Impact of new‐onset and preexisting neurological disorders in COVID‐19 patients

Abstract Background and Purpose Coronavirus disease (COVID‐19) is still considered a global pandemic. The prognosis of COVID‐19 patients varies greatly. We aimed to assess the impact of preexisting, chronic neurological diseases (CNDs) and new‐onset acute neurological complications (ANCs) on the disease course, its complications, and outcomes. Methods We conducted a monocentric retrospective analysis from all hospitalized COVID‐19 patients between May 1, 2020 and January 31, 2021. Employing multivariable logistic regression models, we explored the association of CNDs and ANCs separately with hospital mortality and functional outcome. Results A total of 250 among 709 patients with COVID‐19 had CNDs. We found a 2.0 times higher chance of death (95% confidence interval [CI]: 1.37–2.92) for CND patients than for non‐CND patients. The chance for an unfavorable functional outcome (modified Rankin Scale > 3 at discharge) was 1.67 times higher in patients with CNDs than those without (95% CI: 1.07–2.59). Furthermore, 117 of all patients had 135 ANCs in total. We observed a 1.86 times higher chance to die (95% CI: 1.18–2.93) for patients with ANCs than without. The chance for a worse functional outcome was 3.6‐fold higher in ANC patients than without (95% CI: 2.22–6.01). Patients with CNDs had 1.73 times higher odds for developing ANCs (95% CI: 0.97–3.08). Conclusion Preexisting neurologic disorders or ANCs in COVID‐19 patients were associated with higher mortality and poorer functional outcome at discharge. Furthermore, development of acute neurologic complications was more frequent in patients with preexisting neurologic disease. Early neurological evaluation appears to be an important prognostic factor in COVID‐19 patients.


INTRODUCTION
At the end of 2019, a highly transmissible novel coronavirus, designated as SARS-CoV2 (severe acute respiratory syndrome corona virus type 2), was discovered to be the pathogen causing the new worldwide pandemic, subsequently named COVID-19 (coronavirus disease 2019) (Hu et al., 2021).
The clinical course of this infection is highly variable, ranging from being asymptomatic or with mild respiratory symptoms to multiple organ failure and even death (Al-Tawfiq et al., 2020). Therefore, mortality rates among COVID-19 patients vary greatly (Zhou et al., 2020).
Researchers worldwide attempt to identify possible risk factors for a severe course of infection. This is considered important for predicting the hospitalization as well as the intensive care unit (ICU) admission need. Age, male sex, cardiovascular risk factors such as hypertension, diabetes, and obesity as well as chronic respiratory disease were identified as risk factors for severe/fatal COVID-19 (Zhou et al., 2020). Recent studies showed an association between preexisting cerebrovascular diseases and a severe disease course as well as higher mortality in COVID-19 patients (Aggarwal et al., 2020;Pranata et al., 2020).
Moreover, COVID-19 was shown to be associated with various acute neurological complications (ANCs). Multiple studies described COVID-19-associated cerebrovascular complications and inflammatory syndromes of the central and peripheral nervous systems as well as encephalopathy and miscellaneous disorders (Leven & Bösel, 2021).
In this study, we aimed to investigate the role of preexisting chronic neurological diseases (CNDs) as a potential risk factor for a more severe course or fatal outcome of SARS-CoV2 infection. In addition, we aimed to detect the association of such preexisting diseases with ANCs and their impact on the clinical course of COVID-19 disease.

Patient selection and characteristics
All patients older than 18 years admitted to a regional German secondary hospital (SRH Waldklinikum Gera) between May 1, 2020 and January 31, 2021 with a positive polymerase chain reaction (PCR) test for SARS-CoV2 from a combined nose and throat swab were included in this monocentric retrospective analysis. Patients in psychiatry were excluded for medical records not being available. Patients admitted to hospital for another reason than SARS-CoV2 only were not included if they developed no symptoms and were soon released. Patients who were infected in hospital or developed symptoms on a later date were included.
The study protocol was approved by the

Outcomes
For each COVID-19 patient, data on various complications were collected. This included nonneurologic complications, such as acute respiratory distress syndrome (ARDS) and sepsis (for complete list, see Tables 2 and 4 Outcome variables for regression analyses were survival and the dichotomized mRS at discharge. A favorable outcome was defined as mRS 0-3 at discharge, varying from asymptomatic to moderate disability requiring partial assistance but able to walk unassisted.

Statistical analysis
Descriptive statistics (median, interquartile range, and absolute and relative frequencies) were applied to compare groups. The rate of missing data was 0.07%, so imputations techniques for missing data were not applied. For pairwise comparisons, we used hypothesis TA B L E 1 Demographic data of all patients divided into those with and without preexisting chronic neurological disorders.

Study population and clinical characteristics
Tables 1 and 2 show data of the study population. In total, 709 patients  also Table S3).
Compared to non-CND patients, those with CNDs were significantly older (median age: 82 vs. 74 years; p < .001) and had a higher prehospitalization mRS (3 vs. 0; p < .001) ( Table 1). Table 2 shows that comorbidities such as hypertension (p < .001) and renal insufficiency (p < .001) were significantly more frequent in patients with CNDs in comparison to those without. Patients with CNDs developed ANCs more often than non-CND patients (21.6% vs.

Comorbidities and complications
13.7%; p = .008). The most common ANCs in CND patients were delirium with known dementia (10.4%; n = 26) and encephalopathy (6.8%; n = 31). Less frequent were epileptic seizures (n = 8), ischemic TA B L E 2 Comorbidities, complications, and hospital outcomes of all patients divided into those with and without preexisting chronic neurological disorders. Note: Descriptive data are given as median (IQR-interquartile range) or absolute numbers (%). Corresponding odds ratios (OR) and confidence intervals are presented, where LCI means lower boundary and UCI means upper boundary of a confidence interval. The significance level has been individually adapted to the number of pairwise comparisons done in the related subgroups (α adj ). Remarkable ratios and p-values with respect to α adj are in bold numbers. Abbreviations: ARDS, acute respiratory defiance syndrome; CND, chronic neurological disease; COPD, chronic obstructive pulmonary disease; ICU, intensive care unit; NIV, noninvasive ventilation; non-CND, no chronic neurological disease; NSTEMI, non-ST segment elevation myocardial infarction; PAOD, peripheral artery occlusive disease; STEMI, ST segment elevation myocardial infarction.

Prediction models
In the first multiple logistic regression analysis employing survival as primary endpoint, we found CNDs, male sex, renal failure as well as higher age as independent significant predictors of hospital mortality ( Figure 2; Table S4).

TA B L E 3
Demographic data of patients divided into those with and without acute neurological complications.

ANC Non-ANC p-value
Demographic data α adj = .0071 Number ( Length of stay in ICU-days (IQR) 9.5 (4-24) 4 (2-8) <.001 Duration of mechanical ventilation-days (IQR) 12.5 (7-17) 8 (4-11) <.001 Length of hospital stay-days (IQR) 17 (11-26) 9 (6-15) <.001 mRS-discharge 4 3 <.001 Note: Descriptive data are given as median (IQR-interquartile range) or absolute numbers (%). The significance level has been adapted to the number of pairwise comparisons done in this subanalysis (α adj ). Remarkable p-values with respect to α adj are in bold numbers. Modified Rankin scale (mRS) is defined as mRS 0-no symptoms; 1-symptoms without disability; 2-unable to carry out all previous activities; 3-moderate disability, still able to walk; 4-unable to walk unassisted; 5-bedridden; and 6-dead. Abbreviations: ANC, acute neurological complications; ICU, intensive care unit; non-ANC, no acute neurological complication. Figure S2 and Table S5 show the details of the regression analysis for our secondary endpoint, the dichotomized functional outcome. The abovementioned significant predictors for survival were also significant in this regression analysis, except for severe renal insufficiency. In addition, we observed a significantly worse outcome for patients with hypertension and a baseline mRS higher than 3.

Patients with new-onset ANCs
Of all patients, 117 patients developed a total of 135 ANCs. Tables 3   and 4 show the demographic data, comorbidities, complications, and hospital outcomes for the ANC patients. These had a significantly higher rate of cardiovascular risk factors, such as atrial fibrillation (p < .001) and hypertension (p = .001) compared to non-ANC patients. Among ANCs, cerebrovascular events such as ischemic stroke occurred in 1.3% (n = 9), intracerebral hemorrhage in 0.8% (n = 6), TIA in 0.6% (n = 4), and subarachnoid hemorrhage in 0.1% (n = 1) and one patient showed a hypoxic brain damage after cardiopulmonary resuscitation. Encephalopathic complications occurred in 11.8% (n = 84) patients including 5.5% (n = 39) with encephalopathy, 3.7% (n = 26) with delirium with dementia, 2.5% (n = 18) with delirium without dementia, and 0.1% (n = 1) with alcohol withdrawal with delirium tremens. Epileptic seizures occurred in 1% (n = 7) with the manifestation of a status epilepticus in one of them. One patient had myoclonia.
Patients with ANCs had a higher rate of ICU admission (51.3% vs. 18.4%; p < .001). They received more often intubation (p < .001) and NIV (p < .001). Duration of mechanical ventilation was also significantly longer in patients with ANCs (p < .001). ANC patients had in total a significantly longer ICU treatment in comparison to those without ANCs (9.5 vs. 4 days; p < .001).
In tendency, more ANC patients (n = 38; 32.5%) had a strictly conservative therapy or de-escalation of therapy (comfort care) (p = .012) in comparison to non-ANC patients (n = 126; 21.3%), although the difference was not significant.

Prediction models
In the first multiple logistic regression analysis employing survival as primary endpoint, we found ANCs, male sex, moderate COPD, renal insufficiency as well as higher age as independent significant predictors of hospital mortality (Figure 4; Table S6).
ANCs increased the odds for death 1.9-fold (p = .007; OR = 1.9; Note: Descriptive data are given as median (IQR-interquartile range) or n (%). Corresponding odds ratios (OR) and confidence intervals are presented, where LCI means lower boundary and UCI means upper boundary of a confidence interval. The significance level has been individually adapted to the number of pairwise comparisons done in the related subgroups (α adj ). Remarkable ratios and p-values are in bold numbers. In 117 individual patients, 135 acute neurological complications (ANCs) occurred. For our comparison, we only refer to independent data; multiple ANCs are not further considered. Abbreviations: ANC, acute neurological complications; ARDS, acute respiratory defiance syndrome; COPD, chronic obstructive pulmonary disease; ICU, intensive care unit; NIV, noninvasive ventilation; non-ANC, no acute neurological complication; NSTEMI, non-ST segment elevation myocardial infarction; PAOD, peripheral artery occlusive disease; STEMI, ST segment elevation myocardial infarction.
increased the odds for death by the factor of 1.046 (p < .001; 95% CI: 1.025-1.068). Figure S3 and Table S7 show the details of this regression.

DISCUSSION
In this study, we could show that CNDs in hospitalized COVID-19 patients were associated with a higher mortality and a worse func-tional outcome; however, they were less frequently admitted to the ICU.
Patients with CNDs such as dementia, Parkinson's disease, history of stroke, myasthenia, and multiple sclerosis showed a strong association with a fatal or more severe course of COVID-19. Similar results were demonstrated in previous studies (Liotta et al., 2020). We further found that these patients had a higher mRS at admission and were also significantly older. This might explain the lower rate of admission to   Kleineberg et al., 2021;Pranata et al., 2020;Zhou et al., 2020). In line with Garcia-Azorin et al. (2020), preexisting neurological disorder was a predictor for death in COVID-19. A higher mRS (>3) at admission was also a predictor of mortality. However, most studies did concentrate on the presence of cardiovascular risk factors and not CNDs in general. Moreover, patients with CNDs were discharged with a worse mRS. Therefore, we assume that neurologic deterioration in the course of COVID-19 infection is to be expected in these patients.
COVID-19 patients with CNDs are shown to have higher chance to receive a strictly conservative therapy or de-escalation of therapy (comfort care). It remains unclear whether these patients would have survived COVID-19 if they received a full medical treatment including invasive ventilation. We need to further address this question in order to facilitate the decision-making in this patient population regarding offering them invasive or extended therapeutic options or de-escalation.
In this large-scale cohort study of hospitalized COVID-19 patients, ANCs were detected in 19% of our study population and were associated with higher rates of other severe medical complications, such as sepsis, higher mortality, and significantly worse functional outcome at discharge compared to those without ANCs. The most common neurological complications were encephalopathy, delirium, and ischemic stroke. The incidence of ANCs was 16.5% corresponding to 135 distinct disorders in 117 of 709 patients. This is in line with previous findings in the literature (Frontera, Sabadia, et al., 2021;Maury et al., 2021;Nannoni et al., 2021;Rifino et al., 2021;Travi et al., 2021). Among these complications, ischemic stroke was observed in 1.3% and intracerebral hemorrhage in 0.84% of the patients. This is also consistent with previous publications (Ellul et al., 2020;Frontera, Sabadia, et al., 2021;Rifino et al., 2021). However, we could not determine if the strokes were due to COVID-19 and the higher coagulopathy or if they were coincidently. Previous studies showed limitations in that regard (Tsivgoulis et al., 2020).
Earlier studies reported a higher rate of hemorrhagic stroke in COVID-19 patients associated with extracorporeal membrane oxygenation therapy (ECMO) (Kleineberg et al., 2021). In our cohort, only six patients received ECMO, in which only one of them developed intracerebral hemorrhage. Another study (Schmidbauer et al., 2022) reported a prevalence of intracranial hemorrhages in 0.85% of critically ill patients with COVID-19. In their cohort, 20.8% of ICU patients underwent ECMO therapy. The reported prevalence of encephalopathy varies in different studies between 7% (Ellul et al., 2020) and 31% (Liotta et al., 2020). One study examining COVID-19 patients in ICU reported delirium in 84.3% (Helms et al., 2020); however, our results showed 11.8%. One possible reason for these discrepancies might be the lack of reporting of altered mental state by doctors without a neurological background and differences in preferred definitions and established scoring procedures.
Moreover, we also found a significant association between the development of ANCs and increased mortality rate in  patients. Others also demonstrated a similarly worse mRS at discharge in conjunction with a significantly longer stay in hospital and especially in ICU in ANC patients . Therefore, ANC seems to be an important risk factor for a delayed recovery in COVID-19 patients.
In contrast to our findings, others reported a lower mortality rate in patients with isolated neurological symptoms (Travi et al., 2021).
However, it is difficult to compare the prevalence of neurological complications between studies due to varying definitions. Some studies included dysgeusia and anosmia as well as headache and syncope as neurological complications (Travi et al., 2021), whereas others included less specific symptoms as myalgia and dizziness (Liotta et al., 2020).
We deliberately only included severe complications with the necessity of further neurological assessment such as stroke, epileptic complications, and neuropathy. This might explain the higher mortality rate and less favorable outcome in our study.
Based on our results and literature findings, a thorough neurological evaluation for all hospitalized patients with SARS-CoV2 infection is recommended. Preexisting neurological diseases or the development of ANCs might indicate a significantly higher risk for a poor or even fatal outcome. This entails not only short-lasting goals like hospital mortality, but also long-lasting disabilities and the necessity for rehabilitation, although this was not addressed in our analyses. The need for prophylactic measures, including immunization in CND patients and advanced care planning, should be strongly encouraged on the basis of our study.

Limitations
This was a monocentric retrospective study. The number of ANCs may be underrepresented because imaging was not regularly performed in patients who did not show obvious focal neurological deficits. Furthermore, patients in our sample with preexisting neurological disorders were significantly older than the non-CND group and had higher prevalence in cardiovascular risk factors. This might have led to a confounding bias due to co-existing age-related diseases. Moreover, we did not adapt the type I error for the multiple testing procedure; hence, some of the significant results may be false positive findings. This study could not consider the effects of the type of COVID-19 variants nor the effect of vaccination against the virus on mortality or functional outcome due to the dataset period.

CONCLUSION
COVID-19 in patients with preexisting neurological conditions is associated with higher mortality and survival with worse functional outcomes at discharge, as well as a lower rate of ICU admission. ANCs are more frequent in patients with preexisting neurological conditions in COVID-19 patients. They were also associated with a significantly worse neurological outcome, higher rate of severe complications such as sepsis and ARDS, and thus higher mortality. Early neurological evaluation appears of great importance for prognostication in the treatment course of COVID-19 infection.

ACKNOWLEDGMENTS
Open access funding enabled and organized by Projekt DEAL.

CONFLICT OF INTEREST STATEMENT
Albrecht Günther received speaker's honoraria from Boehringer Ingelheim, Daichii Sankyo, Pfizer, Occlutech, and Ipsen as well as research grants from MERZ Pharma and IPSEN. The other authors declare no conflicts of interest.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.