Guideline for Epidemiologic and Laboratory Surveillance of viral respiratory diseases, August 2020] [Spanish]. Mexican Secreatariat of Health. https://www.gob.mx/salud/documentos/lineamiento-estandarizado-para-la- vigilancia-epidemiologica-y-por-laboratorio-de-la-enfermedad-respiratoria-viral. Published August 2020. Accessed December 8, 2020. 21. [Dataset] Directorate General of Epidemiology of Mexico City; 2020; [Weekly Epidemiological Surveillance Reports of Influenza 2020] [Spanish]; Directorate General of Epidemiology; https://www.gob.mx/salud/documentos/informes- semanales-para-la-vigilancia-epidemiologica-de-influenza-2020; Accessed December 8, 2020. 22. Benlloch J-M, Cortés J-C, Martínez-Rodríguez D, Julián R-S, Villanueva R-J. Effect of the early use of antivirals on the COVID-19 pandemic. A computational network modeling approach. Chaos, Solitons & Fractals. 2020;140:110168. doi:10.1016/j.chaos.2020.110168 23. Lipsitch M, Perlman S, Waldor MK. Testing COVID-19 therapies to prevent progression of mild disease. Lancet Infect Dis. 2020;20(12):1367. doi:10.1016/S1473-3099(20)30372-8 24. Liu J, Zhang S, Wu Z, et al. Clinical outcomes of COVID-19 in Wuhan, China: a large cohort study. Ann Intensive Care. 2020;10(1):99. doi:10.1186/s13613-020- . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 31 00706-3 25. Horby PW, Mafham M, Bell JL, et al. Lopinavir–ritonavir in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial. Lancet. 2020;396(10259):1345-1352. doi:10.1016/S0140- 6736(20)32013-4 26. Mancilla-Galindo J, Vera-Zertuche JM, Navarro-Cruz AR, et al. Development and validation of the patient history COVID-19 (PH-Covid19) scoring system: a multivariable prediction model of death in Mexican patients with COVID-19. Epidemiol Infect. 2020;148:e286. doi:10.1017/S0950268820002903 27. López-Manning M, García-Díaz R. Doctors Adjacent to Private Pharmacies: The New Ambulatory Care Provider for Mexican Health Care Seekers. Value Heal Reg Issues. 2017;14:81-88. doi:10.1016/j.vhri.2017.08.010 28. Pérez-Cuevas R, Doubova S V., Wirtz VJ, Servan-Mori E, Dreser A, Hernández- Ávila M. Effects of the expansion of doctors’ offices adjacent to private pharmacies in Mexico: secondary data analysis of a national survey. BMJ Open. 2014;4(5):e004669. doi:10.1136/bmjopen-2013-004669 29. Wirtz VJ, Díaz-Portillo SP, Idrovo álva ro J, Dreser A, Bonilla FR, Matías-Juan B. Clinics adjacent to private pharmacies in Mexico: infrastructure and characteristics of the physicians and their remuneration. Salud Publica Mex. 2015;57(4):320. doi:10.21149/spm.v57i4.7575 30. Shi Q, Zhou Q, Wang X, et al. Potential effectiveness and safety of antiviral agents in children with coronavirus disease 2019: a rapid review and meta- . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 32 analysis. Ann Transl Med. 2020;8(10):624-624. doi:10.21037/atm-20-3301 31. Tan Q, Duan L, Ma Y, et al. Is oseltamivir suitable for fighting against COVID-19: In silico assessment, in vitro and retrospective study. Bioorg Chem. 2020;104(June):104257. doi:10.1016/j.bioorg.2020.104257 32. Hama R. The mechanisms of delayed onset type adverse reactions to oseltamivir. Infect Dis (Auckl). 2016;48(9):651-660. doi:10.1080/23744235.2016.1189592 33. Moore ML, Chi MH, Zhou W, et al. Cutting Edge: Oseltamivir Decreases T Cell GM1 Expression and Inhibits Clearance of Respiratory Syncytial Virus: Potential Role of Endogenous Sialidase in Antiviral Immunity. J Immunol. 2007;178(5):2651-2654. doi:10.4049/jimmunol.178.5.2651 34. Long B, Brady WJ, Koyfman A, Gottlieb M. Cardiovascular complications in COVID-19. Am J Emerg Med. 2020;38(7):1504-1507. doi:10.1016/j.ajem.2020.04.048 35. Jefferson T, Jones MA, Doshi P, et al. Neuraminidase inhibitors for preventing and treating influenza in adults and children. Cochrane Database Syst Rev. April 2014. doi:10.1002/14651858.CD008965.pub4 36. Harwood R, Allin B, Jones CE, et al. A national consensus management pathway for paediatric inflammatory multisystem syndrome temporally associated with COVID-19 (PIMS-TS): results of a national Delphi process. Lancet Child Adolesc Heal. 2020;19(20):1-10. doi:10.1016/S2352-4642(20)30304-7 37. Ye Z, Rochwerg B, Wang Y, et al. Treatment of patients with nonsevere and severe coronavirus disease 2019: an evidence-based guideline. Can Med Assoc . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 33 J. 2020;192(20):E536-E545. doi:10.1503/cmaj.200648 38. RECOVERY Collaborative Group, Horby P, Lim WS, et al. Dexamethasone in Hospitalized Patients with Covid-19 — Preliminary Report. N Engl J Med. July 2020:NEJMoa2021436. doi:10.1056/NEJMoa2021436 39. Roser M, Ritchie H, Ortiz-Ospina E, Hasell J. Coronavirus Pandemic (COVID-19). OurWorldInData.org. https://ourworldindata.org/coronavirus. Published September 2020. Accessed September 20, 2020. 40. Karagiannidis C, Mostert C, Hentschker C, et al. Case characteristics, resource use, and outcomes of 10 021 patients with COVID-19 admitted to 920 German hospitals: an observational study. Lancet Respir Med. 2020;8(9):853-862. doi:10.1016/S2213-2600(20)30316-7 . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 34 Table 1. Baseline characteristics of patients with laboratory confirmed COVID-19 who were treated with or without antivirals/antibiotics, in 688 accredited COVID-19 medical units in Mexico City. All patients n=136855 No antiviral / antibiotic n=114143 Antiviral only n=4044 Acyclovir n=36 Amantadine n=319 Lopinavir- Ritonavir n=100 Oseltamivir n=8414 Rimantadine n=61 Zanamivir n=39 Antibiotic only n=13743 Sex Women 66683 (48.7) 56999 (49.9) 1813 (44.8) 19 (52.8) 182 (57. 1) 31 (31) 3407 (40.5) 28 (45.9) 17 (43.6) 6000 (43.7) Men 70172 (51.3) 57144 (50.1) 2231 (55.2) 17 (47.2) 137 (42. 9) 69 (69) 5007 (59.5) 33 (54.1) 22 (56.4) 7743 (56.3) Age, mean (SD) 44.2 (16.8) 43.1 (16.6) 50.5 (16.5) 46.9 (14.9) 43.9 (14.8) 56.9 (15.9) 51.8 (15.9) 46 (15.1) 50 (14.1) 48.3 (16.8) Age categories 0-19 years 7558 (5.5) 6963 (6.1) 57 (1.4) 0 (0) 12 (3.8) 1 (1) 97 (1.2) 2 (3.3) 0 (0) 483 (3.5) 20-29 years 20098 (14.7) 18027 (15.8) 375 (9.3) 6 (16.7) 38 (11.9) 2 (2) 638 (7.6) 5 (8.2) 3 (7.7) 1379 (10) 30-39 years 29434 (21.5) 25586 (22.4) 707 (17.5) 6 (16.7) 86 (27.0) 10 (10) 1286 (15.3) 16 (26.2) 7 (17.9) 2437 (17.7) 40-49 years 29553 (21.6) 24683 (21.6) 837 (20.7) 10 (27.8) 71 (22.3) 21 (21) 1780 (21.2) 14 (23) 8 (20.5) 2966 (21.6) 50-59 years 24928 (18.2) 20011 (17.5) 852 (21.1) 5 (13.9) 63 (19.7) 23 (23) 1895 (22.5) 15 (24.6) 10 (25.6) 2906 (21.1) 60-69 years 15070 (11.0) 11441 (10.0) 632 (15.6) 7 (19.4) 32 (10) 18 (18) 1515 (18.0) 3 (4.9) 6 (15.4) 2048 (14.9) 70-79 years 7183 (5.2) 5213 (4.6) 418 (10.3) 2 (5.6) 14 (4.4) 17 (17) 855 (10.2) 5 (8.2) 5 (12.8) 1072 (7.8) 80-89 years 2594 (1.9) 1902 (1.7) 146 (3.6) 0 (0) 2 (0.6) 8 (8) 292 (3.5) 1 (1.6) 0 (0) 389 (2.8) 90-99 years 419 (0.3) 303 (0.3) 20 (0.5) 0 (0) 1 (0.3) 0 (0) 56 (0.7) 0 (0) 0 (0) 59 (0.4) ≥ 100 years 18 (0.01) 14 (0.01) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 4 (0.03) Indigenous self-identification 713 (0.5) 537 (0.5) 37 (0.9) 0 (0) 2 (0.6) 1 (1) 79 (0.9) 0 (0) 2 (5.1) 92 (0.7) Occupation Technical services 1916 (1.4) 1527 (1.3) 70 (1.7) 0 (0) 3 (0.9) 3 (3) 145 (1.7) 0 (0) 1 (2.6) 237 (1.7) Education 10006 (7.3) 9129 (8) 105 (2.6) 1 (2.8) 24 (7.5) 1 (1) 187 (2.2) 2 (3.3) 0 (0) 662 (4.8) Healthcare 17281 (12.6) 14910 (13.1) 655 (16.2) 3 (8.3) 47 (14.7) 9 (9) 1029 (12.2) 7 (11.5) 2 (5.1) 1274 (9.3) Agricultural activities 302 (0.2) 232 (0.2) 5 (0.1) 0 (0) 0 (0) 0 (0) 19 (0.2) 0 (0) 0 (0) 51 (0.4) Commerce 50450 (36.9) 42625 (37.3) 1078 (26.7) 15 (41.7) 111 (34.8) 36 (36) 2569 (30.5) 27 (44.3) 12 (30.8) 5055 (36.8) Other 24630 (18) 19906 (17.4) 911 (22.5) 6 (16.7) 63 (19.7) 20 (20) 2021 (24) 9 (14.8) 6 (15.4) 2599 (18.9) Unemployed 5685 (4.2) 4747 (4.2) 277 (6.8) 1 (2.8) 5(1.6) 2 (2) 463 (5.5) 3 (4.9) 7 (17.9) 457 (3.3) Stay-at-home 26585 (19.4) 21067 (18.5) 943 (23.3) 10 (27.8) 66 (20.7) 29 (29) 1981 (23.5) 13 (21.3) 11 (28.2) 3408 (24.8) Last-season flu vaccination 27087 (19.8) 22972 (20.1) 695 (17.2) 9 (25) 85 (26.6) 9 (9) 1244 (14.8) 13 (21.3) 4 (10.3) 2751 (20 ) Special populations Pregnancy 583 (0.9) 530 (0.9) 12 (0.3) 0 (0) 2 (1.1) 0 (0) 16 (0.5) 0 (0) 0 (0) 35 (0.6) Age during pregnancy, mean (SD) 29.8 (7.4) 29.5 (6.9) 30.3 (5.2) - 32 (4.2) - 30 (5.9) - - 34.8 (12.4) Last-season flu vaccination 161 (27.6) 153 (28.9) 1 (8.3) 0 (0) 0 (0) 0 (0) 1 (6.3) 0 (0) 0 (0) 7 (20) Pregnancy age group Early adolescent (≤14 years) 2 (0.3) 2 (0.4) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) Late adolescent (15-19 years) 34 (5.8) 33 (6.2) 1 (8.3) 0 (0) 0 (0) 0 (0) 1 (6.3) 0 (0) 0 (0) 0 (0) Normal age (20-34 years) 404 (69.3) 373 (70. 4) 9 (75) 0 (0) 1 (50) 0 (0) 12 (75) 0 (0) 0 (0) 18 (51.4) Advanced maternal age (≥35 years) 143 (24.5) 122 (23) 2 (16.7) 0 (0) 1 (50) 0 (0) 3 (18.8) 0 (0) 0 (0) 17 (48.6) Trimester of pregnancy First trimester 114 (19.6) 102 (19.2) 5 (41.7) 0 (0) 1 (50) 0 (0) 5 (31.3) 0 (0) 0 (0) 6 (17.1) Second trimester 177 (30.4) 161 (30.4) 3 (25) 0 (0) 0 (0) 0 (0) 4 (25) 0 (0) 0 (0) 12 (34.3) Third trimester 292 (50.1) 267 (50.4) 4 (33.3) 0 (0) 1 (50) 0 (0) 7 (43.8) 0 (0) 0 (0) 17 (48.6) Puerperium 103 (0.2) 64 (0.1) 2 (0.05) 0 (0) 0 (0.0) 0 (0) 7 (0.2) 0 (0) 0 (0) 32 (0.5) Days of puerperium 1 day 33 (32) 21 (32.8) 1 (50) 0 (0) 0 (0) 0 (0) 3 (42.9) 0 (0) 0 (0) 9 (28.1) 2-7 days 33 (32) 16 (25) 1 (50) 0 (0) 0 (0) 0 (0) 3 (42.9) 0 (0) 0 (0) 14 (43.8) . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 35 8-42 days 37 (35.9) 27 (42.2) 0 (0) 0 (0) 0 (0) 0 (0) 1 (14.3) 0 (0) 0 (0) 9 (28.1) Age during puerperium, mean (SD) 31.9 (9.7) 31.1 (9.2) 33.5 (3.5) - - - 31.1 (2.7) - - 33.6 (11.5) Last-season flu vaccination 22 (21.4) 14 (21.9) 14 (21.9) 0 (0) 0 (0) 0 (0) 1 (14.3) 0 (0) 0 (0) 7 (21.9) Children and adolescents (<18 years) 5791 (4.2) 5336 (4.7) 40 (1) 0 (0) 9 (2.8) 1 (1) 67 (0.8) 2 (3.3) 0 (0) 376 (2.7) Age, mean (SD) 10.9 (5.2) 10.9 (5.2) 10.9 (5.9) - 13.4 (4.5) - 11 (5.8) 14.5 (3.6) - 9.9 (5.9) Last-season flu vaccination 1213 (20.9) 1113 (20.9) 4 (10) 0 (0) 1 (11.1) 0 (0) 9 (13.4) 0 (0) 0 (0) 90 (23.9) Non-pregnant/puerperal adults (≥18 years) 130378 (95.3) 108213 (94.8) 3990 (98.7) 36 (100) 308 (96.6) 99 (9) 8324 (98.9) 59 (96.7) 39 (100) 13300 (96.8) Age, mean (SD) 45.7 (15.5) 44.7 (15.3) 50.9 (16.1) 46.9 (14.9) 44.8 (14) 57.5 (15.1) 52.1 (15.6) 47.1 (14.1) 50 (14.1) 48.5 (15 .7) Last-season flu vaccination 25691 (19.7) 21692 (20) 690 (17.3) 9 (25) 84 (27.3) 9 (9.1) 1233 (14.8) 13 (22) 4 (10.3) 2647 (19.9 ) Comorbidities Diabetes 18229 (13.3) 13458 (11.8) 910 (22.5) 2 (5.6) 36 (11.3) 35 (35) 2007 (23.9) 8 (13.1) 6 (15.4) 2677 (19.5) COPD 1741 (1.3) 1273 (1.1) 119 (2.9) 0 (0) 3 (0.9) 2 (2) 212 (2.5) 2(3.3) 1 (2.6) 248 (1.8) Asthma 3035 (2.2) 2561 (2.2) 96 (2.4) 1 (2.8) 13 (4.1) 4 (4) 165 (2) 0 (0) 3 (7.7) 288 (2.1) Immunosuppression 1758 (1.3) 1368 (1.2) 84 (2.1) 0 (0) 7 (2.2) 3 (3) 145 (1.7) 0 (0) 0 (0) 235 (1.7) Hypertension 22185 (16.2) 16799 (14.7) 1074 (26.6) 10 (27.8) 40 (12.5) 35 (35) 2211 (26.3) 12 (19.7) 5 (12.8) 3073 (22.4) HIV/AIDS 573 (0.4) 462 (0.4) 168 (4.2) 0 (0) 0 (0) 2 (2) 47 (0.6) 0 (0) 0 (0) 62 (0.5) Cardiovascular disease 2724 (2.0) 2064 (1.8) 153 (3.8) 2 (5.6) 4 (1.3) 5 (5) 277 (3.3) 2 (3.3) 0(0) 370 (2.7) Obesity 23848 (17.4) 18924 (16.6) 837 (20.7) 12 (33.3) 83 (26) 18 (18) 1817 (21.6) 5 (8.2) 6 (15.4) 2983 (21.7) Chronic kidney disease 2067(1.5) 1471 (1.3) 150 (3.7) 0 (0) 2 (0.6) 7 (7) 286 (3.4) 1 (1.6) 0 (0) 300 (2.2) Smoker 14727 (10.8) 12214 (10.7) 461 (11.4) 2 (5.6) 46 (14.4) 6 (6) 885 (10.5) 8 (13.1) 5 (12.8) 1561 (11.4) Type of medical attention Ambulatory 109902 (80.3) 98060 (85.9) 2012 (49.8) 30 (83.3) 282 (88.4) 8 (8) 3187 (37.9) 48 (78.7) 19 (48.7) 8268 (60.2) Hospitalization 26953 (19.7) 16083 (14.1) 2032 (50.2) 6 (16.7) 37 (11.6) 92 (92) 5227 (62.1) 13 (21.3) 20 (51.3) 5475 (39.8) Severity of the disease Non-critical 129658 (94.7) 110009 (96.4) 3518 (87) 34 (94.4) 310 (97.2) 69 (69) 7126 (84.7) 58 (95.1) 34 (87.2) 12018 (87.4) Critical 7197 (5.3) 4134 (3.6) 526 (13) 2 (5.6) 9 (2.8) 31 (31) 1288 (15.3) 3 (4.9) 5 (12.8) 1725 (12.6) Time from symptom onset to medical attention 4.5 (3.8) 4.4 (3.8) 4.49 (3.9) 6.36 (4.6) 4.9 (3.6) 6.8 (4.1) 4.7 (3.7) 5.6 (4.2) 5.33 (4.7) 5.5 (3.8) Baseline symptoms Fever 83120 (60.7) 66011 (57.8) 3332 (82.4) 25 (69.4) 198 (62.1) 81 (81) 6963 (82.8) 42 (68.9) 31 (79.5) 9769 (71.1) Cough 96206 (70.3) 78367 (68.7) 3406 (84.2) 24 (66.7) 245 (76.8) 67 (67) 7055 (83.8) 42 (68.9) 35 (89.7) 10371 (75.5) Sore throat 59040 (43.1) 48845 (42.8) 2014 (49.8) 12 (33.3) 159 (49.8) 31 (31) 3701 (44) 29 (47.5) 22 (56.4) 6241 (45.4) Shortness of breath 42942 (31.4) 31061 (27.2) 2234 (55.2) 14 (38.9) 85 (26.6) 66 (66) 5170 (61.4) 22 (36.1) 23 (59) 6501 (47.3) Irritability 24098 (17.6) 19460 (17) 1079 (26.7) 5 (13.9) 81(25.4) 7 (7) 1993 (23.7) 14 (23) 13 (33.3) 2525 (18.4) Diarrhea 31649 (23.1) 25821 (22.6) 1152 (28.5) 11 (30.6) 96 (30.1) 19 (19) 2180 (25.9) 15 (24.6) 13 (33.3) 3494 (25.4) Chest pain 36851 (26.9) 29524 (25.9) 1662 (41.1) 12 (33.3) 105 (32.9) 22 (22) 2979 (35.4) 25 (41) 15 (38.5) 4169 (30.3) Chills 48282 (35.3) 39405 (34.5) 2138 (52.9) 16 (44.4) 158 (49.5) 34 (34) 3616 (43) 33 (54.1) 17 (43.6) 5003 (36.4) Headache 95018 (69.4) 78893 (69.1) 3284 (81.2) 22 (61.1) 227 (71.2) 59 (59) 6348 (75.4) 40 (65.6) 29 (74.4) 9400 (68.4) Myalgias 70666 (51.6) 57192 (50.1) 2633 (65.1) 27 (75) 201 (63.0) 60 (60) 5074 (60.3) 43 (70.5) 24 (61.5) 8045 (58.5) Arthralgias 64381 (47) 51792 (45.4) 2377 (58.8) 22 (61.1) 179 (56.1) 48 (48) 4846 (57.6) 42(68.9) 23 (59) 7429 (54.1) Abrupt deterioration 62460 (45.6) 48991 (42.9) 2704 (66.9) 20 (55.6) 180 (56.4 66 (66) 5367 (63.8) 34 (55.7) 20 (51.3) 7782 (56.6) Rhinorrhea 38288 (28) 32135 (28.2) 1350 (33.4) 6 (16.7) 118 (37.0) 20 (20) 2283 (27.1) 27 (44.3) 11 (28.2) 3688 (26.8) Polypnea 15868 (11.6) 11977 (10.5) 1081 (26.7) 3 (8.3) 46 (14.4) 13 (13) 1847 (22) 7 (11.5) 6 (15.4) 1969 (14.3) Vomit 10123 (7.4) 8094 (7.1) 479 (11.8) 3 (8.3) 23 (7.2) 7 (79 870 (10.3) 8 (13.1) 2 (5.1) 1116 (8.1) Abdominal pain 17338 (12.7) 14080 (12.3) 1025 (25.3) 3 (8.3) 60 (18.8) 8 (8) 1561 (18.6) 10 (16.4) 7 (17.9) 1609 (11.7) Conjunctivitis 16941 (12.4) 14277 (12.5) 513 (12.7) 4 (11.1) 55 (17.2) 7 (7) 962 (11.4) 11 (18) 6 (15.4) 1619 (11.8) Cyanosis 5917 (4.3) 4461 (3.9) 463 (11.4) 1 (2.8) 12 (3.8) 4 (4) 816 (9.7) 8 (13.1) 3 (7.7) 612 (4.5) Sudden onset of symptoms 46723 (34.1) 37607 (32.9) 1574 (38.9) 10 (27.8) 96 (30. 1) 48 (48) 3831 (45.5) 26 (42.6) 16 (41) 5089 (37) Concomitant use of antibiotics 18840 (13.8) 172 (0.2) - 29 (80.6) 151 (47.3) 77 (77) 4627 (55) 26 (42.6) 15 (38.5) - Data expressed as Frequency (%) or mean (SD) SD: Standard deviation, COPD: Chronic obstructive pulmonary disease, HIV/AIDS: Human immunodeficiency virus/acquired immune deficiency syndrome. . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 36 Table 2. Adjusted mortality risk in laboratory-confirmed COVID-19 patients receiving antivirals, antibiotics, both, or none in 688 accredited COVID-19 medical units in Mexico City. All patientsa Ambulatoryb Hospitalizedc Non-Criticald Criticale HR (95%CI) p HR (95%CI) p HR (95%CI) p HR (95%CI) p HR (95%CI) p Models for type of treatment No antiviral / antibiotic Reference Reference Reference Reference Reference Antiviral only 1.72 (1.61-1.84) <0.0001 4.7 (3.94-5.62) <0.0001 1.07 (0.99-1.15) 0.07 2.03 (1.86-2.21) <0.0001 1.09 (0.99-1.21) 0.09 Antibiotic only 1.13 (1.08-1.19) <0.0001 0.72 (0.58-0.89) 0.003 0.81 (0.77-0.86) <0.0001 1.05 (0.98-1.14) 0.2 0.67 (0.63-0.72) <0.0001 Antiviral + antibiotic 1.57 (1.47-1.67) <0.0001 1.91 (1.47-2.49) <0.0001 0.91 (0.86-0.97) 0.004 1.63 (1.49-1.77) <0.0001 1.02 ( 0.93-1.11) 0.7 Models for type of antiviral No antiviral / antibiotic Reference Reference Reference Reference Reference Acyclovir 1.37 (0.51-3.65) 0.5 Not estimable 2.75 (1.03-7.33) 0.04 1.19 (0.29-4.75) 0.8 2.85 (0.71-11.4) 0.1 Amantadine 0.73 0.44-1.21) 0.2 0.08 0.24-2.36) 0.6 0.88 (0.5-1.55) 0.7 0.67 (0.33-1.34) 0.3 1.05 (0.49-2.21) 0.9 Lopinavir-Ritonavir 1.04 (0.69-1.55) 0.9 4.28 (0.59-30.7) 0.1 0.59 (0.4-0.89) 0.01 0.69 (0.33-1.46) 0.3 0.66 (0.41-1.04) 0.08 Oseltamivir 1.66 (1.58-1.75) <0.0001 3.52 (3.01-4.11) <0.0001 0.98 (0.93-1.03) 0.4 1.84 (1.72-1.96) <0.0001 1.06 (0.99-1.14) 0.1 Rimantadine 1.39 (0.66-2.92) 0.4 2.54 (0.36-18.1) 0.4 1.11 (0.49-2.46) 0.8 1.48 (0.56-3.95) 0.4 1.63 (0.52-5.09) 0.4 Zanamivir 1.66 (0.83-3.32) 0.2 2.49 (0.35-17.8) 0.4 0.84 (0.39-1.76) 0.6 1.43 (0.46-4.43) 0.5 0.7 (0.29-1.69) 0.4 Antibiotic only 1.14 (1.08-1.19) <0.0001 0.72 (0.58-0.9) 0.004 0.81 (0.77-0.86) <0.0001 1.06 (0.98-1.14) 0.2 0.68 (0.63-0.72) < 0.0001 Models for Acyclovir No antiviral / antibiotic Reference Reference Reference Reference Reference Acyclovir only 8.1 (1.14-57.6) 0.04 Not estimable 8.98 (1.26-63.9) 0.03 Not estimable 2.85 (0.39-20.3) 0.3 Antibiotic only 1.11 (1.05-1.17) <0.0001 0.71 (0.57-0.89) 0.002 0.82 (0.77-0.86) <0.0001 1.03 (0.96-1.11) 0.4 0.67 (0.63-0.72) <0.0001 Acyclovir + antibiotic 1.07 (0.35-3.33) 0.9 Not estimable 2.28 (0.74-7.08) 0.2 1.23 (0.31-4.92) 0.8 3.11 (0.44-22.2) 0.3 Models for Amantadine No antiviral / antibiotic Reference Reference Reference Reference Reference Amantadine only 1.78 (1.03-3.06) 0.04 1.69 (0.42-6.79) 0.5 1.62 (0.89-2.93) 0.1 1.63 (0.78-3.42) 0.2 1.39 (0.62-3.1) 0.4 Antibiotic only 1.11 (1.05-1.17) <0.0001 0.71 (0.57-0.89) .002 0.82 (0.77-0.86) <0.0001 1.03 (0.96-1.12) 0.4 0.67 (0.63-0.72) < 0.0001 Amantadine + antibiotic 0.15 (0.04-0.59) 0.007 0.34 (0.05-2.39) 0.3 0.15 (0.02-1.06) 0.06 0.13 (0.02-0.9) 0.04 0.44 (0.06-3.11) 0.4 Models for Lopinavir-Ritonavir . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 37 No antiviral / antibiotic Reference Reference Reference Reference Reference Lopinavir-Ritonavir only 0.68 (0.26-1.82) 0.4 56.9 (7.87-412) <0.0001 0.39 (0.15-1.05) 0.06 0.47 (0.07-3.37) 0.5 0.41 (0.15-1.08) 0.07 Antibiotic only 1.11 (1.05-1.17) <0.0001 0.71 (0.57-0.89) 0.002 0.82 (0.77-0.86) <0.0001 1.04 (0.96-1.12) 0.4 0.67 (0.63-0.72) <0.0001 Lopinavir-Ritonavir + antibiotic 1.1 (0.71-1.7) 0.7 Not estimable 0.67 (0.43-1.05) 0.08 0.69 (0.31-1.53) 0.4 0.79 (0.47-1.34) 0.4 Models for Oseltamivir No antiviral / antibiotic Reference Reference Reference Reference Reference Oseltamivir only 1.72 (1.61-1.84) <0.0001 4.79 (4.01-5.75) <0.0001 1.07 (0.99-1.15) 0.06 2.05 (1.88-2.23) <0.0001 1.11 (1.0-1.2 3) 0.05 Antibiotic only 1.13 (1.08-1.19) <0.0001 0.72 (0.58-0.89) 0.003 0.81 (0.77-0.86) <0.0001 1.06 (0.98-1.14) 0.2 0.67 (0.63-0.72) <0.0001 Oseltamivir + antibiotic 1.61 (1.51-1.71) <0.0001 2.1 (1.65-2.8) <0.0001 0.92 (0.87-0.98) 0.01 1.68 (1.55-1.83) <0.0001 1.02 (0.94-1.12) 0.6 Models for Rimantadine No antiviral / antibiotic Reference Reference Reference Reference Reference Rimantadine only 1.88 (0.85-4.21) 0.1 4.9 (0.69-34.9) 0.1 1.21 (0.5-2.91) 0.7 1.81 (0.58-5.62) 0.3 1.69 (0.54-5.27) 0.4 Antibiotic only 1.11 (1.05-1.17) <0.0001 0.71 (0.57-0.89) <0.0001 0.82 (0.77-0.86) <0.0001 1.03 (0.96-1.12) 0.4 0.67 (0.63-0.72 ) <0.0001 Rimantadine + antibiotic 0.51 0.07-3.6) 0.5 Not estimable 0.77 (0.11-5.45) 0.8 0.88 (0.12-6.23) 0.9 - - Models for Zanamivir No antiviral / antibiotic Reference Reference Reference Reference Reference Zanamivir only 1.9 (0.85-4.25) 0.12 3.99 (0.55-28.9) 0.2 0.9 (0.37-2.17) 0.8 1.2 (0.17-8.49) 0.9 0.72 (0.29-1.74) 0.7 Antibiotic only 1.11 (1.05-1.17) <0.0001 0.71 (0.57-0.89) <0.0001 0.82 (0.77-0.86) <0.0001 1.03 (0.96-1.12) 0.4 0.67 (0.63-0.72 ) <0.0001 Zanamivir + antibiotic 1.14 (0.28-4.55) 0.9 Not estimable 0.74 (0.18-2.94) 0.7 1.57 (0.39-6.29) 0.5 - - HR: Hazard ratio, 95%CI: 95% confidence intervals a: Model adjusted by : Sex (men), Age, Indigenous self-identification, Diabetes, Chronic obstructive pulmonary disease, Immunos uppression, Hypertension, Human immunodeficiency virus/acquired immune deficiency syndrome, Cardiovascular disease, Obesity, Chronic kidney disease, Unemployed, Time from symptom onset to medical attention, Fever, Cough, Shortness of breath, Irritability, Chest pain, Chills, Myalgias, Arthralgias, Abrupt deterioration, Po lypnea, Abdominal pain, Cyanosis. b: Model adjusted by: Sex (men), Age, Indigenous self-identification, Diabetes, Chronic obstructive pulmonary disease, Immunosu ppression, Hypertension, Cardiovascular disease, Obesity, Chronic kidney disease, Unemployed, Time from symptom onset to medical attention, Fever, Cough, Shortness of breath, I rritability, Chest pain, Chills, Myalgias, Arthralgias, Abrupt deterioration, Polypnea, Vomit, Abdominal pain, Cyanosis. c: Model adjusted by: Sex (men), Age, Diabetes, Chronic obstructive pulmonary disease, Immunosuppression, Hypertension, Cardiov ascular disease, Chronic kidney disease, Smoker, Unemployed, Cough, Shortness of breath, Chest pain, Chills, Myalgias, Arthralgias, Abrupt deterioration, Polypnea, Cyan osis. d: Model adjusted by: Sex (men), Age, Indigenous self-identification, Diabetes, Chronic obstructive pulmonary disease, Immunosu ppression, Hypertension, Hypertension, Human immunodeficiency virus/acquired immune deficiency syndrome, Cardiovascular disease, Obesity, Chronic kidney disease, Unemployed, Time from symptom onset to medical attention, Fever, Cough, Shortness of breath, Irritability, Chest pain, Myalgias, Arthralgias, Polypnea, Vomit, Abdominal pain, Cyanosis. e: Model adjusted by: Sex (men), Age, Diabetes, Chronic obstructive pulmonary disease, Immunosuppression, Hypertension, Cardiov ascular disease, Chronic kidney disease, Unemployed, Time from symptom onset to medical attention, Cough, Sore throat, Shortness of breath, Chest pain, Headache, Myalgi as, Arthralgias, Rhinorrhea, Polypnea, Abdominal pain, Cyanosis. . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 38 Table 3. Adjusted mortality risk in laboratory-confirmed COVID-19 patients receiving antivirals, oseltamivir, or antibiotics after propensity score matching. All patients Ambulatory Hospitalized Non-Critical Critical Models for antivirals n=8088 n=4794 n=3294 n=7232 n=856 HR (95%CI) p HR (95%CI) p HR (95%CI) p HR (95%CI) p HR (95%CI) p Antiviral only 1.77 (1.61-1.95) <0.0001 4.83 (3.55-6.56) <0.0001 1.13 (1.01-1.25) 0.03 1.99 (1.74-2.27) <0.0001 1.10 (0.95-1.28 ) 0.2 Models for oseltamivir n=7574 n=4412 n=3162 n=6760 n=814 HR (95%CI) p HR (95%CI) p HR (95%CI) p HR (95%CI) p HR (95%CI) p Oseltamivir 1.78 (1.61-1.97) <0.0001 4.32 (3.19-5.84) <0.0001 1.13 (1.01-1.26) 0.03 1.87 (1.64-2.13) <0.0001 1.22 (1.05-1.43) 0 .01 Models for antibiotics n=27486 n=18548 n=8938 n=24844 n=2642 HR (95%CI) p HR (95%CI) p HR (95%CI) p HR (95%CI) p HR (95%CI) p Antibiotic only 1.14 (1.07-1.22) <0.0001 0.78 (0.61-1.00) 0.05 0.79 (0.74-0.85) <0.0001 0.98 (0.89-1.08) 0.7 0.61 (0.56-0.68) < 0.0001 HR: Hazard ratio, 95%CI: 95% confidence intervals . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint 39 Figure 1. Flow diagram of patients assessed for eligibility. Figure 2. Survival of patients (general population, ambulatory, and hospitalized) treated with antivirals and/or antibiotics. Survival curves are shown according to treatment modality in the general population (a), ambulatory (b), and hospitalized (c) patients. Survival in patients receiving specific antivirals, antibiotics, both, or none in the general population (d), ambulatory (e), and hospitalization (f) settings. Figure 3 . Survival of patients (non-critical and critical) treated with antivirals and/or antibiotics. Survival curves are shown according to treatment modality in non-critical (a) and critical (b) patients. Survival in critical (c) and non-critical (d) patients receiving specific antivirals, antibiotics, both, or none. Figure 4 . Weekly prescription of oseltamivir and newly diagnosed patients with influenza (a) and COVID-19 (b) during the COVID-19 pandemic in Mexico City. *Data of prescription of oseltamivir for patients tested only for influenza was not collected. . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint . CC-BY-NC-ND 4.0 International licenseIt 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 January 1, 2021. ; https://doi.org/10.1101/2020.10.13.20211797doi: medRxiv preprint