BACKGROUND: The coronavius disease 2019 (COVID-9) caused by the severe acute respiratory syndrome coronavirus 2 reached Spain by 31 January 2020, in April 2020, the Comunidad de Madrid suffered one of the world's highest crude mortality rate ratios. This study aimed to detect risk factors for mortality in patients with COVID-19. METHODS: Our cohort were all consecutive adult patients with laboratory-confirmed COVID-19 at a secondary hospital in Madrid, March 3-16, 2020. Clinical and laboratory data came from electronic clinical records and were compared between survivors and non-survivors, with outcomes followed up until April 4. Univariable and multivariable logistic regression methods allowed us to explore risk factors associated with in-hospital death. FINDINGS: The cohort comprised 562 patients with COVID-19. Clinical records were available for evaluation for 392 patients attended at the emergency department of our hospital, of whom 199 were discharged, 85 remained hospitalized and 108 died during hospitalization. Among 311 of the hospitalized patients, 34.7% died. Of the 392 patients with records, the median age was 71.5 years (50.6-80.7); 52.6% were men. 252 (64.3%) patients had a comorbidity, hypertension being the most common: 175 (44.6%), followed by other cardiovascular disease: 102 (26.0%) and diabetes: 97 (24.7%). Multivariable regression showed increasing odds of in-hospital death associated with age over 65 (odds ratio 8.32, 95% CI 3.01-22.96; p<0.001), coronary heart disease (2.76, 1.44-5.30; 0.002), and both lower lymphocyte count (0.34, 0.17-0.68; 0.002) and higher LDH (1.25, 1.05-1.50; 0.012) per 1-unit increase and per 100 units respectively. INTERPRETATION: COVID-19 was associated in our hospital at the peak of the pandemic with a crude mortality ratio of 19.2% and a mortality ratio of 34.7% in admitted patients, considerably above most of the ratios described in the Chinese series. These results leave open the question as to which factors, epidemiological or intrinsically viral, apart from age and comorbidities, can explain this difference in excess mortality.