Long-Term Evolution of Activities of Daily Life (ADLs) in Critically Ill COVID-19 Patients, a Case Series

Background: The most common long-term symptoms of critically ill COVID-19 patients are fatigue, dyspnea and mental confusion. Adequate monitoring of long-term morbidity, mainly analyzing the activities of daily life (ADLs), allows better patient management after hospital discharge. The aim was to report long-term ADL evolution in critically ill COVID-19 patients admitted to a COVID-19 center in Lugano (Switzerland). Methods: A retrospective analysis on consecutive patients discharged alive from ICU with COVID-19 ARDS was performed based on a follow-up one year after hospital discharge; ADLs were assessed through the Barthel index (BI) and the Karnofsky Performance Status (KPS) scale. The primary objective was to assess differences in ADLs at hospital discharge (acute ADLs) and one-year follow-up (chronic ADLs). The secondary objective was to explore any correlations between ADLs and multiple measures at admission and during the ICU stay. Results: A total of 38 consecutive patients were admitted to the ICU; a t-test analysis between acute and chronic ADLs through BI showed a significant improvement at one year post discharge (t = −5.211, p < 0.0001); similarly, every single task of BI showed the same results (p < 0.0001 for each task of BI). The mean KPS was 86.47 (SD 20.9) at hospital discharge and 99.6 at 1 year post discharge (p = 0.02). Thirteen (34%) patients deceased during the first 28 days in the ICU; no patient died after hospital discharge. Conclusions: Based on BI and KPS, patients reached complete functional recovery of ADLs one year after critical COVID-19.


Introduction
The SARS-CoV-2 virus emerged in December 2019 in Wuhan, China, and spread rapidly across the globe [1]; on 20 February 2020, the first Swiss patient with SARS-CoV-2related pneumonia was accepted in our center in Lugano [2]. In March 2020, the World Health Organization (WHO) declared a global pandemic due to this new coronavirus   [3]; in the following two years, John Hopkins University's Coronavirus Resource Center has recorded more than 630 million cases and more than 6 million deaths [4].
Following the acute phase, emotional and psychological distress may persist for an extended period, further contributing to multifactorial disabilities requiring continuous care and multidisciplinary rehabilitation management [26][27][28]. A recent meta-analysis reported a significant incidence of post-traumatic stress disorder (39%), symptoms of depression (33%) and anxiety (30%) affecting survivors of the most severe forms of COVID-19 six months after hospital discharge [29]. Careful evaluation of long-term morbidity outcomes in critically ill COVID-19 patients is essential to optimize patient management after hospital discharge [30,31]; an essential aspect regarding physical sequelae concerns the ability to regain independence in activities of daily life (ADLs).
The aim of this study was to describe the long-term physical consequences in critically ill COVID-19 patients, identifying long-term sequelae in patients' ADLs through a followup analysis of the Barthel index (BI) and Karnofsky Performance Status (KPS) scale one year after hospital discharge.

Materials and Methods
A retrospective analysis of consecutive critically ill COVID-19 patients discharged alive from the ICU in our COVID-19 center during the first pandemic wave was performed, from 16 March to 10 April 2020. Patients transferred to other ICUs were excluded from the analysis. We included patients with ARDS due to a SARS-CoV-2 infection confirmed by a positive result from real-time reverse transcriptase-polymerase chain reaction (RT-PCR) on nasal or pharyngeal swabs [32]. Clinical data regarding the acute phase, defined as the hospitalization time in the ICU and acute medicine department, from electronic health records were retrieved. We collected the following demographic and clinical characteristics: age, sex, body mass index (BMI), comorbidities (arterial hypertension (HTA), diabetes mellitus (DM), obstructive sleep apnea syndrome (OSAS), chronic obstructive pulmonary disease (COPD) and ischemic heart disease (IHD)), pulmonary embolism (VTE), days of symptoms before hospital admission, ICU length of stay (ICU LOS), Simplified Acute Physiology Score II (SAPS) during the first 24 h, admitting Sequential Organ Failure Assessment (SOFA), days of mechanical ventilation, pronation sessions, presence of tracheostomy and development of ventilator-associated pneumonia (VAP). Further, we collected the following laboratory test results: ASAT, ALAT, leucocytes, lymphocytes, lactates (all at admission in the ICU), minimum platelet level and maximum total bilirubin, CK, CRP, LDH, ferritin and creatinine (during the entire ICU stay).
A follow-up evaluation one year after hospital discharge was performed by contacting patients by telephone to gather information on survival and their performance status.
The BI measures ten essential self-care and physical dependency aspects, rating each ADL's element on a semiquantitative scale with high inter-rater and test-retest reliability [38]. A score of 100 denotes normality, and lower scores indicate increasing disability. Notably, there is a direct correlation between the ability to carry out these everyday activities and the degree of autonomy necessary to live at home after hospital discharge [33].
Single ADLs, such as getting out of bed, going to the toilet, dressing and eating, can be used as markers of an individual's functional status during standard and serial ADL screening to detect the presence and the degree of a specific disability [39].
The KPS was primarily developed to assess a patient's ability to survive chemotherapy [40][41][42][43]. However, it is helpful for assessing functional impairment by measuring a patient's general performance status or ability to carry out ADLs [44].

Outcomes
The primary objective was to evaluate the difference in ADLs at hospital discharge (acute ADLs) and 12 months later (chronic ADLs) by administering the BI and KPS. We looked for correlations between acute/chronic ADLs and clinical and biological characteristics observed during the ICU stay as a secondary objective.

Statistical Analysis
Descriptive statistics were used to summarize the clinical data. Data are presented as mean (SD) or median (IQR) for continuous variables according to the data distribution and as absolute numbers (percentage) for categorical variables. The data distribution was verified by the Kolmogorov-Smirnov and Shapiro-Wilk tests (Table S3). Differences between continuous variables by the paired t-test or the Mann-Whitney test for independent groups requiring a nonparametric analysis were investigated. Clinical evolution over time was compared using the paired t-test or the nonparametric Wilcoxon test depending on data distribution, and we analyzed the relationships between continuous variables by linear regression. All intervals of confidence (CI) were established at 95%. The type 1 error rate was 0.05. Statistical significance was considered with a p value less than 0.05. Statistical data analysis was performed using the SPSS.26 package (SPSS Inc., Armonk, NY, USA). The mean age was 59 years (SD 12), 21 (84%) were men and the mean BMI was 29 kg/m 2 (SD 5); 11 (44%) patients had arterial hypertension, 8 (32%) had diabetes, 3 (12%) had OSAS, 1 (4%) had COPD and 1 (4%) patient presented pulmonary embolism during the SARS-CoV-2 infection. The survivors' mean ICU length of stay (LOS) was 12 days (SD 8). The mean duration of mechanical ventilation (MV) was 11.8 (SD 9) days; 20 (80%) The mean age was 59 years (SD 12), 21 (84%) were men and the mean BMI was 29 kg/m 2 (SD 5); 11 (44%) patients had arterial hypertension, 8 (32%) had diabetes, 3 (12%) had OSAS, 1 (4%) had COPD and 1 (4%) patient presented pulmonary embolism during the SARS-CoV-2 infection. The survivors' mean ICU length of stay (LOS) was 12 days (SD 8). The mean duration of mechanical ventilation (MV) was 11.8 (SD 9) days; 20 (80%) patients received invasive MV, 4 (16%) patients received a tracheostomy and 4 (16%) patients required continuous renal replacement therapy (CRRT). At ICU admission, no patients required vasopressors. Demographic, clinical and laboratory characteristics are shown in Table 1. Patients' characteristics during the ICU stay regarding clinical, laboratory and intra-ICU data. Data on SAPS II, SOFA, NEMS, systolic and diastolic arterial blood pressure, heart rate and temperature were reported from ICU admission; *: at ICU admission. Continuous measurements are presented as mean ± SD (min-max) and otherwise as median (25th-75th) if they are not normally distributed. Categorical variables are reported as counts and percentages. No patients were on vasopressor at ICU admission. CRRT = continuous renal replacement therapies, ICU = intensive care unit, ICU LOS = intensive care unit length of stay, MV = mechanical ventilation, ASAT = aspartate-aminotransferase, ALAT = alanine-aminotransferase, CRP = C-reactive protein, LDH = Lactate dehydrogenase and CK = creatine kinase.

Primary Outcome
The median BI was 75 (IQR 55-97.5, min/max 5-100) at hospital discharge (acute ADLs) and 100 (IQR 100-100; diff = 25, Z = −3.823, p < 0.0001) at one year post discharge (chronic ADLs); chronic ADLs showed a complete recovery in all analyzed activities ( Figure 2). Data regarding the single BI activities at hospital discharge (acute ADLs) and at oneyear follow-up (chronic ADLs) were compared and show significant differences for several items ( Table 2).  Data regarding the single BI activities at hospital discharge (acute ADLs) and at oneyear follow-up (chronic ADLs) were compared and show significant differences for several items ( Table 2).

Secondary Outcomes
No significant correlations between the BI at discharge and demographic, biological and ICU-specific data were found (Table 3) except for a slight trend between the SOFA score at ICU admission and acute BI (r 2 = 0.2).    A slight reverse correlation between the BI at hospital discharge and ICU LOS (r 2 = 0.08), MV days (r 2 = 0.117) and the SOFA score at ICU admission (r 2 = 0.2, Figure 3) was observed. A slight reverse correlation between the BI at hospital discharge and ICU LOS (r 2 = 0.08), MV days (r 2 = 0.117) and the SOFA score at ICU admission (r 2 = 0.2, Figure 3) was observed. Patients with OSAS (t = −0.833, dF 2.207, p = 0.485) and those treated with invasive MV (t = −2.173, dF 13.764, p = 0.048) and with nosocomial infections (t = −0.383, dF 4.972, p = 0.718) presented a slightly different BI at hospital discharge than the rest of the cohort, without any significant differences. A similar reverse correlation between KPS at hospital discharge and ICU LOS was found (r 2 = 0.661, p = 0.02), and simple correlation trends between KPS and MV days (r 2 = 0.536, p = 0.385) and between KPS at hospital discharge and age were similarly encountered (r 2 = 0.185, p = 0.139, Figure 4). Patients with OSAS (t = −0.833, dF 2.207, p = 0.485) and those treated with invasive MV (t = −2.173, dF 13.764, p = 0.048) and with nosocomial infections (t = −0.383, dF 4.972, p = 0.718) presented a slightly different BI at hospital discharge than the rest of the cohort, without any significant differences. A similar reverse correlation between KPS at hospital discharge and ICU LOS was found (r 2 = 0.661, p = 0.02), and simple correlation trends between KPS and MV days (r 2 = 0.536, p = 0.385) and between KPS at hospital discharge and age were similarly encountered (r 2 = 0.185, p = 0.139, Figure 4). Further, no correlations between any single acute ADL element of the BI and clinical, biological or intra-ICU data were found (Table 4).  Further, no correlations between any single acute ADL element of the BI and clinical, biological or intra-ICU data were found (Table 4).  In particular, there was no correlation between feeding and invasive MV/tracheostomy (Z = −1.065, p = 0.287 and Z = −0.639, p = 0.523, respectively), between walking and invasive MV/tracheostomy (Z = −1.432, p = 0.152 and Z = −0.329, p = 0.742, respectively) or walking and ICU LOS/MV days (r 2 = 0.038, p = 0.739 and r 2 = 0.018, p = 0.536, respectively).

Discussion
Our data showed that critically ill COVID-19 patients discharged from the hospital presented significant improvements within the first year of follow-up, fully recovering ADLs and KPS functional status.
With a mean BI of 75 and a mean KPS of 86 at hospital discharge, critically ill COVID-19 patients presented a moderate degree of ADL dependency at hospital discharge and could perform regular activities with some effort despite some signs or symptoms of disease. One year later, our entire cohort achieved a BI of 100, meaning complete functional autonomy, and a KPS of 99.6, which indicates the absence of complaints and the ability to carry on regular activities and work. These results are encouraging because they showed that COVID-19-related ARDS survivors can cope with the massive catabolic state inherent to the acute critical illness and the numerous symptoms and limitations reported by patients up to 12 months after acute COVID-19 illness, called the post-COVID-19 syndrome [9,17,19,45,46].
In a Spanish multicentric cohort of 113 patients, the mean BI was 99 at a median of 240 days from the first positive PCR test for SARS-CoV-2 despite 80.5% having at least one residual symptom [9]. One of the most frequent symptoms was fatigue, which affects about 60% of patients one year after discharge [19,47]. It may persist as a direct consequence of some level of immune activation with or without persistent viral infection [48][49][50][51][52][53][54], long-term lung tissue damage, lasting neurological complications [55,56], myocardial injury [57,58] and other extrapulmonary involvement [59]. Impaired muscle function and deconditioning may also explain the compromised functional ability, impacting the 6 min walking distance. Notably, virtually all ARDS patients exhibited severe muscle waste and weakness in the acute phase, and only 70% returned to their baseline weight by one year [25]. Nine months after SARS-CoV-2-induced ARDS, about half of the patients completed less than 80% of the theoretical reference distance in the 6 min walk distance test (6MWDT) [9]. In two other studies, the 6MWDT showed significant recovery between 3 and 12 months in most ICU survivors after SARS-CoV-2 pneumonia but remained below the predicted value for 20-25% of patients with no association with D LCO alterations. The functional improvement occurred despite significant radiological lung parenchyma alterations (reticulations, traction bronchiectasis, honeycombing, ground-glass opacities and emphysema) persisting in most patients (80-95%) 12 months after COVID-19 related ARDS [47,60].
Additionally, 40-45% of patients report some degree of breathlessness, mainly modified Medical Research Council Dyspnea scale (mMRC) grades 1 and 2, one year after acute SARS-CoV-2 pneumonia, in some cases with a worsening trend compared to the 6-month follow-up [9,19,47,61,62]. At this time point, reduced D LCO and pulmonary restriction still affect approximately 50-60% and 7-30% of patients and are proportional to the severity of lung failure expressed as the level of respiratory support during the acute SARS-CoV-2 infection [19,47].
According to reports on COVID-19-related ARDS survivors, the health-related quality of life (HRQoL) was worse than in the general population [9,60], which is also consistent with previous studies among patients affected by ARDS of different etiologies [25,63]. Mean Short-Form 36 (SF36) scores were significantly worse than the general population for each of the eight dimensions in a Spanish multicentric cohort eight months post discharge [9]. In a second French monocentric observational study, SF36's emotional role domain normalized three months after hospital discharge and the physical role within 12 months, but the other six domains continued to show reduced scores [60]. Psychological and emotional dysfunction is known to persist for up to five years after ICU discharge [64]. Accordingly, many patients develop anxiety disorders (30%), depression (33%) and post-traumatic stress disorder (PTSD) symptoms (39%) [29], potentially aggravated by the past pandemic period.
Regardless of the primary illness, survivors of an extended ICU stay may experience medium-and long-term morbidity related to critical illness, the necessary support and the environment. This condition, which may include new or worsening cognitive, psychiatric and physical impairment, is now recognized as post-intensive care syndrome (PICS) [65][66][67][68][69].
Therefore, survivors of critical COVID-19 may experience a range of sequelae related to their critical condition (i.e., PICS), the SARS-CoV-2 infection (i.e., post-acute COVID-19 syndrome (PACS)) or both. In the present study, no patients reported any specific or persistent symptoms 1 year after hospital discharge.
This study is part of the constantly growing group of scientific evidence reporting the medium-long term clinical conditions of critical COVID-19 patients, not only confirming that the post-discharge mortality rate is confirmed to be low but also, above all, underlining how functional recovery-despite potential residual symptoms-appears to be almost complete, with the possibility of having a quality of life completely similar to the situation prior to admission. We recently showed that critically ill COVID-19 patients could recover certain physical functions (swallowing) significantly faster than different critically ill patients [70]. This information appears even more useful observing the correlation between the BI at hospital discharge and the SOFA score at ICU admission, suggesting that even for patients initially considered most critical, the chances of long-term recovery are optimal once they are discharged from the ICU. Precisely in line with this peculiarity of COVID-19 patients, Biehl et al. [71] analyzed non-COVID-19 critically ill patients with/without ARDS at 6 months using the Barthel index, finding no significant differences between baseline and 6-month ADLs. These findings suggested that critically ill COVID-19 patients have the potential to recover better after the acute phase of damage than critical patients with ARDS of another nature for a number of causes related to the disease and the characteristics of patients discharged alive from the ICU, which will certainly be the subject of future studies.
This study presented some limitations. Firstly, it was a single-center retrospective study enrolling a fairly small number of patients; although our results are in line with other groups [9,62], further confirmatory studies are needed. Second, the follow-up assessment was performed over the phone and not face-to-face in the clinic. The information collected was based on the patient's self-assessment and not on a direct medical evaluation, although it is important to note that the BI and the KPS can be administered reliably via telephone conversation [72]. Again, we do not have any homogenous information about the rehabilitation programs the patients participated in, as they were transferred to different institutions with patient-specific nonstandardized programs. Finally, we did not look for the presence of psychiatric or psychological disorders either in the acute or in the chronic phase, although they can affect ADLs [73,74] both positively and negatively; the Barthel index is designed to monitor pure functional autonomy for ADLs and does not cover cognitive decline, which of course could affect the global functional result.

Conclusions
Critically ill COVID-19 patients showed complete recovery of ADLs and performance status one year after acute illness. Despite PICS and PACS, this long-term perspective justifies prolonged hospitalization in the ICU and the use of invasive and aggressive techniques to overcome the most acute and dangerous phase.  Institutional Review Board Statement: According to the local federal rules, the study was submitted to the Ethics Committee of Canton Ticino (Comitato Etico Cantonale, project ID 2021-01342, CE_TI_3906), which identified it as a case series. No funding was required.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.