30-day mortality in patients hospitalized with COVID-19 during the first wave of the Italian epidemic: A prospective cohort study

1. Introduction

In late December 2019, an outbreak of an emerging disease (COVID-19) caused by a novel coronavirus that was later named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was recorded in Wuhan, China [1], and subsequently rapidly spread to a substantial number of Asian and non-Asian countries. It was declared a pandemic by the World Health Organisation on 12 March 2020 [2].

The increasing number of studies conducted in Chinese hospitals over the last few months have contributed to delineating the characteristics of the disease and its lethality [[3], [4], [5], [6]]. They describe COVID-19 as an atypical SARS-like pneumonia that requires intensive care in 26–33 % of patients, 4–15 % of whom eventually die [4,5]. However, as the epidemic moves outside China, there is a need to verify the clinical features and lethality of COVID-19 in countries with different demographic characteristics.

Italy was the first European country to be hit hard by the COVID-19 epidemic, with Lombardy in northern Italy being the region in which the first autochthonous cases were identified and the largest epidemic foci developed [7]. Italy is also the European country in which the highest number of COVID-19 deaths have so far been recorded (24,780 as of 27 April 2020) [8].

The Department of Infectious Diseases of Luigi Sacco Hospital (the national reference centre for epidemiological emergencies and bioterrorism in northern Italy) has been admitting SARS-CoV-2 patients (particularly those coming from the “red zone” municipalities first involved in the epidemic) since the night of 20 February 2020, when the first COVID-19 case was identified in a town about 50 km from Milan [9].

This paper describes the demographic and clinical characteristics of the COVID-19 patients admitted to our hospital between 21 February and 19 March 2020 in the early stage of the Italian epidemic, and the factors associated with the risk of COVID-19-related death.

2. Materials and methods

3. Results

Between 21 February and 19 March 2020, a total of 233 COVID-19 patients were admitted to L. Sacco Hospital, Milan, Italy. Most were males (69.1 %) of Italian nationality (92.8 %), and their median age was 61 years (IQR 50–72). Twenty-six (11.2 %) were healthcare workers. A total of 133 (57.1 %) were resident in the city or metropolitan area of Milan (the first case in Milan was recorded on 23 February 2020). Twenty-seven (11.6 %) came from the “red zones” and 71 were transferred from other provinces in Lombardy (Lodi, Cremona, and Bergamo) whose hospitals were overwhelmed by the sudden explosion of the epidemic.

Forty-eight patients (20.6 %) died during the study period, and 185 survived, including 162 (69.5 %) patients who were discharged, and 23 (9.9 %) who were still hospitalised on the censoring date.

Table 1 shows the differences in the baseline demographic and clinical characteristics of the survivors and non-survivors. The non-survivors included a higher proportion of subjects aged 66−75 (39.6 % vs 19.5 %), 76−85 (20.8 % vs 13.0 %), and 86−95 years (10.4 % vs 2.7 %) (p < 0.001); a higher proportion of patients transferred from other hospitals (62.5 % vs 36.8 %, p < 0.002). The non-survivors were more frequently being treated with anti-platelet agents (p = 0.009), calcium channel blockers (p = 0.023) and angiotensin II receptor blockers (p = 0.001). Conversely, the survivors were more frequently without any co-medication on hospital admission when compared to the non-survivors (37.8 % vs 18.8 %, p = 0.016).

There were no significant between-group differences in terms of time from the onset of symptoms to hospital admission (overall median 7 days, IQR 4−9), and no significant differences in symptoms and signs upon admission. The most frequent presenting signs were fever, cough and dyspnea in both groups.

A higher proportion of survivors had chest X-rays without any pathological findings (16.2 % vs 4.2 %; p = 0.033), but there was no difference in the pathological X-ray patterns between the two groups. A higher proportion of non-survivors presented with severe or critical disease (p < 0.001).

Table 2 shows the baseline laboratory findings. The non-survivors had significantly lower median lymphocyte (p = 0.008), hemoglobin (p = 0.009) and albumin levels (p < 0.001), and significantly higher median CRP (p < 0.001), D-dimer (p < 0.001) and creatinine levels (p < 0.001).

During the hospitalization 172 (73.8 %) patients received a combination of LPV/r plus HCQ, 39 (16.7 %) with remdesivir of whom 33 (14.2 %) after LPV/r plus HCQ and 42 (18 %) with tocilizumbab of whom 35 (83.3 %) after LPV/r plus HCQ. Ten patients received all the 3 combinations. One hundred and forty-four (61.8 %) patients were treated with at least one antibiotic during the hospital stay.

The median follow-up of the cohort as a whole was 40 days (IQR 33–47): eleven days (IQR 6–18) for the non-survivors and 44 days (IQR 38–49) for the survivors. The median hospital stay was 12 days (IQR 8–21). Kaplan Meier curve analysis showed that the time-dependent probability of death ten, 20 and 30 days after hospital admission was respectively 10.3 % (95 % confidence interval [CI] 6.4–14.2 %), 16.3 % (95 % CI 11.6–21.1) and 19.7 % (95 % CI 14.6–24.9 %) (Fig. 1 ).

Table 3 shows the uni- and multivariable Cox analysis of the factors associated with the risk of death. Age (adjusted hazard ratio [aHR] 2.08, 95 % CI 1.48−2.92 per ten years more) and obesity (aHR 3.04, 95 % CI 1.42−6.49) were independently associated with an increased risk of death, which was also associated with critical disease (aHR 8.26, 95 % CI 1.41−48.29), CRP levels (aHR 1.17, 95 % CI 1.02−1.35 per 50 mg/L more) and CK levels above 185 U/L (aHR 2.58, 95 % CI 1.37−4.87) upon admission. Conversely, the multivariable model did not confirm the univariable findings of an increased risk of death in patients receiving at least one anti-hypertensive agent, those with anemia, or those with D-dimer levels of >1000 μg/L upon admission.

4. Discussion

We studied the characteristics and outcome of 233 adult COVID-19 patients hospitalised in Milan during the early dramatic days of the Italian epidemic. Forty-eight patients (20.6 %) died during the study period, with the probability of dying at 30 days from hospital admission being of 19.7 %. The overall case fatality rate observed in our cohort was similar to that found in a recent study of 201 COVID-19 patients hospitalised in Wuhan [13], but higher when compared to the 14 % estimated by Wu et al. in the early period of the epidemic in China [14].

In line with the findings of some Chinese studies [13,15], our patients were prevalently male, which suggests a gender-based need for different hospital care during SARS-CoV-2 infection. It has been suggested that males may be more prone to developing severe and fatal COVID-19 [16,17], and recent data regarding the epidemic in Europe shows a male-to-female death ratio of 2.1 that increases to 3.9 in patients aged 50–65 years [18]. However, our multivariable analysis did not reveal a significant gender-based difference in the survival rate. The male/female ratio among the patients admitted to our hospital during study period was 2.24, and so the reasons for the findings of the preliminary European report may be that females are less likely to be hospitalised with COVID-19.

The median age of our patients was 61 years which was similar to that reported in a recent large case series from the New York city area (63 years, IQR 52–75) [19], as against the 47–56 years reported in studies of Chinese hospitals [20,21]; like the fact that the proportion of our patients aged >75 years was 18.9 %, this difference probably mirrors the demographic differences between Italy and China [22]. Preliminary data from Chinese studies indicated that COVID-19 was more lethal in the elderly than younger people [13,15], and we also found that an older age was an independent risk factor for death, as in the case of SARS and Middle East respiratory syndrome-related coronavirus [23,24]. It has been speculated that older patients may be more likely to die of COVID-19 because age-related alterations in immunological functions and type 2 cytokine production lead to deficiencies in controlling SARS-CoV-2 replication and pro-inflammatory responses [15].

Our analysis showed that obese patients had a 3-fold higher risk of dying as compared to those with a body mass index below 30. This finding is in line with recently published studies suggesting that obesity represents one of the most important factors related to COVID-19 severity as evidenced by higher need of hospitalization and of invasive mechanical ventilation [25,26]. Several mechanisms have been proposed to explain the increased severity of COVID-19 in obese patients, including the combination of reduced cardiorespiratory reserve and impairment of adaptive immune response to infections [27] It should be notice that the proportion of obese patients in our cohort was much lower than that recently described in a large case series by Richardson et al. (16.3 % vs 41.7 %) and it cannot be excluded that in countries with high prevalence rate of obesity the effect of this condition on the burden of COVID-19 related mortality may be greater [19].

In line with the findings of previous Chinese studies [13,15], fever, cough and dyspnoea were the most frequent signs/symptoms at the time of hospital admission; however, there was no significant difference between survivors and non-survivors in terms of symptoms upon admission.

The most frequent chest X-ray alterations were monolateral consolidations and interstitial alterations, followed by bilateral lung consolidations; however, no lung alterations were detected at x-ray in 13.7 % of cases.

Most of our patients had mild or moderate disease [12] but, as expected, greater disease severity at the time of admission was strongly associated with an increased risk of death; in particular, the patients presenting with critical disease requiring assistance in ICU were at 8 times higher risk than those who did not.

In line with the study by Luo et al. we found that serum CRP level upon admission was independently associated with adverse outcome of COVID-19 [28]. It has been shown that in pulmonary diseases marked by inflammatory features there is a typical raise in serum CRP level in response to inflammatory cytokines such as IL-6, IL-1 or TNF-α. [29]. Thus, higher CRP level in non-survivors of our study may indicate excessive and dangerous inflammatory response. We also found a significant correlation between higher CK levels upon admission and the risk of death. Elevated level of CK in COVID-19 patients might be a sign of respiratory muscle injury resulting from the increased demands placed on the respiratory system. There also previous evidence suggesting that serum levels of CK may rise in patients with pneumonia and pulmonary embolic disease [30]. Furthermore, it cannot be excluded that the increase in CK values in COVID-19 may be related to a damage to CK-rich tissues, such as skeletal and cardiac muscle and brain, directly induced by the virus or maladaptive immune responses. Autopsy studies are required to define the damage of organs that can cause CK increase during COVID-9.

Our study has a number of limitations. Firstly, as in China, the dramatically evolving scenario of the epidemic in Italy required continuous structural, organisational and staff changes that exposed the study to maturation bias. In particular, the number of laboratory examinations performed upon admission was limited (i.e. Interleukine-6 determination became available only after 15 days from the start of the study). Secondly, only 14.8 % of our patients underwent a chest computed tomography upon admission because of barriers in our infrastructure created in order to ensure a dedicated COVID-19 service. Thirdly, it was difficult to ascertain the different effects on outcomes of the miscellaneous and often concomitant drug treatments given to our patients because of the absence of a standard of care for COVID-19 (excluding oxygen supplementation). Nevertheless, as all of the patients admitted to our Infectious Diseases Department were enrolled in this study and the study population was probably representative of the COVID-19 patients hospitalised in Italy in the early stage of the epidemic.

In conclusion, case-fatality rate of patients hospitalized with COVID-19 in the early days of the Italian epidemic was about 20 %. Older age, obesity, disease severity upon admission were factors related with increased risk of death. Further studies are needed to evaluate pathogenic mechanisms of SARS-CoV-2 and the effect of the several proposed therapeutic approaches in reducing its lethality. Moreover, long term post discharge follow-up is warranted to provide a more accurate estimate of the morbidity and mortality attributable to this infection.

Declaration of Competing Interest

The authors declare that there are no conflicts of interest.

Authors contribution

AG, ALR and MG designed the study. LO, AG were responsible for the statistical analysis. All authors contributed in the patient’s enrolment, data collection and interpretation. AG and ALR drawn a preliminary draft of the manuscript. MG, SA, SR, LM critically revised the manuscript. All authors approved the final version of the manuscript.