key: cord-0982766-sq01hlaa
authors: Giannouchos, Theodoros V.; Sussman, Roberto A; Mier, José M; Poulas, Konstantinos; Farsalinos, Konstantinos
title: Characteristics and risk factors for COVID-19 diagnosis and adverse outcomes in Mexico: an analysis of 89,756 laboratory–confirmed COVID-19 cases
date: 2020-07-30
journal: Eur Respir J
DOI: 10.1183/13993003.02144-2020
sha: 49d1ac7f0b917bcc133220f4bb965afc6bb95b04
doc_id: 982766
cord_uid: sq01hlaa

BACKGROUND: There is insufficient information about risk factors for COVID-19 diagnosis and adverse outcomes from low and middle-income countries (LMICs). OBJECTIVES: We estimated the association between patients’ characteristics and COVID-19 diagnosis, hospitalisation and adverse outcome in Mexico. METHODS: This retrospective case series used a publicly available nation-level dataset released on May 31, 2020 by the Mexican Ministry of Health, with patients classified as suspected cases of viral respiratory disease. Patients with COVID-19 were laboratory-confirmed. Their profile was stratified by COVID-19 diagnosis or not. Differences among COVID-19 patients based on two separate clinical endpoints, hospitalisation and adverse outcome, were examined. Multivariate logistic regressions examined the associations between patient characteristics and hospitalisation and adverse outcome. RESULTS: Overall, 236 439 patients were included, with 89 756 (38.0%) being diagnosed with COVID-19. COVID-19 patients were disproportionately older, males and with increased prevalence of one or more comorbidities, particularly diabetes, obesity, and hypertension. Age, male gender, diabetes, obesity and having one or more comorbidities were independently associated with laboratory-confirmed COVID-19. Current smokers were 23% less likely to be diagnosed with COVID-19 compared to non-smokers. Of all COVID-19 patients, 34.8% were hospitalised and 13.0% experienced an adverse outcome. Male gender, older age, having one or more comorbidities, and chronic renal disease, diabetes, obesity, COPD, immunosuppression and hypertension were associated with hospitalisation and adverse outcome. Current smoking was not associated with adverse outcome. CONCLUSION: This largest ever case series of COVID-19 patients identified risk factors for COVID-19 diagnosis, hospitalisation and adverse outcome. The findings could provide insight for the priorities the need to be set, especially by LMICs, to tackle the pandemic.

complications and neurological symptoms. [1] [2] [3] Risk factors for adverse outcomes include age, hypertension, diabetes, cardiovascular disease and respiratory disease. 4 The pandemic represents a big challenge particularly for low-and middle-income countries (LMIC). The cost of epidemiologic surveillance and of infection prevention and control, the sudden flow and the need to enhance the constrained critical care capacity to treat patients and the implementation of non-medical interventions such as social distancing measures are expected to significantly stress the limited financial resources of these countries. 5 Therefore, understanding the factors associated with COVID-19 susceptibility and adverse prognosis is crucial to guide local authorities towards more efficient allocation of their scarce resources to avoid exceeding the limited capacity of the healthcare system.

In this study we present evidence and information about patients who were screened for COVID-19 due to suspected viral respiratory infection through the respective surveillance system implemented in Mexico. The objective of this study was to examine the association between individuals' sociodemographic and clinical characteristics and COVID-19 diagnosis.

Additionally, we explored similar associations with two clinical outcomes, hospitalization and adverse outcome, within the COVID-19 diagnosed cohort. This study extends the current literature by presenting novel information and evidence about COVID-19 patients in Mexico using a large and recent cohort of such patients, with potentially important implications at the clinical and policy level.

We performed a cross-sectional secondary data analysis using a publicly available individual level dataset which included patients classified as ‗suspected cases of viral respiratory disease' during point of service at medical facilities in Mexico. The 

Upon admission, patients were screened by healthcare professionals who were expected to verify that the subjects show specific symptoms documented as inclusion criteria for the dataset. Additionally, they recorded data about the medical history on a specific DGE form.

After the case was evaluated and confirmed at the district, state and national level surveillance system, it was added to the dataset. Both USMER and Non-USMER units had to fill the same forms which were sent to an online platform (SISVER platform Laboratory testing to confirm SARS-CoV-2 infection was performed according to WHO interim guidance. 8 Combined nasopharyngeal and oropharyngeal swabs were obtained and placed in a container. For intubated patients, bronchoalveolar lavage was obtained. In case of death, lung biopsies were obtained during autopsy, from an area visibly affected by disease.

The samples were sent to the nearest Laboratory of Respiratory Virus (InDRE) for testing with RT-PCR.

The dataset included information on COVID-19 testing results, categorized as positive, negative and pending, as well as individual level data on sociodemographic and clinical characteristics and facility specific information. Sociodemographic information included patients' age, gender and nationality (Mexican or not). Clinical characteristics included existing comorbidities, namely diabetes, chronic obstructive lung disease (COPD), asthma, immunosuppression, hypertension, cardiovascular disease, obesity, chronic renal disease and other comorbidities (not defined). No data on the duration or time of diagnosis of comorbidities, pharmacotherapy or clinical condition of patients relative to the comorbidities were available. Additionally, comorbidity classifications were not further defined according to specific diagnoses. For example, no information on specific cardiovascular diseases (coronary artery disease, heart failure, arrhythmias etc.) were reported. Finally, comorbidities were recorded based on past medical history; thus, there was no record of the diagnostic criteria used for each comorbidity, similarly to other studies. [9] [10] [11] [12] Smoking was also recorded, with participants classified as smokers or non-smokers. No data on former smokers was available. All information was recorded on a specific -Respiratory Triage form by the attending physicians. 7 For patients with laboratory-confirmed COVID-19 diagnosis, the dataset included additional information related to clinical endpoints, namely whether the patient was admitted into an intensive care unit, intubated, or died. Facility level information included a dichotomous indicator on whether the healthcare unit was part of the USMER network or not, and the type of facility where the patient was diagnosed.

The first outcome variable of interest in the study was whether a patient was diagnosed with COVID-19 or not, defined as a dichotomous indicator. We thus excluded 36,803 patients (13.4%) with pending results, resulting in a final sample of 238,194 patients. We also explored two outcomes within the subgroup of patients with COVID-19 diagnosis, hospitalization and severity. Both outcomes were dichotomous; the first was hospitalization and the second was adverse outcome defined as intubation, intensive care unit admission or death.

We included individual level sociodemographic and clinical characteristics, and facility information based on data availability. Sociodemographic patient-level information included age, gender, and Mexican nationality (or not). Clinical information included number of comorbidities, whether the patient had a particular clinical condition or not, namely asthma, chronic obstructive pulmonary disease (COPD), diabetes, obesity, hypertension, immunosuppression, cardiovascular condition, and chronic renal disease. We also included information on whether the patient was a current smoker and whether there was previous contact with someone who was diagnosed with COVID-19. Facility specific information included the type of facility by ownership and whether the medical unit is a monitoring unit for respiratory diseases (USMER). USMER consists of medical facilities which monitor the incidence of infectious respiratory diseases as part of the government's Epidemiological Surveillance system.

We initially conducted descriptive analyses of all patients to characterize the overall study population. We then performed stratified bivariate analyses to compare patients based on whether they were diagnosed with COVID-19 or not. Subsequently, laboratory-confirmed COVID-19 cases were stratified by the two outcomes of interest (hospitalization and adverse outcome) using similar bivariate analyses. We tested for statistical differences in the stratified analyses using Pearson's chi-square for categorical variables and the non-parametric Mann Whitney U test on the age variable (as numeric).

To estimate the association of all the independent variables with the three outcomes of interest (COVID-19 positive diagnosis for all patients, hospitalization and adverse outcome only for COVID-19 positive patients), we then conducted two multivariate logistic regressions for each of the three outcomes (six regressions in total). For each outcome, both models included the same sociodemographic and facility specific variables. The first model also included clinical comorbidities as dichotomous indicators, while the second included the number of clinical diagnoses (comorbidities) only, due to multicollinearity, as the later was derived from the clinical diagnoses. We also included area-of-residence fixed-effects to control for unobserved regional variations. Finally, standard errors were clustered at the residence level. All analyses were conducted using Stata (version 16.1; StataCorp, College Station, TX).

Our final analytic sample included 236,439 patients who were suspects of a viral respiratory disease, after excluding 1,755 patients (0.7%) due to missing variables. The majority were 18 to 44 years of age, Mexicans, and around 40% of those had one or more comorbidities, while 9.0% were current smokers ( Table 1) . The most prevalent clinical conditions included hypertension, diabetes, and obesity. The age distribution of comorbidities is presented in Supplementary Table 1. About 37% of the patients used a USMER facility and more than half used hospitals of the Ministry of Health (Secretaria de Salubridad y Asistencia -SSA).

Around one-third (38.0%) of patients were diagnosed with COVID-19 (Table 1) . These patients had higher shares of males (56.4% versus 47.5% for non-COVID-19 patients, p<0.001), and were 6 years older on average (p<0.001). COVID-19 patients had also disproportionately higher shares of one or more comorbidities (p<0.001), and chronic conditions particularly related to diabetes, hypertension, and obesity (p<0.001 for all). We also observed greater shares for USMER related COVID-19 cases at the facility level (p<0.001 for both). Table 2 indicates the results of the two regressions on COVID-19 diagnosis for all patients in the data. Across both models, females and younger patients (0 to 17) were significantly less likely to be diagnosed with COVID-19 compared to males and to their 18 to 44 counterparts (p<0.001for all). In addition, current smokers were approximately 23% less likely to be diagnosed with COVID-19 compared to non-smokers (p<0.001 in both models). In contrast, older patients (45 years of age or older) and those with one or more comorbidities were more likely to be diagnosed with COVID-19 compared to those aged 18 to 44 and those without comorbidities respectively. Diabetes and obesity were particularly associated with COVID-19 diagnosis compared to patients without such conditions (p<0.001 for all).

Among the 89,756 patients who were diagnosed with COVID-19, about 35% were hospitalized and 13% had high clinical severity ( Table 3) . Across both subgroups, hospitalization and adverse outcome were more frequent in males and older patients, with a mean age difference of more than 12 years. Patients with one or more comorbidities, particularly those with hypertension, obesity, diabetes, and COPD were also more prevalent in both the hospitalized and the adverse outcome groups. Table 4 indicates the results of the two regressions for hospitalization and adverse outcome, respectively (Table 4) . Across both models, males and older patients were significantly more likely to be hospitalized and to experience adverse outcome compared to females and to their 18 to 44 old counterparts (p<0.001 for all). Those 0 to 17 years of age were also less likely to experience adverse outcome compared to the 18 to 44 years age group. In addition, patients with chronic renal disease, diabetes, immunosuppression, COPD, obesity, and hypertension were up to 121% (adjusted OR: 2.21, 95% CI: 1.91-2.55, for those with chronic renal disease) more likely to experience hospitalization and more severe composite endpoints compared to those without such conditions (p<0.001 for all). Similarly, having one or more comorbidities increased the likelihood of these outcomes, as expected ( Table 5) . Finally, we did not observe any significant differences among current smokers compared to non-smokers across both outcomes.

To the best of our knowledge, this study presents a case series with the highest number of laboratory-confirmed COVID-19 patients, and the first of this size from a LMIC. The Chinese Centers for Disease Control (CDC) recently presented data from 44,672 confirmed cases, however information about comorbidities was available for only 20,812 patients. 9 Another study in the UK examined risk factors for critical care and mortality in hospital among 20,133 hospitalized COVID-19 patients. 10 In the US, patient characteristics, comorbidities and outcomes were presented among 5,700 patients hospitalized for COVID-19 in New York City area. 11 Our study adds to the current evidence by presenting information from laboratory-confirmed cases in a LMIC using a large publicly available dataset.

In agreement with previous studies, 1,10-14 age was a strong risk factor for hospitalization, adverse outcome among COVID-19 patients. It was recognized early during the pandemic that the elderly had higher rates of hospitalization and infection fatality ratios compared to younger people. 15 Recently, the US CDC reported that the best estimates for the COVID-19 symptomatic case fatality ratio was 0.4% for the whole population but ranged from 0.05% for those aged 0-49 years to 1.3% for those ≥ 65 years, a 26-fold difference. 16 Hospitalization rates were estimated to be approximately 7-fold higher for patients aged ≥ 65 years compared to those aged 18 to 44 years in our study. Thus, targeted interventions tailored at the higher needs and risk of older people are clearly needed, to protect them from SARS-CoV-2 infection and to reduce the COVID-19 morbidity and mortality burden.

Our results indicate that cardiovascular and endocrine conditions were the most common comorbidities identified among confirmed COVID-19 patients. Hypertension, obesity and diabetes were not only common comorbidities but also independent correlates of hospitalization and adverse outcomes. These findings are in-line with case series from China, the US and Europe. 1, [9] [10] [11] [12] [13] These conditions are very common worldwide and in Mexico. Approximately 1.4 billion adults are estimated to suffer from hypertension globally, with the prevalence being higher in LMICs. 17 In Mexico the prevalence of hypertension was 25.5% in 2016. 18 Obesity is also a major healthcare issue in Mexico. In a random sample of 2,511 adults, 38.3% of Mexicans were overweight and almost 25% were obese. 19 The Organization for Economic Cooperation and Development (OECD) reports that Mexico is the one of the countries with the highest rates of obesity in the population. 20 Obesity is a risk factor for diabetes, and Mexico had an estimated 10.4% prevalence of diabetes in 2016 with a continuously increasing trend. 21, 22 The latter, in combination with the increased prevalence of these conditions among COVID-19 patients suggests that such patients represent another population subgroup where targeted interventions and guidance are needed to prevent SARS-CoV-2 transmission.

In contrast, while cardiovascular disease and COPD were risk factors for hospitalization, and adverse outcomes, only a small proportion of patients suffered from these comorbidities. A case series of 1,590 patients from China reported a similarly low prevalence of these comorbidities among Chine patients. 12 The COPD prevalence in Mexico City was 3.4% in a study defining airflow obstruction as FEV1/FEV 6 below the 5th percentile or Lower Limit of Normal, 23 but it has been reported that COPD is highly underdiagnosed in Mexico and in other countries, mainly because of lack of spirometry evaluation. 24 In a 2009 study, ischemic heart disease and stroke prevalence in Mexico City ranged from 0.4% to 5.4% and from 0.4% to 3.5%, respectively, depending on age. 25 In addition, other risk factors for adverse outcomes were immunosuppression and chronic renal disease. Our findings are supported by a recent systematic review and meta-analysis which found a higher risk for adverse COVID-19 outcomes among patients with chronic renal disease. 26 Having more than one comorbidity was strongly associated with hospitalization and adverse outcome. This is not unexpected considering that multiple comorbidities contribute to disease complexity and such patients are more susceptible and vulnerable to adverse events.

Approximately 1 in 5 patients with laboratory-diagnosed COVID-19 had more than 1 comorbidity, and they had approximately 3-fold higher odds for hospitalization and adverse outcome. Therefore, it is necessary to prioritize the assessment of these patients, offering early diagnosis and proper hospital care, while primary preventive measures to reduce disease transmission to these patients are warranted.

Notably, smoking was not associated with a higher risk for adverse outcomes and hospitalization. Smokers were also less likely to be diagnosed with COVID-19 compared to non-smokers. The latter is in agreement with a recent observational population study from Israel. 27 Some studies have found that smokers are under-represented among COVID-19 patients and presented a hypothesis that nicotine may exert protective effects, 28-31 while others have found that nicotine and smoking causes ACE2 up-regulation which may increase viral invasion. 31, 32 Recent meta-analyses reported that hospitalized smokers with COVID-19 had higher odds for adverse outcomes, 33-35 but very few smokers relative to the population smoking rates appear to be hospitalized for COVID-19. 35 It is possible that more smokers have been tested for COVID-19 compared to non-smokers, which could explain the lower odds for positive diagnostic test. This cannot be directly addressed in this study since all participants were by definition subjects who were tested for COVID-19. However, according to the latest data (2016), the smoking prevalence in Mexico was 14.0% in the population aged ≥ 15 years. 36 In our study, 9.0 % of the sample were smokers. Even if we assume that none of the participants aged 0-17 were smokers (4.8% of the total sample), still smokers would represent approximately 9.6% of the adult sample, lower than the population smoking rates.

Thus, it is unlikely that smokers were more likely to be tested for COVID-19 based on the proportion of smokers in the study sample and the population smoking rates. It is currently unclear whether nicotine exerts any positive effect, however, there is no doubt that smoking cannot be used as a protective measure and smoking cessation should be encouraged during the COVID-19 pandemic. 29 It has been established that SARS-CoV-2 uses angiotensin converting enzyme 2 (ACE2) as a receptor for cell entry and viral replication. 37 While this would imply that ACE2 upregulation would be associated with COVID-19 severity and adverse outcome, there is evidence that the opposite is the case. Risk factors for adverse outcomes identified in this and other studies, such as age, male gender, endocrine and cardiovascular diseases, are associated with lower levels of ACE2. [38] [39] [40] [41] [42] Therefore, it has been hypothesized that ACE2 deficiency may in fact be detrimental for COVID-19. 43 Additionally, severe COVID-19 represents a hyper-inflammatory response with patients developing cytokine storm and exhibiting ineffective regulation of the immune response. 44 Risk factors identified in this study also represent inflammatory conditions. [45] [46] [47] [48] Thus, there is a relevant pathophysiological basis explaining the association between hospitalized and adverse COVID-19 outcomes and the comorbidities identified in our analysis.

This study is not without limitations. First, given the nature and the availability of the data, we were not able to use more detailed clinical and laboratory information for the patients. For example, immunodeficiency may include a vast array of different disease conditions, however no specific information was provided in the dataset. No information was available on pharmacotherapy for comorbidities or the clinical condition of patients in relation to these comorbidities (controlled or decompensated at the time of COVID-19 diagnosis), which could have affected the outcome. Additionally, we could not examine whether specific disease conditions (e.g. coronary heart disease vs. congestive heart failure, emphysema vs. chronic bronchitis) would differently affect the outcome. We were not able to include facility and regional specific information, due to the lack of such information in the dataset.

However, we believe that we sufficiently addressed the heterogeneity between hospitals and regions using the appropriate variables in our analyses, given the research question of interest. The study sample of laboratory-confirmed COVID-19 cases was not derived from a random sample of the general population but from cases with suspected respiratory disease.

This type of selection bias is almost universally applicable to case series of COVID-19 patients considering that laboratory testing capacities are limited worldwide and are usually prioritized for those with more severe symptoms or in patients with comorbidities and at risk for severe disease. Still, caution is advised in generalizing the conclusions to the general population or to population-representative samples. Finally, some of the patients may have not recovered by the time the dataset was released and, thus, the outcome is unknown, while it is also possible that some outpatients or patients with mild disease may experience disease progression and will thus require hospitalization in the future. This limitation would have been particularly problematic had the epidemic wave been at an early stage, with many new disease cases but limited outcomes due to the time lag from disease diagnosis to final outcome. However, Mexico confirmed its first COVID-19 cases on February 28 while the first death was recorded on March 18. 49 Therefore, the epidemic wave in Mexico was sufficiently advanced at the time we obtained the dataset, and substantial discrepancies in the patient characteristics between new disease cases and outcomes are probably excluded. Still, this should be considered as a limitation, and further analysis as the epidemic progresses is warranted.

In conclusion, this large retrospective case series from Mexico, the largest ever presented for COVID-19, identified risk factors for laboratory-confirmed COVID-19 diagnosis as well as for hospitalization and adverse outcome among COVID-19 patients. These findings could provide valuable insight for Mexico and other LMICs in setting priorities, allocating healthcare resources and establish disease transmission preventive strategies in order to protect vulnerable groups, particularly the elderly and people with comorbidities. 

Zhong NS; China Medical Treatment Expert Group for Covid-19. Clinical Characteristics of Coronavirus Disease 2019 in China

Prevalence of venous thromboembolism in patients with severe novel coronavirus pneumonia

Neurologic Features in Severe SARS-CoV-2 Infection

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Managing COVID-19 in Low-and Middle-Income Countries

Lineamiento estandarizado para la vigilancia epidemiológica y por laboratorio de la enfermedad respiratoria viral (Standardized guidelines for epidemiological and laboratory surveillance of the respiratory viral disease). Secretaría de Salud. Abril 2020

Preparación y respuesta frente a casos de SARS-CoV2-2019 para la atención primaria a la salud(Preparation and response to SARS-CoV2-2019 cases for primary health care

Clinical management of severe acute respiratory infection when novel coronavirus (2019-nCoV) infection is suspected: interim guidance

The Epidemiological Characteristics of an Outbreak of 2019 Novel Coronavirus Diseases (COVID-19) -China

Semple MG; ISARIC4C investigators. Features of 20 133 UK patients in hospital with covid-19 using the ISARIC WHO Clinical Characterisation Protocol: prospective observational cohort study

Presenting Characteristics, Comorbidities, and Outcomes Among 5700 Patients Hospitalized With COVID-19 in the

He JX; China Medical Treatment Expert Group for Covid-19. Comorbidity and its impact on 1590 patients with Covid-19 in China: A Nationwide Analysis

Clinical Characteristics of Covid-19 in

Renin-Angiotensin-Aldosterone System Inhibitors and Risk of Covid-19

Estimates of the severity of coronavirus disease 2019: a model-based analysis

US Centers for Disease Control (CDC)

The global epidemiology of hypertension

Obesity in Mexico: prevalence, comorbidities, associations with patient outcomes, and treatment experiences

Obesity update

Ending diabetes in Mexico

Burden of type 2 diabetes in Mexico: past, current and future prevalence and incidence rates

Chronic Obstructive Pulmonary Disease in Latin America. Ann Glob Health

BOLD Collaborative Research Group, the EPI-SCAN Team, the PLATINO Team, and the PREPOCOL Study Group. Determinants of underdiagnosis of COPD in national and international surveys

The prevalence of chronic diseases and major disease risk factors at different ages among 150,000 men and women living in Mexico City: cross-sectional analyses of a prospective study

Prognosis of COVID-19 in Patients with Liver and Kidney Diseases: An Early Systematic Review and Meta-Analysis

Smoking and the risk of COVID-19 in a large observational population study

COVID-19 and the nicotinic cholinergic system

Systematic review of the prevalence of current smoking among hospitalized COVID-19 patients in China: could nicotine be a therapeutic option?

Current smoking is not associated with COVID-19

Editorial: Nicotine and SARS-CoV-2: COVID-19 may be a disease of the nicotinic cholinergic system

ACE-2 Expression in the Small Airway Epithelia of Smokers and COPD Patients: Implications for COVID-19

Smoking is Associated with COVID-19 Progression: A Meta-Analysis

Impact of smoking status on disease severity and mortality of hospitalized patients with COVID-19 infection: a systematic review and meta-analysis

Current smoking, former smoking, and adverse outcome among hospitalized COVID-19 patients: a systematic review and meta-analysis

Smoking prevalence, total (ages 15+) -Mexico

Smoking Upregulates Angiotensin-Converting Enzyme-2 Receptor: A Potential Adhesion Site for Novel Coronavirus SARS-CoV-2 (Covid-19)

Cell Entry Depends on ACE2 and TMPRSS2 and Is Blocked by a Clinically Proven Protease Inhibitor

Age-and gender-related difference of ACE2 expression in rat lung

COVID-19, Diabetes Mellitus and ACE2: The conundrum

Loss of angiotensin-converting enzyme 2 accelerates maladaptive left ventricular remodeling in response to myocardial infarction

COVID-19 infection and the predictive ACE2 soluble levels: the favourable protection of children and women. Front. Pediatr

The pivotal link between ACE2 deficiency and SARS-CoV-2 infection

The cytokine release syndrome (CRS) of severe COVID-19 and Interleukin-6 receptor (IL-6R) antagonist Tocilizumab may be the key to reduce the mortality

Obesity, inflammation and diet. Pediatr Gastroenterol Hepatol Nutr

Inflammation and hypertension: new understandings and potential therapeutic targets

Inflammation and cardiovascular disease: from pathogenesis to therapeutic target

Glycemic Control and Risk of Infections Among People With Type 1 or Type 2 Diabetes in a Large Primary Care Cohort Study

COVID-19 coronavirus pandemic