This prospective study was conducted at Sakakibara Heart Institute, the hospital with the highest number of cardiovascular surgeries and the largest cardiac rehabilitation center in Japan. We included 236 patients (median age, 73 years; 34% female) who required an ICU stay of at least 72 hours after undergoing cardiovascular surgery and underwent postoperative rehabilitation at our hospital between May 2018 and November 2020. The exclusion criteria were ICU stay 72 hours (n = 1392), postoperative cerebral infarction or spinal cord infarction (n = 38), hospital transfer within 7 days after surgery (n = 4), in-hospital death (n = 7), and no assessment of physical function at ICU discharge (n = 294) (Fig. 1). The criteria for ICU discharge was according to the ICU guideline [8].

Fig. 1.

Schematic presentation of study procedure. This study included 236 patients who required an ICU stay of at least 72 hours after undergoing cardiovascular surgery and underwent postoperative rehabilitation.

The Barthel index (BI) was used to assess ADL levels preoperatively and at discharge, and HAD was defined as a decrease in the BI score between discharge and prior to surgery of at least 5 points. The BI consists of 10 items: (1) feeding, (2) moving from wheelchair to bed and return, (3) personal toilet, (4) getting on and off the toilet, (5) bathing, (6) transferring, (7) ascending and descending stairs, (8) dressing, (9) controlling bowels, and (10) controlling bladder. Independence is scored as 0, 5, 10, and 15 points according to the degree, and the total score was calculated for each patient [9]. The preoperative ADL level was obtained either directly from the patient or from relatives, about the condition that was not exacerbated before admission.

Physical Function ICU Test-scored (PFIT-s), Functional Status Score for the ICU (FSS-ICU), and Medical Research Council (MRC)-sumscore were used to assess physical function in the ICU. PFIT-s is composed of four items: (1) sit-to-stand assistance, (2) cadence, (3) shoulder flexion strength, and (4) knee extension strength; each item was scored to a sum of 0 to 3 points and evaluated on a 10-point scale converted to an interval scale [10]. FSS-ICU was scored from 0 to 7 (35 points total) for each of the five following actions: (1) rolling, (2) transfer from supine position to sitting, (3) sitting at the edge of bed, (4) transfer from sitting to standing, and (5) walking according to the degree of independence [11, 12]. The MRC-sumscore classifies the muscular strength in shoulder abduction, elbow flexion, wrist dorsiflexion, hip flexion, knee extension, and ankle dorsiflexion, attributing a score from 0 to 5 [13]. Each assessment was measured one day before or on the day of ICU discharge. Physical function assessment at ICU discharge was performed by three physiotherapists with at least 5 years’ clinical experience. Preoperative and discharge BIs are routinely assessed by our hospital’s well-trained physiotherapists. Data were collected by researchers who were blinded to the predictive factors.

We also examined arousal (Richmond Agitation-Sedation Scale) [14], pain (Numerical Rating Scale) [15], circulatory dynamics (mean blood pressure, heart rate, and atrial fibrillation), respiratory status (percutaneous oxygen saturation), and ICU mobility scale [16] at the time of physical function assessment.

Postoperative cardiac rehabilitation was performed in compliance with the Guidelines for Rehabilitation in Cardiovascular Diseases established in 2012 [17] and in consultation with the attending physician. The days on which a patient was able to independently sit, stand, practice walking, and walk 100 m are referred to as the starting date of each of these activities. In addition, we assessed the extent of exercise performed in the cardiac rehabilitation room, consisting mainly of aerobic exercise using a bicycle ergometer and treadmill as well as resistance training.

The following data were collected from medical records: age, sex, body mass index (BMI), medical history of comorbidities, preoperative blood biochemistry examination, intraoperative records, Acute Physiology and Chronic Health Evaluation (APACHE) II score, duration of ventilator intubation, intra-aortic balloon pumping, continuous renal replacement therapy, noninvasive positive pressure ventilation (NPPV), and postoperative complications. Postoperative delirium was assessed using the Confusion Assessment Method for the ICU [18].

Continuous variables are reported as the median (interquartile range [IQR]), whereas categorical variables are expressed as percentages. The Shapiro–Wilk test was used to verify a normal distribution. In group comparisons, the Mann–Whitney U test were performed for continuous variables and the $\chi$${}^2$ test was used for categorical variables. Receiver operating characteristic (ROC) analyses were used to calculate cut-offs of physical functional evaluation as a predictor of HAD at ICU discharge, and continuous variables were converted to categorical data. The cut-off value was calculated using the tangential method from the AUC. Multivariate logistic regression analysis was performed with the presence of HAD as the dependent variable and factors, both continuous and categorical variables. Age, BMI, APACHEII score, delirium, day of onset of standing, and each physical function that was significantly different in the univariate analysis were included as adjusted variables. The level of significance was set at ˂0.05, and all statistical analyses were performed using IBM SPSS Statistics 21 (IBM Corp., Armonk, NY, USA). There was no missing information.

In this study, the incidence of HAD was 24.6%, and age, postoperative delirium, the starting date of standing, and physical function at ICU discharge were identified as associated factors. There have been many previous reports of age-related decline in physical function [19, 20], and the age-related decline in physical reserve and mental function, such as cognitive decline and depression, have been shown to be risk factors for HAD [6]. In addition, patients with HAD had significantly longer duration of ventilator intubation, increased NPPV usage, and longer ICU stay. In cases of unstable postoperative respiratory and circulatory dynamics, patients are forced to rest in bed and require long-term ICU management. Under these circumstances, postoperative delirium is more likely to develop and has been shown to be a risk factor for ADL reduction [21, 22]. In the older adult population, environmental changes due to hospitalization may lead to decreased ADL independence and activity, resulting in HAD [6]. Recently, it was reported that if physical function and ADL were impaired during ICU stay, these impairments remained following ICU discharge [23, 24]. It has also been shown that low physical function during the ICU stay is a predictor of a decrease in QOL and short-term and long-term prognosis after ICU discharge [4, 5]. Although early mobilization can help to prevent a decline in physical function and ADL and improve the QOL [25, 26, 27], usually patients with HAD require postoperative recumbency and long-term ventilator management due to major surgery, leading to delays in the ability of standing and walking. Therefore, in addition to the marked decrease in physical function at the time of ICU discharge and delayed rehabilitation progression thereafter, the failure to transition from exercise therapy in the ward to that in the cardiac rehabilitation room may have prevented a re-acquisition of function until the time of discharge.

In this study, the traditionally widely used PFIT-s, FSS-ICU, and MRC-sumscore were utilized as an assessment index for physical function in the ICU. Although the FSS-ICU was identified as a significant factor in univariate analysis, the significance was lost in multivariate analysis. Drains and infusions are often placed in the long term after cardiovascular surgery, inhibiting physical activity and movement in ICU patients. FSS-ICU was scored based on the degree of independence in movement, but the postoperative placement decreased the movement at ICU discharge, resulting in a lower score indicating that the assessment as a physical function lacked validity. The MRC-sumscore evaluates limb muscle strength, whereas the PFIT-s comprehensively assesses and scores physical performance and endurance in addition to muscle strength. Therefore, we believe that the PFIT-s was more helpful as a method to predict HAD in this study.

Generally, safety and validity are considered when assessing physical function in the ICU [28]. In the immediate postoperative period, exacerbation of pain and variations in respiratory and circulatory dynamics are likely to occur during the evaluation, and these effects may be reflected in the evaluation results. In this study, at the time of evaluating physical function, the arousal level and analgesic control were close to normal and respiratory and circulatory dynamics were stable. In addition, 97.5% of the patients with HAD were able to secure until transferring, suggesting that the restriction of activities did not affect the evaluation of physical functions. Altogether, to prevent HAD, we believe that assessing physical function at ICU discharge could be suitable to design a rehabilitation program following major surgery and a subsequent stay in the ICU.