Clinical characteristics, risk factors and complications of COVID-19 among critically ill older adults - A case control study [version 1; peer review: awaiting peer review]

Background: The older population is often disproportionately and adversely affected during humanitarian emergencies, as has also been seen during the COVID-19 pandemic. Data regarding COVID-19 in older adults is usually over-generalized and does not delve into details of the clinical characteristics in them. This study was conducted to analyze clinical and laboratory characteristics, risk factors, and complications of COVID-19 between older adults who survived and those who did not. Methods: We conducted a case-control study among older adults(age> 60 years) admitted to the Intensive Care Unit(ICU) during the COVID-19 pandemic. The non-survivors(cases) were matched with age and sex-matched survivors (control) in a ratio of 1: 3. The data regarding socio-demographics, clinical characteristics, complications, treatment, laboratory data, and outcomes were analyzed. Results: The most common signs and symptoms observed were fever (cases vs controls)(68.92 vs. 68.8%), followed by shortness of


Introduction
There is evidence of higher death among the elderly with severe COVID-19. [1][2][3][4][5] The USA reported 8 out of 10 deaths among those 65 years or older. 6 In addition, compared to 18-29-year-olds, the risk of death among the age groups are as follows: Age 30-39(4× higher), 40-49(10× higher), 50-64(30× higher), 65-74(90× higher), 75-84(220× higher), 85 and above (630× higher). In India, the mortality data showed a similar trend of being higher among older adults. Around 50% of deaths in India were among those 60 years and above, compared to 37% of deaths (45-60 years) and 11% of deaths (26-44 years). 7 Nevertheless, data among the older adults were not disaggregated, leading to the overgeneralization of clinical characteristics and outcomes among this vulnerable group. The socio-demographic factors and comorbidity profiles of older adults in LMIC 8 (Low and Middle-Income Countries) such as India vary greatly from countries such as the US, UK, and Europe. Most older adults in India live in families with younger working adults, who form the primary disease source. Nursing home residents of most older adults in the US 9 and Europe have a person-to-person spread between older adults. In northern Italy, around 600 elderly individuals died due to COVID-19 spreading within a nursing home. 10 These social factors may also impact the time of presentation of the patients to hospitals. There is a lack of quality data on the characteristics, risk factors, and outcomes of COVID pneumonia among older adults. This deficiency is exceptionally high in LMICs. 8 In addition, we assert that older adults are not a single homogenous group. 11 Thus, data disaggregation is needed to identify modifiable risk factors and to address and anticipate complications; this would ensure better clinical management of older adults with COVID pneumonia. Furthermore, data disaggregation will help understand the health status of older adults, identify factors affecting their well-being, and identify drivers of inequality during the COVID pandemic. Older adults are known to present with atypical symptoms. Thus, we have examined the presence of the same in this study.
The study was undertaken to analyze the clinical and laboratory characteristics of older adults admitted to the Intensive Care Unit (ICU), with COVID-19 and to analyze the risk factors and their association with adverse outcomes among these critically ill older adults.

Methods
A hospital-based case-control study was undertaken. Data was collected from the Intensive Care Unit (ICU) from December 2020 to September 2022. The sample size was calculated with a two-sided confidence level(1-α) of 95, 80% power, and with a ratio of controls to cases at 3:1. A sample size of 260 was calculated consisting of 195 controls and 65 cases.
A Case was defined as a COVID-19-positive individual older than 60 years who, after being admitted or transferred to the ICU, did not survive, i.e., non-survivor. A Control was defined as a COVID-19-positive individual with an age greater than 60 years who was admitted or transferred to the ICU, following which the patient recovered (survived) and was discharged alive from the hospital, i.e., survivor. Those patients who were admitted for post-COVID-19 complications or COVID-19 unrelated medical conditions following discharge after initial treatment for COVID-19 pneumonia were excluded.
The cases (non-survivors) were recruited according to the inclusion and exclusion criteria mentioned above and were then matched with an age and sex-matched control (survivor) in a ratio of 1: 3, respectively. The data regarding sociodemographics, clinical characteristics, complications, treatment, laboratory data, and outcomes were collected using a modified ISARIC 12 form. The patient's identity was anonymized by assigning a code. The comorbidities and risk factors recorded in the study were chronic cardiac disease (including hypertension), chronic pulmonary disease (including asthma), chronic kidney disease, obesity, liver disease, asplenia, chronic neurological disorder, malignant neoplasm, chronic hematological disease, AIDS/HIV, diabetes mellitus, rheumatological disorder, dementia, tuberculosis, malnutrition, and smoking. Before the study's launch and data collection, approval was acquired from the Institutional Ethics Committee and the medical directors of the participating hospitals. The study was done after clearance from the Institutional ethics committee (Ref No. KMCXMLR-11/2020/338). Written informed consent was obtained from all patients.
Data collection was done using Microsoft Excel. Data was analyzed using the IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY: IBM Corp. The data was expressed as mean and SD for continuous variables. Based on the type of distribution of data, a t-test or Mann-Whitney U test was applied. The categorical variables were analyzed using Pearson's chi-square or Fisher's exact test based on the data distribution.

Demographic details
Out of 260 patients in our study, 186 were survivors (controls), and 74, non-survivors (cases) due to COVID-19 (Table 1). Among them 42.6% were females, and 57.4% were males. The majority belonged to the age group of 60 to 70 years (70.4%), 70 to 80 years (32.4%), and more than 80 years (1.6%). 55.7% belonged to APL (Above Poverty Line), 13 and the remaining 44.23% belonged to BPL. The age, sex, and social category were not significant among the two groups (p-value>0.05).

Signs and symptoms at admission
The heart rate and respiratory rate were higher among the non-survivors, and the difference was statistically significant when compared to the survivors (p value=0.0001). Further, systolic and diastolic blood pressure was higher among the non-survivors than among survivors, and the difference was not statistically significant (p-value>0.05). Saturation at room air was lower among the non-survivors than survivors, and the difference was statistically significant (p value=0.0001) ( Table 2).
The most common signs and symptoms observed among the non-survivors and survivors: were fever (68.92 vs. 68.8%), followed by shortness of breath (62.2% vs. 52.2%), and cough (47.3% vs. 60.2%). Our study revealed that wheezing was significantly higher among the non-survivors (28.4%) than among survivors (16.7%). The association was statistically significant, with a p-value of 0.03. Other symptoms across the two groups did not statistically differ from one another (p-value>0.05) ( Table 3).

Risk factors/comorbidities
The most common comorbidities or risk factors among both groups were hypertension (51.54%), followed by diabetes mellitus (46.54%), chronic cardiac disease (12.31%), chronic pulmonary disease, obesity, chronic neurological disorder, asthma, chronic kidney disease, malignant neoplasm, moderate or severe liver diseases, smoking, and mild liver disease. Our analysis found no significant association between the two groups regarding the presence of any of the above comorbidities (p-value>0.05) ( Table 4).

Laboratory investigations
At admission, neutrophil count, LDH (Lactate dehydrogenase), WBC (White Blood Count), creatinine, CRP(C-Reactive Protein), D-dimer, ferritin, and IL-6 (Interleukin-6) were significantly higher among the non-survivors than in the survivors. The difference was statistically significant with a p-value<0.05 (Table 5). Low levels of sodium, lymphocyte counts, and procalcitonin were seen among the non-survivors, and this association was found to be significant with a p-value<0.05. At admission, variables such as HbAlc (glycosylated hemoglobin), and potassium had no statistically significant association between the two groups (non-survivors and survivors).

Complications
Complications such as the development of seizure, bacteremia, acute renal injury, respiratory failure, and septic shock were significantly associated with cases (non-survivors) (p-value<0.05) ( Table 6).

Discussion
Our study had 260 patients, with 186 controls (survivors) and 74 cases (non-survivors). The principal residence was at home (99.23%) than nursing homes (0.77%). However, in most developed countries, the older population resides at facilities like nursing homes; this had an essential role in the transmission of COVID-19 in countries like Italy, 10 Canada, 14 and the United States. 15 Males were 57.4%, and females 42.6%. Few studies suggested males had a higher risk of mortality, 2,16,17 but our study did not find any such significance.
Most patients came under the ages 60-70(70.4%), followed by 70-80-year-olds(24.5%), least by above 80-year-olds age group(1.6%). We found that death was seen with increasing age of the individuals after stratification, and it is concurrent with available evidence in international studies. 16,[18][19][20][21] Our study found that hypoxia, tachycardia, and tachypnoea at admission were essential factors relevant to in-hospital mortality. These features align with current evidence. 22,23 Jasmine Ming Gan et al. 22 concurred that fever, cough, and shortness of breath were the commonest symptoms at presentation in decreasing order. There was no difference in mortality with gender in their study. Other atypical presentations of COVID-19 in the elderly in their study included the presence of falls, delirium, and reduced mobility. These patients first showed unusual symptoms but later were found to have typical COVID-19. symptoms such as fever and breathlessness with concurrent hypoxia. Such patients were seen to have worse clinical outcomes, but it was not statistically significant in their study.
Aman Nanda et al. 23 found that older adults with COVID-19 presented with fever, fatigue, and dry cough. Cough with sputum, headache, gastrointestinal symptoms, and anosmia was infrequent. Sore throat, hypoxia, tachypnoea, tachycardia, and delirium were atypical presentations in their study. Many studies found that patients admitted with COVID-19 had higher incidence of fever, 1,3 fatigue, 24 dry cough, 22,23 shortness of breath. 19 Fever, shortness of breath, and cough were the most common symptoms at presentation in our study. The atypical presentation features studied were confusion, inability to walk, nausea, vomiting, and diarrhea. Confusion/altered consciousness in our study was seen more among non-survivors (20.27%) than survivors (11.83%) but without statistical significance. Reduced mobility, i.e., inability to walk, was an infrequent feature in our study, seen more among survivors (9.68%) than the non-survivors (5.41%), with no statistical significance. Overall, in our study, atypical presentation of COVID-19 was not frequent. The patients in our study presented with more of the typical symptoms of COVID-19 than the atypical symptoms usually expected in older adults.
Among the clinical features at presentation, our study showed higher mortality with wheezing with statistical significance (p-value = 0.033). Nonetheless, there is limited evidence stating the role of asthma or wheezing due to COVID-19. 25 In addition, earlier data suggested no increase in the development of severe disease in patients with pre-existing asthma. 26 The most common comorbidities or risk factors among both groups were hypertension followed by diabetes mellitus, chronic cardiac, pulmonary disease, obesity, chronic neurological disorder, asthma, and chronic kidney disease. Incidence of chronic cardiac(16.22% vs 10.75%), pulmonary(16.22 vs 9.68), kidney disease(8.11% vs 7.53%) was higher among non-survivors, similarly for asthma(8.1% vs 4.84%), obesity(9.46% vs 5.38%), and others. These findings were consistent with many studies. 16,18,27 Richardson, S found that the frequent comorbidities noted were hypertension, followed by obesity and diabetes. 16 Attaway described factors associated with adverse outcomes as male gender, higher age, diabetes, hypertension, cardiovascular disease, and chronic pulmonary, kidney, or liver disease. 18 Giacomo Gracelli and others found hypertension to have the highest incidence, followed by cardiovascular diseases in individuals admitted with COVID-19 in Italy. Hypercholesterolemia was the next in order. Older patients had comorbid illnesses, but chronic pulmonary disease had a minor incidence. 27 The presence of comorbidities in our study did not predict mortality, contrary to the available evidence. 1,18-21,23 Nanchen Chen 20 found that COVID-19 patients have higher odds of death with comorbid illnesses and male sex. Annemarie and colleagues 19 suggested chronic cardiac disease was the most frequent comorbidity, followed by diabetes and chronic pulmonary and kidney disease. Older age, male sex, obesity, and chronic cardiac, pulmonary, kidney, and liver diseases were associated with higher mortality. Amy H Attaway and colleagues 18 described factors associated with adverse outcomes as male sex, older age, and the presence of diabetes, hypertension, cardiovascular disease, and chronic pulmonary, kidney, or liver disease. Aman Nanda 23 found higher odds of death among males and those with comorbid illnesses. Williamson 1 suggested an increased risk of death in patients with higher age, BMI, HbA1c, social deprivation, and presence of comorbidities (diabetes, severe asthma, respiratory disease, chronic cardiac disease, liver disease, stroke, reduced renal function, neurological diseases, autoimmune diseases, and immunosuppressive conditions).
Our study found a higher prevalence of diabetes(51.54%) and hypertension(46.54%) in the study population. The numbers were significantly higher than the average prevalence as per a representative study in India done among 1.3 million adults(7.5% and 25.3%, respectively), 28 which may suggest the increased incidence of these comorbidities with older age. Their presence may be associated with an increased need for ICU among those admitted. However, there was no statistical significance of their presence with death in our study. Our study found that laboratory investigations at admission, that were significantly higher among the non-survivors were LDH, WBC, Creatinine, CRP, D-dimer, Ferritin, and IL-6. The difference was statistically significant with a p-value <0.05. Investigations, such as serum sodium, lymphocyte count, and procalcitonin, were lower among non-survivors. The difference was statistically significant with a p-value<0.05. Parameters such as HbAlc, and potassium did not differ significantly between the two groups. In a study by Nanshan Chen, patients with lower leucocyte counts had higher mortality. 20 Aman Nanda found leucopenia, lymphopenia, elevated ferritin, LDH, PT (prothrombin time), and liver enzymes among patients with COVID-19. Those with a higher D-dimer and worse lymphopenia had higher odds of death. 23 A study by Zaishu Chen revealed that LDH elevation was significant for mortality among patients with or without comorbidities. In their study, CRP, D-dimer, lymphopenia, and lower sodium were seen more among nonsurvivors. Our study yielded similar findings. There was no statistical significance among creatinine and neutrophil counts, contrary to the findings in our study.
Various studies showed a role of elevated D-dimer and CRP in a higher risk of death, similar to our findings. [29][30][31] Our finding of a lower procalcitonin among the non-survivors is against the available evidence, 3,20,[29][30][31][32] but the number of patients with these values was low.
There is evidence of increased interleukins and chem0okines in patients with severe COVID. Our study showed that higher values of IL-6 were seen among non-survivors than survivors, but the number of samples was low. The same article suggested that lymphopenia was associated with death, similar to what was observed in our study. 32 Complications following hospitalization, such as acute renal injury, respiratory failure, seizure, bacteremia, and septic shock, were significantly associated with mortality (p-value <0.05). Lil Chen and others suggested that acute kidney injury (AKI) was associated with significant mortality among those admitted with COVID-19. The death rate was high in those with AKI in the ICUs. Around one-third of these patients did not recover kidney function in their study. 33 Rosenthal found sepsis, acute kidney injury, hyperkalemia, acidosis, acute liver injury, and neurological disorders as complications with higher odds of death. 34 The evidence further points toward critically ill patients with COVID-19 having bacteremia and septic shock in the hospital. 20 Another study found that acute respiratory distress syndrome and cardiac injury were related to more significant risks of death in critically ill individuals with COVID-19, 23 which was not observed in our study.
We found that those that developed in-hospital seizures (new onset) had higher mortality, with statistical significance (pvalue 0.03). Although our study's rate of convulsive seizures was low(8 of 260 patients), Amir Emami and colleagues found a similar outcome with COVID patients who developed new-onset seizures. 35 The underlying pathogenesis may be related to underlying hypoxia or metabolic causes like hyponatremia, which were independent risk factors for death in our study. Thus, we emphasize that further clarity is required in this matter.
Our study had a higher death rate in patients with respiratory failure. Pierachille Santus et al. 35 found that a greater death rate was independently linked with age >65, respiratory failure, and a PaO 2 /FiO 2 under 200 mmHg upon admission. However, our study did not measure the severity of the same with the outcome.
The other limitation of the study would be the insufficient data regarding the causes of bacteremia and sepsis in critically ill older adults. In addition, the role of sepsis in acute kidney injury needs further evaluation.

Conclusion
We found hypoxia, tachycardia, and tachypnoea on admission among the non-survivors of COVID-19. Therefore, there is a need for vigilance at admission by monitoring these vital parameters in older adults and the need for intensive care following the identification of hypoxia, tachypnea, and tachycardia. Clinical comorbidities were found to be equally distributed among survivors and non-survivors and were not associated with higher mortality. Worse lymphopenia was significantly seen among non-survivors. In addition, inflammatory markers such as LDH, CRP, ferritin, D-dimer, and IL-6 were significantly higher among non-survivors and thus can be used as markers of severity in critically older adults. Hyponatremia and elevated creatinine were more common among non-survivors. Among the clinical complications studied, older adults who did not survive were found to have seizures, bacteremia, AKI, septic shock, and respiratory failure than the survivors.
The onset of sepsis with Multi Organ Dysfunction Syndrome (MODS) probably leads to mortality in older adults, and prevention of the same may improve outcomes. Merely, chronological age or comorbidities alone do not signify an added risk for mortality among older adults in our study. The onset of sepsis with multiorgan failure can serve as a poor prognostic indicator in these patients.
In critically ill older adults with COVID-19, there is a need to identify the development of bacteremia and sepsis to reduce mortality. Identifying and using appropriate antimicrobials for concurrent infections may help control poorer outcomes. In addition, prevention of early hypoxia may reduce the incidence of in-hospital seizures and respiratory failure, further improving prognosis.
This project contains the following underlying data: • COVID_study_F1000.xlsx (the data collected in an Excel sheet, in a sheet labeled "Data", Legend explaining the data in sheet "Legend") • README.md (containing key information for understanding and reuse of our data) Data are available under the terms of the Creative Commons Zero "No rights reserved" data waiver (CC0 1.0 Public domain dedication).