INTRODUCTION

Since the beginning of the COVID-19 pandemic, emerging evidence from the US and the UK has revealed significant disparities between people belonging to ethnic/racial minorities and those of White race, with the first being at higher risk of infection, severe disease and adverse outcomes^1-4. Additionally, several studies have shown that COVID-19 affects disproportionally the migrant population^4,5, although relevant research is still limited. So far, the impact of SARS-CoV-2 infection has not been investigated in people belonging to ethnic/racial minorities in Greece. Thus, we aimed to explore the outcomes of COVID-19 patients hospitalized in a large Greek General Hospital according to their racial/ethnic origin. We hypothesized that patients of European origin would present lower mortality rates compared to those of African/Asian origin.

METHODS

This retrospective study included 628 patients admitted in the COVID-19-dedicated unit (common isolation wards) of Evangelismos General Hospital, between 10 September and 31 December 2020. Data were retrieved from patients’ medical records. Patients were classified as being either of European origin or African/Asian origin. The reasons why we divided our population into these two groups are described below: Greece in the early 1990s and 2000s became a migrant-hosting country for mainly other Balkan countries, like Albania, Romania and Bulgaria, and some European states of the former Soviet Union⁶. The most recent migration flows to Greece have been mainly from the Middle East, central Asia and different parts of Africa^7,8. As immigrants’ health tends to converge with the local population over time⁹, our study focuses on the comparison of two groups, Europeans (either born in Greece or other countries) and non-Europeans, which in our case are patients of African/Asian origin and who represent the vast majority of the recent immigrant flows. Patients were treated according to the national and international guidelines which were progressively developing during the period of the study.

We compared data concerning gender, age, comorbidities and outcome between patients of European and non-European origin. Continuous variables are presented as medians with IQR, and were evaluated with Mann-Whitney U-test, while contingent analysis was performed using Fisher’s exact test. Moreover, we performed a univariate analysis with the race being the independent variable and in-hospital mortality being the dependent. In-hospital mortality was assessed against race using a multiple logistic regression model including gender and the Charlson Comorbidity Index (CCI) score, as additional co-variates. All analyses used the SPSS version 23.0 (IBM SPSS, IBM Corp., Armonk, NY, USA).

RESULTS

A total of 628 hospitalized patients, with a positive QT-RT-PCR nasopharyngeal swab test for SARS-CoV-2, with a median (IQR) age of 59 years (range: 28–73), were included in the study (Table 1). Among them, 84% were of European origin. Compared with Europeans, non-Europeans were younger (p<0.001), with a greater proportion of males (p<0.001) and never smokers (p=0.001). In terms of comorbidities, fewer non-Europeans were obese (p<0.001), had hypertension (p=0.001) and cancer (p=0.041), but more were suffering from chronic liver disease (p=0.021). In addition, the median CCI score was 2 (range: 1–4) for Europeans, which was significantly higher (p<0.001) compared with non-Europeans, which was 1 (range: 0–1.75). On admission, Europeans had more advanced disease, as defined by the 2022 WHO criteria¹⁰, than non-Europeans. Compared with non-Europeans, more Europeans received remdesivir (38.3% vs 20%, p<0.001) or corticosteroids (51.5% vs 26%, p<0.001), which more likely reflect differences in disease severity. By the end of the study, immune-modulating agents were not recommended as routine treatment for severe and critical disease. Therefore, only two patients of European origin were given tocilizumab. Compared to Europeans, a lower ICU admission (p=0.001), intubation (p=0.007) and mortality (p<0.001) rate was observed in the non-Europeans.

In the unadjusted race-only logistic regression model, non-Europeans were less likely than European patients to die in the hospital (OR=0.057; 95% CI: 0.008–0.411, p=0.005). However, using a logistic regression analysis adjusted for sex (OR=0.533; 95% CI: 0.297–0.956, p=0.035), age (OR=1.068; 95% CI: 1.039–1.097, p<0.001) and CCI score (OR=1.328; 95% CI: 1.040–1.098, p<0.001), we could not find significant differences in the death rate between Europeans and patients of other racial/ethnic groups (OR=0.492; 95% CI: 0.062– 3.930, p=0.503).

DISCUSSION

This study examined the clinical characteristics and outcomes of 628 hospitalized COVID-19 patients, according to their ethnic/racial origin. Europeans were older with a larger proportion of female patients, compared to non-Europeans. Additionally, they had a heavier burden of comorbidities (except chronic liver disease which was more common among the African/Asian) and on admission, they had more advanced disease. More Europeans than non-Europeans received remdesivir and dexamethasone, which more likely reflects the differences in disease severity between the two groups. Non-Europeans presented a lower in-hospital mortality rate, but after adjusting for age, gender and CCI score, the ethnic origin did not seem to be associated with the patients’ outcome.

To the authors’ best knowledge, this is the first study to assess ethnic/racial differences in the clinical characteristics and outcomes of hospitalized COVID-19 patients, in a Greek General Hospital, demonstrating that patients of African/Asian origin had excellent outcomes. This observation may be explained by the well-documented ‘healthy immigrant effect’¹¹; immigrants often report better health status compared to local individuals, as they are young and healthy on arrival. In general, non-Europeans were younger with a better co-morbidity profile. Thus, a more meaningful comparison had to eliminate the possibility that the differences in the COVID-19 outcomes were a result of the differences in the population composition, since older age and comorbidities have been linked to a higher COVID-19 mortality¹². Evaluation of age, gender and CCI as co-variates revealed no significant differences in mortality. Similarly, while early observations in the US showed increased mortality among African-Americans or Hispanics, compared to Caucasians, this observation was later attributed to increased infection rate rather than worse disease outcomes^13,14. In agreement with our findings, others have reported no effect of being of ethnic/racial origin in hospitalized COVID-19-patient outcomes, when accounting for base-line patients’ characteristics^14-17.

CONCLUSIONS

Asian and African immigrants with SARS-CoV-2 infection, hospitalized in a General Greek hospital during autumn, had similar mortality rates as patients of European origin (local or immigrant). Older age, male gender and higher Charlson Comorbidity Index were independent risk factors of in-hospital mortality.