Neurological and neuropsychiatric manifestations in hospitalised patients with COVID-19

Background Although literature globally indicates varied neurological and/or neuropsychiatric manifestations (NNM) and complications associated with coronavirus disease 2019 (COVID-19), information about NNM in infected hospitalised patients on the African continent remains limited. Aim To describe the presentation of NNM and compare patients with and without NNM considering demographic and clinical profiles, treatment, and outcomes. Setting Tygerberg Hospital, Cape Town, South Africa. Methods Retrospective medical record review of the first 100 consecutively admitted COVID-19 patients (64 females, mean age 47.6 years) between March and June 2020. Results Of the 98 patients included in the analysis, 56.1% had at least one NNM. The most common NNM were myalgia (32.7%), headache (21.4%), loss of smell and/or taste (15.3%), and delirium (10.2%). Patients with and without NNM did not differ with respect to demographic characteristics. Patients with NNM had significantly more constitutional symptoms (p = 0.017) and were more likely to have neurological and/or neuropsychiatric comorbid conditions (10.9% vs. 0.0%, p = 0.033) than those without NNM. Patients without documented NNM were more likely to have abnormalities on chest X-ray (p = 0.009) than those with NNM. Coronavirus disease 2019 related treatment and mortality did not differ between the groups. Conclusion Neurological and/or neuropsychiatric manifestations were common in hospitalised patients with COVID-19. The results suggest that while COVID-19 patients with NNM may have less of a respiratory phenotype they nonetheless have equivalent mortality rates. Contribution This study highlights the common NNM in patients with COVID-19 admitted to Tygerberg Hospital early in the pandemic and adds to the growing evidence of COVID-19 NNM.


Introduction
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) emerged in Wuhan City, China, in December 2019. 1,2The virus spread rapidly in China and other countries, and was soon characterised as a pandemic. 3South Africa remains the most affected country on the African continent with 4 070 434 confirmed cases and 102 595 deaths as at 18 March 2023. 4

Study setting
The study was conducted at Tygerberg Hospital, a 1380-bed public sector tertiary hospital in the Eastern Metropole of Cape Town that provides services to approximately 3.5 million people. 16The hospital serves mostly vulnerable populations from densely populated low-income communities and rural areas.Tygerberg Hospital was the first designated COVID-19 treatment centre in the Western Cape province. 17

Study sample
The first 100 laboratory confirmed COVID-19 patients admitted at Tygerberg Hospital.The inclusion criteria were: • Confirmed RT-PCR positive SARS-CoV-2 with results available on the National Health Laboratory Service database.• Must have been admitted to Tygerberg Hospital between 01 March 2020 to 30 June 2020 for COVID-19 and have complete medical records -either hard copy in the Medical Record Department or electronic copy uploaded on the Enterprise Content Management (ECM) of Tygerberg Hospital (complete medical record entails demographic, admission assessment, treatment, and outcome data).

Data collection
A thorough literature review was done to ascertain the common NNM in patients with COVID-19.A pilot review of medical records of 10 patients with COVID-19 admitted to Tygerberg Hospital was conducted to understand the design of the medical records and to determine the commonly recorded demographic, clinical, laboratory, treatment, and outcome data.A standardised form was designed to extract the data based on previous studies and available variables in the medical records of patients at Tygerberg Hospital.Data were collected regarding headache, loss of smell and/or taste, myalgia, delirium, focal neurology, altered consciousness, seizures and any other NNM reported in the clinical notes.
Because of the retrospective nature of the data and the uncertainty of symptom attribution to COVID-19 or other conditions, any NNM reported during the admission was included.Fatigue as a neurological symptom was not included as it is generally considered to be a constitutional symptom.Patients were defined as having NNM if they had any NNM noted in their records at any point during their admission.The presence of symptoms was recorded and not the duration and severity as these data were frequently not possible to extract from the medical records.The data were obtained from the electronic version of patient medical records uploaded on the ECM website of Tygerberg Hospital.Blood biochemistry results were accessed from the National Health Laboratory System of South Africa website, and radiological images and reports were accessed on the Picture Archiving Communication System (PACS) of the Western Cape Department of Health.Data were retrieved and entered into REDCap. 18

Non-neurological and/or neuropsychiatric symptoms and comorbidities associated with neurological and/or neuropsychiatric manifestations
Patients had between one and nine non-neurological and/ or neuropsychiatric symptoms (Table 2), grouped into respiratory (range 0-4), gastrointestinal (range 0-4),  constitutional (range 0-3), and other (range 0-2) symptoms.Those without NNM were more likely to have any respiratory symptom, demonstrating a trend towards significance (p = 0.075), whereas patients with NNM had significantly more constitutional symptoms than those without NNM (p = 0.017).When comparing individual symptoms between the groups, we found that patients with NNM were significantly more likely to have weakness and/or malaise (34.5% vs. 4.7%, p < 0.001) and upper respiratory tract (URT) symptoms (32.7% vs. 11.6%,p = 0.014), such as sore throat, than those without NNM.

Characteristics
It was theorised that weakness and/or malaise was likely associated with the NNM myalgia and that URT symptoms were likely associated with the NNM of loss of taste and/or smell.Therefore, post-hoc tests were conducted to investigate the associations between these symptoms, and statistically significant associations between weakness and/or malaise and myalgia (p < 0.001), and between URT symptoms and loss of taste and/or smell (p < 0.001), were indeed found.
In the sample, 91.8% had at least one comorbidity (range 0-6), with patients without NNM more likely to have a comorbidity than those with NNM, demonstrating a trend towards significance (p = 0.075).Comorbidities (Table 2) were grouped into respiratory (range 0-3), metabolic (range 0-4), and other (range 0-4).Patients with NNM were more likely to have a pre-existing neurological and/or psychiatric disorder (10.9% vs. 0.0%, p = 0.033), such as depression, Parkinson's disease, and dementia.Human immunodeficiency virus (HIV) was found in 20.0% of patients with NNM versus 20.9% of patients without NNM (p = 0.968).The only specific pre-existing disease that differed significantly between the groups was tuberculosis (TB), with patients without NNM being significantly more likely to have had TB in their lifetime (18.6% vs. 3.6%, p = 0.020).

Vital signs and special investigations in relation to neurological and/or neuropsychiatric manifestations
There were no significant group differences on any vital signs, although patients without NNM were more likely to have tachypnoea (p = 0.065) and a higher respiratory rate (p = 0.069), demonstrating a trend towards significance (Table 3).Blood gas values and derangements as well as special investigations did not differ significantly between patients with and without NNM (Table 4).Urea and electrolytes (UE) and full blood counts (FBC) were carried out on all patients and the results did not differ between the groups.C-reactive protein (CRP), lactate dehydrogenase (LDH), Alanine transaminase (ALT), and glycated haemoglobin (HbA1C) were carried out on at least 60% of participants and LDH was more likely to be requested on patients without NNM (p = 0.023).However, none of the results differed between the groups.Clinicians were more likely to request Vitamin B12 (25.5% vs. 9.3%, p = 0.064) and syphilis serology (25.5% vs. 9.3%, p = 0.064) on patients with NNM, demonstrating a trend towards significance.The majority of participants had a chest X-ray (96.9%) and patients without NNM were more likely to have abnormalities on chest X-ray (p = 0.010).They also had a greater number of abnormalities (range 0-4, p = 0.051), demonstrating a trend towards significance (Table 4).

Treatments and outcome in relation to neurological and/or neuropsychiatric manifestations
Patients with and without NNM did not differ significantly in terms of treatments received for COVID-19 (Table 5).
Patients with NNM were more likely to have been referred for specialist neurology and/or psychiatry input (10.9% vs. 0.0%, p = 0.033), but were not more likely to have received psychotropic drugs (5.7% vs. 2.5%, p = 0.632).Patients with and without NNM did not differ in terms of survival or length of stay in hospital (range 1-84 days).

Discussion
Observation indicated that NNM was commonly reported in hospitalised patients with acute COVID-19.Respiratory, gastrointestinal, and constitutional symptoms co-occurred with NNM, and most patients had at least one pre-existing condition.Patients with NNM had significantly more constitutional symptoms and underlying neurological and psychiatric disorders than those without NNM.Patients without NNM were more likely to have had pre-existing TB and abnormalities on chest X-ray than those with NNM.Supportive treatment was the mainstay of management, and the average length of hospitalisation was 10.0 ± 11.2 days with 13.3% mortality.Patients with NNM received equivalent COVID-19 supportive treatments and had similar mortality to those without NNM.
This study found that 56.1% of patients had at least one NNM and 13.3% had a severe or serious NNM.A similar retrospective study conducted in Iran among hospitalised patients showed a 63.9% prevalence of neurological symptoms in patients with COVID-19. 10The higher prevalence of NNM reported by the Iranian study may potentially be because of those investigators having access to more detailed clinical data, such as scales being used within the hospital environment.Furthermore, studies conducted in hospital and community-based settings in Egypt and Tunisia found a prevalence of 50.2% and 72.1%, respectively, of neurological manifestations among confirmed and probable COVID-19 patients. 13,14,15The prevalence found in our study is roughly similar to studies conducted in other settings, although methodological differences limit direct comparisons.
There were no differences between patients with and without NNM in relation to age, sex, ethnicity, and marital status in our study.In a previous study, females were more likely to endorse headache than males. 10The researchers did not analyse for sex differences in relation to individual NNM in this study because of the limited sample size, which was also predominantly female.Post-hoc testing revealed that headache did not differ according to sex in our sample.A recent systematic review and meta-analysis assessing sex differences in the prevalence of confirmed cases of COVID-19 showed a higher pooled prevalence in males globally.http://www.sajpsychiatry.orgOpen Access
It was found that patients with NNM had significantly more constitutional symptoms than those without NNM.This was likely because of an association between myalgia and malaise and/or weakness.Other studies have included fatigue under NNM, which we conceptualised as a constitutional symptom. 21esearchers also found that patients without NNM were more likely to have respiratory symptoms, although there was only a trend towards significance.It was postulated, and confirmed, that URT symptoms were associated with the NNM loss of smell and/or taste as infection and inflammation in URT tissues would also be more likely to influence taste and smell.
The majority of patients had at least one comorbid medical condition and patients without NNM were more likely to have a comorbid condition than those with NNM, demonstrating a trend towards significance.Metabolic syndrome related conditions and respiratory conditions were the most common comorbid conditions, and these did not differ between patients with and without NNM.In contrast, Khedr et al., 13 demonstrated that pre-existing hypertension, ischaemic heart disease, diabetes mellitus, impaired liver function, impaired renal function, lung disease, and neurological disease are risks factors for the development of neurological manifestations in patients with COVID-19.
The present study found that COVID-19 infected patients with pre-existing neurological and psychiatric disorders were more likely to have NNM recorded in their clinical notes.Patients who had lifetime TB were also found to be less likely to have NNM as compared to those with no history of TB.This isolated finding is contrary to the findings of Khedr et al. 13 who showed that chest disease was one of the risk factors for development of neurological manifestations in patients with COVID-19.The presence and absence of NNM did not differ with regards to HIV.This finding is consistent with a previous study in the same setting which did not find significant differences between patients with and without HIV in relation to severity and outcome of COVID-19. 17tal signs and special investigations in the majority of patients revealed acute COVID-19 illness as evidenced by tachypnoea, tachycardia, elevated blood pressure, low oxygen saturation (with PF < 300), and elevated CRP and LDH.Studies have shown that lower oxygen saturation and diastolic blood pressure, and higher respiratory rate and blood glucose level in patients with COVID-19 are associated with disease severity and mortality. 23,24,25n addition, severe COVID-19 has been linked to neuropsychiatric manifestations and disorders during acute disease and post-COVID-19. 26,27,28is study did not find differences between patients who had NNM and those who did not with regards to vital signs including oxygen saturation.However, patients without NNM had higher respiratory rates and were more likely to have tachypnoea, demonstrating a trend towards significance.C-reactive protein and LDH were markedly elevated in the sample overall as reported in other literature. 29,30However, the level of CRP and LDH did not differ in patients with and without NNM.Clinicians were, however, more likely to request LDH on patients without NNM.Lactate dehydrogenase is a marker of more severe disease, and it is postulated that this may suggest that clinicians were more likely to consider patients without NNM as being more severely ill and to request LDH on them.
Unlike vital signs and blood biochemistry results, chest X-ray abnormalities were more frequently observed in patients without NNM, and they had more abnormalities on their chest X-ray, demonstrating a trend towards significance.A previous study has linked the degree of chest radiological (lung) abnormalities to the severity of COVID-19 and poor prognosis. 31Several studies have linked the severity of COVID-19 to neuropsychiatric manifestations and sequelae. 26,27,32Though the indirect link of chest radiological abnormalities and severity of COVID-19 to neuropsychiatric sequelae has been established, we did not find any studies directly linking chest X-ray abnormalities to NNM.Finally, treatment, duration of hospitalisation and outcomes did not differ significantly between patients with and without NNM.
Patients without NNM were more likely to have a 'respiratory' phenotype than those with NNM.This is supported by patients without NNM being more likely to have respiratory symptoms, more likely to have had TB, more likely to have tachypnoea and a higher respiratory rate, and more likely to have abnormalities on chest X-ray.Despite this, they were not more likely to be ventilated.However, patients with a poor prognosis may not have been candidates for ventilation.Despite having more severe respiratory features, patients without NNM also did not have a higher mortality rate.Patients with NNM may thus have a phenotype that involves less typical COVID-19 respiratory features but have a similar prognosis and outcome.

Limitations and strengths
The study was conducted at a single medical centre and only included hospitalised patients; thus, the findings may not be representative of the general population.Moreover, some important clinical features may have been omitted from documentation.For instance, early in the pandemic clinicians may have been more focussed on the typical respiratory symptoms related to COVID-19, and may have paid less attention to NNM.Certain laboratory tests were not done in all the patients.Missing data may limit the generalisability of the study findings.Furthermore, it is important to note that NNM that presented at various points during admission cannot necessarily be attributed to COVID-19 infection.Side effects of treatment or comorbid conditions may have caused NNM.
However, even with these limitations, we assume that the study is a representation of hospitalised COVID-19 patients with NNM in the Western Cape province.To the best of our knowledge, this is the first study reporting NNM in hospitalised patients with COVID-19 in South Africa.
Although the sample size is small, the study shows that NNM are common in patients with COVID-19; however, there is a need for other studies in South Africa and on the African continent, and in larger samples, to corroborate these findings.

Conclusion
Neurological and/or neuropsychiatric Manifestations are common in patients with COVID-19.This study highlights that even during acute illness with COVID-19, features of NNM can be identified.Raising awareness, early identification during-and post-hospitalisation, and monitoring and management of NNM are important to potentially avert medium and long-term sequelae, such as long COVID-19.
http://www.sajpsychiatry.orgOpen Access The data are not publicly available due to privacy and/or ethical restrictions.

TABLE 4 (
Continues...):Comparison of special investigations in patients with and without neurological/neuropsychiatric manifestations.

Results Sociodemographic characteristics associated with neurological and/or neuropsychiatric manifestations
Statistical Package for the Social Sciences (SPSS) version 28([computer program].Armonk (NY): IBM Corporation; 2021) was used to analyse the data.Descriptive numerical data with a normal distribution were described using means and standard deviations (s.d.), whereas non-normal data were described using medians and interquartile ranges (IQRs).

TABLE 1 :
Comparison of socio-demographic characteristics in patients with and without neurological and/or neuropsychiatric manifestations.

TABLE 2 :
Comparison of patients with and without neurological/neuropsychiatric manifestations with regards to non-neurological symptoms and comorbidities.
Note: Constitutional symptoms consist of fever, malaise, and body weakness.Metabolic comorbidities: hypertension, dyslipidaemia, diabetes mellitus and obesity.Respiratory comorbidities: current and past pulmonary tuberculosis (TB), pneumonia, obstructive sleep apnoea and other chronic lung diseases.Other comorbidities:human immunodeficiency virus (HIV), gastro-intestinal conditions, kidney disease and cardiac disease; Bold values denotes significant p-values.IQR, Interquartile range; NNM, neurological and/or neuropsychiatric manifestations.†, Mann-Whitney U; ‡, Pearson Chi-square; §, score; ¶, degrees of freedom; † †, Fisher's exact p-values are reported for cells that have less than five observations. 19

TABLE 3 :
Comparison of vital signs in patients with and without neurological and/or neuropsychiatric manifestations: Observations (N = 98).
Note: The highest oxygen saturations recorded during the admission were within reference range ≥ 95%; therefore, it was not reported in the table.NNM, neurological and/or neuropsychiatric manifestations; IQR, Interquartile range.†,Mann-Whitney U; ‡, Pearson Chi-square; §, Z score; ¶, degrees of freedom; † †, Fisher's exact p-values are reported for cells that have less than 5 observations.The highest oxygen saturations recorded during the admission were within reference range ≥ 95%; therefore, it was not reported in the table.http://www.sajpsychiatry.orgOpen Access

TABLE 4 :
Comparison of special investigations in patients with and without neurological neuropsychiatric manifestations.