Sociodemographic and clinical characteristics associated with delirium in hospitalized patients with COVID-19: are immigrants a vulnerable group?

Dear Editor,

Patients with Coronavirus Disease (COVID-19) who require hospitalization could be at higher risk of neuropsychiatric sequelae than individuals not requiring inpatient admission [1]. Among different neuropsychiatric manifestations, it has been reported a considerable rate of delirium among inpatients with COVID-19 [2, 3]. Indeed, the prevalence of delirium in COVID-19 has been estimated to be up to 80% in intensive care units (ICU) and around 20–30% among non-ICU hospitalized patients [4, 5]. Delirium is interpreted as an indicator of systemic critical illness and it is associated with high risk of mortality in subjects affected by COVID-19 [6]. Regarding this aspect, preliminary data suggested that immigrants and ethnic minorities may experience a more severe course of COVID-19 in comparison with the rest of population [7]. However, although confusional state was identified as a predictor of poor clinical course, contrasting results were reported in literature about identification, associated risk factors and clinical impact of delirium in COVID-19 patients.

In the light of the frequent occurrence of delirium in patients hospitalized for COVID-19, the current study was retrospectively designed to investigate the occurrence of delirium among non-ICU inpatients with COVID-19, its impact on in-hospital stay and whether some socio-demographic, clinical and biochemical variables may increase the likelihood of this condition.

We conducted a cross-sectional study, including inpatients consecutively admitted to the isolated COVID-19 wards at two hospitals—Desio and San Gerardo—(Monza, Lombardy region) [8] during the peak of COVID-19 outbreak, between 4th March and 6th May 2020. All patients were admitted to the hospital because of COVID-19 related pneumonia. Nevertheless, the COVID-19 cases were defined as subjects who had a positive transcriptase-polymerase chain reaction (PCR) test. We excluded patients from other hospitals or referred to another clinic, admitted to ICU during the hospital stay, with pre-existing substance use, mental or neurological disorders or taking any psychotropic medication before hospitalization. Patients with any missing data were ruled out, including those for whom was impossible to identify the presence of delirium. Information on socio-demographic characteristics, pre-existing comorbidities, routine blood tests and length of hospitalization were retrieved from medical records. Pre-existing lung diseases included chronic obstructive pulmonary disease, asthma, bronchiectasis, sarcoidosis, extrinsic allergic alveolitis, cystic fibrosis, idiopathic pulmonary fibrosis, other interstitial lung diseases, and lung cancer. Laboratory tests were performed within 24 h after hospitalization and included complete blood cell count, neutrophil to lymphocyte ratio (NLR), international normalized ratio (INR), lactate dehydrogenase (LDH) and C-reactive protein (CRP).

The presence of delirium during stay was carefully checked after review of daily reports on clinical charts and was identified in case of fulfillment of criteria of the Confusion Assessment Method (CAM) tool, shortened version [9]. On the other hand, the presence of anxiety or agitation alone, but not accompanied by alterations of consciousness or thinking, was not defined as delirium.

Descriptive analyses were originally performed for the whole sample. Then, we carried out univariate analysis to detect statistically significant differences between subjects developing or not developing delirium during hospital stay. The normal distribution of quantitative variables was verified using Shapiro–Wilk's test. The groups were compared for qualitative variables by chi-square tests and for continuous variables by Student’s t or Wilcoxon-Mann–Whitney tests as appropriate. Finally, all the variables from the univariate analysis resulting in a p < 0.05, together with age, gender and length of hospitalization, were inserted as independent variables into a logistic regression model with the occurrence of delirium as dependent variable. Statistical significance was set at p < 0.05. Analyses were conducted using Stata Version 13.1 SE.

Data were available for 221 individuals. Fifty-four of them (24.4%) developed delirium during hospital stay. Most patients were male, of Italian origin with a mean age of 67.4 years. Hypertension was the most common comorbidity (52.5%). The average number of days spent in hospital was 15.9. A comparison of the main baseline clinical variables between individuals with versus without development of delirium is reported in Table 1.

Table 1.

Demographic and clinical characteristics of patients with COVID-19 at hospital admission, categorized according to the development of delirium during hospital stay

Variables	Total sample N = 221	Patients without delirium N = 167 (75.6%)	Patients with delirium N = 54 (24.4%)	p-value
Sociodemographic
Age (years) mean (SD)	67.4 (14.5)	66.2 (11.1)	71.4 (14.5)	0.023a
Female gender	58 (26.2%)	46 (27.6%)	12 (22.2%)	0.440c
Non-Italian nationality	25 (11.3%)	12 (7.2%)	13 (24.1%)	0.001c
Existing comorbidities
High blood pressure	116 (52.5%)	83 (49.7)	33 (61.1)	0.144c
Diabetes	48 (21.7%)	37 (22.2)	11 (20.4)	0.782c
Cardiovascular disease	96 (43.4%)	67 (40.1)	29 (53.7)	0.080c
Cancer	31 (14.0%)	19 (11.4)	12 (22.2)	0.046c
Liver disease	17 (7.7%)	11 (6.6)	6 (11.1)	0.210d
Lung disease	50 (22.6%)	32 (19.2)	18 (33.3)	0.030c
Other disease	98 (44.3%)	70(41.9)	28 (51.8)	0.201c
Biochemical factors mean (SD)
Red blood cells (106 /μl)	4.7 (0.7)	4.7 (0.7)	4.4 (0.7)	0.006a
White blood cells (109 /L)	7.9 (4.0)	7.8 (3.4)	8.1 (5.4)	0.457b
Platelets (109 /L)	215.0 (83.9)	222.0 (86.2)	193.4 (72.9)	0.023b
Neutrophils (109 /L)	6.2 (3.8)	5.9 (3.1)	6.9 (5.4)	0.916b
Lymphocytes (109 /L)	1.3 (1.9)	1.2 (1.2)	1.4 (3.2)	0.518b
Monocytes (109 /L)	0.5 (0.3)	0.5 (0.3)	0.4 (0.3)	0.251a
Neutrophil-to-Lymphocyte ratio (%)	8.2 (12.7)	6.9 (6.4)	12.3 (22.7)	0.005b
International normalized ratio (%)	1.2 (0.5)	1.2 (0.4)	1.3 (0.6)	0.008b
Lactate dehydrogenase (U/L)	392.0 (196.6)	365.5 (133.5)	474.0 (309.1)	0.003b
C-reactive protein (mg/dL)	101.1 (78.2)	97.0 (71.8)	113.6 (95.0)	0.407b
Hospital stay (days) mean (SD)	15.9 (11.2)	15.1 (9.1)	18.1 (16.1)	0.674b

dL deciliter, L liter, μl microliter, mg milligram, SD standard deviation, U unit

Values are numbers (%), unless otherwise specified

a. Student’s t test; b. Wilcoxon-Mann–Whitney test; c. Pearson's χ² test;  d. Fisher’s exact test

Statistically significant findings are reported in bold

Patients with delirium resulted to be older, more frequently of non-Italian origin and had higher rate of both cancer and lung diseases. Significant correlations were also found for red cell count, platelet count, NLR, INR and LDH. No statistical differences were observed between the two groups with regard to duration of hospitalization. Finally, after controlling for age, gender and duration of hospitalization, being of non-Italian origin (aOR = 7.89, p = 0.001) and having lung disease (aOR = 2.26, p = 0.043) was positively associated with the development of delirium. Baseline levels of LDH (aOR = 1.00 p = 0.035) were also associated with occurrence of delirium. An inverse association between the occurrence of delirium and platelet count (aOR = 0.99, p = 0.059) at borderline statistical significance was found. Nevertheless, a diagnosis of cancer (aOR = 2.50, p = 0.066), higher levels of NLR (aOR = 1.04, p = 0.084) and a more advanced age (aOR = 1.03, p = 0.088) were all factors showing a trend of increased risk of delirium (Table 2).

Table 2.

Results of logistic regression analysis, exploring the factors associated with delirium in 221 patients hospitalized for COVID-19 pneumonia

	Variables	aOR	95% CI	p-value
Sociodemographic
	Age (years)	1.03	1.00–1.06	0.088
	Female gender	1.27	0.53–3.00	0.591
	Non-Italian nationality	7.89	2.41–25.79	0.001
Existing comorbidities
	Cancer	2.50	0.94–6.61	0.066
	Lung disease	2.26	1.03–4.99	0.043
Biochemical factors
	Red blood cells (106 /μl)	0.74	0.41–1.32	0.304
	Platelets (109 /L)	0.99	0.99–1.00	0.059
	Neutrophil-to-Lymphocyte ratio	1.04	0.99–1.09	0.084
	International normalized ratio	1.08	0.51–2.28	0.842
	Lactate dehydrogenase (U/L)	1.00	1.00–1.00	0.035
Hospital stay (days)
	Hospital stay	1.02	0.99–1.06	0.208

L  liter, μl microlitre, SD standard deviation, U unit

^aOR adjusted odds ratios and their 95% confidence interval (CI)

Statistically significant findings are reported in bold

According to our findings, almost one-fourth of COVID-19 inpatients without any apparent neuropsychiatric history developed delirium. Even though comparisons with available data in literature may be difficult for the variability in study designs, inclusion criteria, definition of delirium, the results of the present research confirm that delirium is a common complication of SARS-COV2, as previously reported by other studies [10, 11].

Regarding vulnerability factors, we found that non being of Italian origin and a history of lung disease was associated with an increased risk of delirium occurrence. Some previous investigations reported that incidence and mortality rates of COVID-19 in Europe were higher in immigrants than in natives [7]. This is the first Italian study showing immigrants as a group associated with a more unfavorable course of COVID-19, probably as a result of socioeconomic frailty. In particular, limited home space, unemployment, low income, language barriers, limited health literacy, genetic and biological susceptibilities to specific diseases may explain higher COVID-19 mortality among immigrants compared to natives in Italy [12]. Noteworthy, age was significantly lower in non-Italian (55.3 years) than in Italian subjects (69.0 years) (p < 0.001). Moreover, presence of concomitant lung disease was found to predict the onset of delirium. Different studies hypothesized that patients with a chronic lung disease, particularly those with chronic obstructive pulmonary disease, may have a predisposition to suffer from severe forms of COVID-19. In particular, biological pathways involving the expression of angiotensin-converting enzyme 2 in bronchial epithelium (especially in smokers) as well as in different cells of central nervous system may explain the considerable rate of delirium among patients with lung diseases [13]. Nevertheless, other predisposing factors, including older age and a history of cancer, were found to be associated with delirium onset in other studies [5]. With regard to the association of delirium with LDH serum levels and the platelet count, some authors argued that higher levels of LDH and low platelet count may reflect more severe forms of COVID-19. Indeed, thrombocytopenia and elevated LDH levels reflect a hypercoagulable state, in turn related to severe illness and high mortality. Overall, reduced platelet counts as well as higher levels of LDH may be common in COVID-19 patients with neuropsychiatric manifestations [14]. Nevertheless, the etiology of neuropsychiatric symptoms is probable multi-factorial. In this regard, patients with delirium were found to have higher NLR levels, thus supporting the hypothesis that inflammation could trigger delirium in COVID-19 patients [11]. In contrast with most studies, the occurrence of delirium was not found to be significantly associated with a longer duration of hospitalization. This finding may be explained by the fact that our sample excluded patients admitted to ICU.

Different limitations of our study should be mentioned. First, some data were not recorded. However, the retrospective design of the study as well as the emergency context of patient care during the first peak of COVID-19 pandemic prevented us to collect complete data regarding delirium subtypes (hyperactive, hypoactive, and mixed), timing of onset, other biochemical parameters and prescription of pharmacological treatments. However, it is important to underline that all these variables as well as the PaO2/FIO2 ratio and physical frailty may be associated with disease severity and poor prognosis in patients with COVID-19 [15]. Furthermore, most of medical records lacked of comprehensive information. In this regard, we cannot entirely exclude that economic status or deprivation may be a confounding variable, particularly in the association between being of non-Italian origin and occurrence of delirium. Second, we analyzed variables associated with delirium in patients admitted to a general ward while they may be different in another context, as the ICU. Finally, the relatively low number of cases and the impossibility of a blinded evaluation of delirium were important limitations. Taken together, all these factors may limit the generalizability of our findings.

In spite of these limitations, we confirm that delirium is a common complication in patients hospitalized for COVID-19 and represents a multi-factorial syndrome caused by a complex interrelationship between predisposing factors and precipitating factors. Particularly, immigrants may represent a group more vulnerable to severe manifestations of COVID-19. Nevertheless, some clinical and biological markers identified in this study may be useful to stratify the risk of developing delirium at admission, thus improving the prognosis of patients with severe forms of COVID-19.

Funding

This research received no external funding.

Declarations

Conflict of Interest

The authors declare no conflict of interest.

Human and animal rights statement

The research was conducted in accordance with the principles embodied in the Declaration of Helsinki and in accordance with local statutory requirements.

Informed consent

All participants gave written informed consent to participate in the study.