Dementia and psychotropic medications are associated with significantly higher mortality in geriatric patients hospitalized with COVID-19: data from the StockholmGeroCovid project

Table 2 Risk of death during the whole follow-up in relation to dementia in patients hospitalized with COVID-19

			Patients with COVID-19 (n=5122)—hazard ratios (95% CI)
			Unadjusted model	Model 1	Model 2
Dementia			1.83 (1.62–2.06)**	1.67 (1.47–1.90)**	1.59 (1.40–1.81)**
Age group	<70		0.73 (0.54–0.99)*	0.74 (0.55–1.01)	0.75 (0.56–1.02)
	70–79		Reference	Reference	Reference
	80–89		1.84 (1.61–2.12)**	1.83 (1.59–2.10)**	1.83 (1.58–2.10)**
	>89		2.81 (2.43–3.24)**	3.02 (2.61–3.50)**	2.90 (2.49–3.37)**
Male sex	Baseline effect		1.33 (1.18–1.51)**	1.35 (1.21–1.49)**	1.55 (1.37–1.76)**
Male sex	Male × time^a		0.99 (0.99–0.99)*	1.35 (1.21–1.49)**	0.99 (0.99–0.99)*
CCI group	0–1		Reference	Reference	Reference
	2–3		1.47 (1.31–1.64)**	1.50 (1.34–1.68)**	1.52 (1.35–1.71)**
	>3		2.31 (2.00–2.66)**	2.50 (2.17–2.90)**	2.56 (2.21–2.97)**
SpO₂	Baseline effect	≥90%	Reference		Reference
	Baseline effect	<90%	3.14 (2.60–3.81)**		3.00 (2.47–3.63)**
	SpO2 × time^a		0.99 (0.99–0.99)**		0.99 (0.99–0.99)**
Beta blockers			1.28 (1.15–1.41)**
Ca2+ channel blocker			0.81 (0.73–0.91)**
RAAS inhibitors			0.86 (0.78–0.95)*
Statins			0.71 (0.63–0.78)**		0.73 (0.65–0.81)**
Antithrombotics			1.13 (0.86–1.48)
Antipsychotics	Baseline effect		2.14 (1.82–2.51)**	1.92 (1.67–2.20)**	1.70 (1.47–1.97)**
Antipsychotics	Antipsychotics × time^a		0.99 (0.99–0.99)**	1.92 (1.67–2.20)**	1.70 (1.47–1.97)**
Antidepressants			1.16 (1.04–1.29)*
Anxiolytics			1.37 (1.23–1.52)**		1.18 (1.05–1.32)*
Hypnotics/sedatives			1.35 (1.22–1.49)**		1.15 (1.04–1.28)*

^aSex, SpO₂, and antipsychotic use violated the proportionality of hazards assumption in some models—presented as interaction mode where appropriate. Interaction with time represents the incremental change in the probability of death with advancing days of survival from the baseline probability; *p-value <0.05; **p-value <0.001. In the unadjusted model, the variables were entered into the equation separately; model 1 was adjusted for dementia, antipsychotics, age, sex, and CCI categories, while model 2 was further adjusted for SpO2 at admission and prescription of statins, hypnotics, and sedatives at hospital admission or day after admission. Results are based on the Cox regression models

ISSN: 1758-9193