2.1 Study design and setting

The trial was conducted between the 1st of September 2018 and the 31st of December 2019 at the department of anesthesiology and intensive care, university hospital of Ostrava.

2.2 Participants

Following their admission to the ICU, all patients on mechanical ventilation expected to last five days and more were screened. At the time of weaning from mechanical ventilation initiation, the study physician and rehabilitation worker assessed the patients’ eligibility using the criteria shown in Table 1. The cause of mechanical ventilation was noted see Table 2. According to the mobility protocol, standard rehabilitation was started within 48 hours after mechanical ventilation initiation.

2.3 Interventions and randomization

The weaning from mechanical ventilation was considered immediately after cardiorespiratory stability was achieved. The stability was defined as absence of low dose of vasopressors (norepinephrine <8 μg/min), inspiratory fraction of oxygen (FiO₂ <60%) and positive end-expiratory pressure (PEEP <10 cm H₂O). At the time of the beginning of weaning the patients were randomly allocated into the in-bed leg cycling group (m-group) undergoing standard rehabilitation together with in-bed cycling exercise and a control group (c-group) receiving only rehabilitation. The intended allocation ratio was 1:1. The randomization was performed by randomly pulling out the unmarked sealed envelopes pre-filled with the fixed number of group-specific designations.

In all randomized patients, the weaning was performed strictly according to local standard of care, including gradual reduction of synchronized intermittent mandatory ventilatory support (pressure control and PEEP), followed by spontaneous breathing on pressure support ventilation and intermittent T piece trials. We followed cooperative analgosedation concepts (target: richmond agitation and sedation score about 0 to –2) to during the weaning period using titrated dosage of opiates (sufentanil, S2CD02, Chilsi Pharmaceuticals, Austria) and short-term sedatives (dexmedetomidine, 2112311, Hameln Pharmaceuticals BBraun Melsungen AG, Spain), (propofol, 211138071, Hameln Pharmaceuticals BBraun Melsungen AG, Germany). Discontinuation of mechanical ventilation was considered after two successful consecutive spontaneous breathing trials.

At the randomization, the informed consent was signed by two study-independent physicians in unconscious patients and obtained directly from the patients conscious enough to approve the study participation. The study physician collected the baseline, clinical, and outcome characteristics in all randomized patients (see Table 3) and stored them in the secure analog database.

Muscle strength was assessed in both groups by study rehabilitation worker at the two-time points: the time of randomization and the end of rehabilitation when each patient was discharged from the intensive care unit. The medical research council manual muscle testing scale was used [18]. The strength was measured by a dynamometer (microFET2, Hoggan Scientific LCC, Salt Lake, UT, USA). The wrist flexors on the upper extremity and the knee extensors were evaluated three times, 4 seconds each measurement, and then the average was calculated.

Rehabilitation was carried out twice a day in both groups. The standard rehabilitation included: (1) chest physiotherapy encompassing manual therapy, ventilation support (contact breathing, activation of diaphragm, reflex stimulation of breathing), and airway clearance techniques (vibration, breathing exercises on the side or in sitting position, cough assistance, instrument support-vest airway clearance system), and (2) functional mobility training using recondition (reeducation of the range of motion and muscle strength, coordination and endurance training), mobilization according to the mobility protocol. In passive patients, we performed a passive exercise to prevent secondary changes. In-bed cycling was performed once daily only in the m-group 5 days a week except for Saturday and Sunday. The cycling device was the MOTOmed Letto 2 (RECK–Technik GmbH, Betzenweiler, Germany). The therapy was targeted to at least 10 minutes of active cycling in (1) servo mode, (2) without resistance, or (3) against dosed resistance, depending on their current ability in the m-group [19].

Before release from the intensive care unit, a cycling test was performed in both groups (10 minutes in the primary mode without resistance), as was the barthel index for activities of daily living.

2.4 Ethics and outcomes

The code of ethics of the world medical association (Declaration of Helsinki) for human subjects experiments was pursued.

We followed (1) overall length of mechanical ventilation (number of ventilator days) and mechanical ventilation length from the randomization, (2) change in intensity of muscle strength and speed of its recovery, (3) presence of adverse alterations of vital (respiratory and cardiovascular) parameters related to use of cycling device.

2.5 Statistical analysis

The sample was characterized using descriptive statistics (the median, arithmetic mean, standard deviation, frequency tables). The distribution of quantitative data was tested for normality by using shapiro-wilk normality test, data was not normal distributed. Selected characteristics (methods) were compared between m-group and c-group using the nonparametric test. Mann-whitney was used for testing the quantitative data and chi-squared test for qualitative data or, where the criteria for chi-squared test using not accepted, fisher’s exact test was performed. The data were graphically depicted as box plots. The level of significance was set at 5%. The analyses were performed by certified statistician using Stata 13 software application.

3.1 Primary outcome

The mechanical ventilation length did not significantly differ between the m-group and the c-group (21.0 ± 9.78 vs. 24.4 ± 10.88 days, respectively, p = 0.915). The mechanical ventilation length from randomization to the discontinuation of mechanical ventilation did not significantly differ between groups (5.4 ± 6.17 vs. 8.7 ± 8.00 days, respectively, p = 0.860; Fig. 2).

Fig. 2.Comparison of total length of mechanical ventilation in m-group and c-group and of length of mechanical ventilation from randomization in m-group and c-group. m-group—MOTOmed group; c-group—control group; MV—mechanical ventilation; MVRAND—mechanical ventilation from randomization.

3.2 Secondary outcomes

M-group and c-group patients’ dynamometric values did not significantly differ in both measured groups of muscles (knee joint extensors, wrist flexors) at the time of randomization. However, comparing values at randomization, we observed a significant improvement in muscle strength at the time of ICU discharge in both groups. Dynamometric evaluation of knee joint extension revealed a significant improvement in the m-group respondents in both the left (LLE) and right lower extremity (RLE) (p = 0.004 and 0.015, respectively). Apart from the RUE (p = 0.576), the same significant improvement in muscle strength was observed in handgrip values: in the LUE in the m-group (p = 0.004) and both upper extremities (LUE and RUE) in the c-group (p = 0.011 and 0.017, respectively), see Table 4. The m-group patients reached the same muscle strength level in a non-significantly shorter time (8.70 ± 6.44 vs. 6.8 ± 4.3 days, p = 0.534). The result was achieved even when the interval of rehabilitation absence was significantly longer in the m-group (3 ± 2.78 vs. 0.5 ± 1.55 days, p = 0.050). The average duration of physiotherapy lasted 18 ± 12.94 days in the m-group compared to 28 ± 37.91 days in the c-group (p = 0.103).

We found no adverse alteration in physiological respiratory and cardiovascular parameters (respiratory rate, pulse oximetry, heart rate, mean arterial blood pressure, frequency of heart arrhythmias) in all patients from m-group, leading to procedure interruption.

The m-group patients had non-significantly higher scores on the Barthel test than the c-group (35 ± 19.1 vs. 33 ± 24.04, p = 0.610). The ICU length of stay was non-significantly shorter in the m-group than in the c-group (25 ± 38.9 days and 28 ± 11.0 days, respectively, p = 0.813). No death was observed in the m-group, while the ICU mortality rate was 8 % (p = 0.234) in c-group.

4.1 Mechanical ventilation and in-bed cycling

Our study belongs to a few interventional studies reporting the influence of in-bed leg cycling on the length of mechanical ventilation in critically ill patients. The most similar trial was performed by Yu et al. [20], enrolling the patients almost simultaneously. They found a statistically significant difference in ICU length of stay, the average length of mechanical ventilation, and the Barthel index favored the intervention group [20]. Our study showed a beneficial effect of in-bed leg cycling exercise in the same parameters. Compared the methodology of the Yu et al. [20] study, the reasons for our non-significant results were most probably as follows: (1) smaller number of patients enrolled, as the main factor, (2) our patients were almost all intubated before admission to ICU, which might bias the results due to the unknown possible adverse effect of previous treatment strategy, and (3) the length of mechanical ventilation in Yu et al. [20] study was almost twice shorter than in our study (8 to 10 days vs. 21 to 24 days, respectively) implying that we enrolled more ill patients possibly influencing the results. Machado et al. [21] evaluated MOTOmed-based passive cycling in a randomized fashion in mechanically ventilated patients. In consent with our results, they found improvement in muscle strength in both study groups, with a more pronounced increase in the intervention group. They also analyzed data from a small number of patients (n = 36) and did not find a significant mechanical ventilation length difference. They even observed that the length of mechanical ventilation was three days longer in the intervention group, while our results showed 3.4 days short interval.

Schweickert et al. [15] pursued the efficacy of combining daily interruption of sedation with physical and occupational therapy on functional outcomes of mechanically ventilated intensive care patients (n = 104). Among other results, they found significantly more ventilator-free days on the 28th day of hospital stay in a study group with daily interruption and physical and occupational activity (2.5 days difference). They also found a significant reduction in the overall length of mechanical ventilation. Compared to our results, the length of mechanical ventilation in the study was almost five times shorter.The possible explanation of why the authors found a significant difference in mechanical ventilation length might include the daily interruption of sedation or that the study included less severely ill patients with better rehabilitation ability.

4.2 Cycling and non-mechanical ventilation outcome

Pires-Neto et al. [10] evaluated physiological changes and safety of an earlier cycling intervention (<72 hours of mechanical ventilation) in critically ill patients. They found that early passive cycling exercise in sedated, critically ill, mechanically ventilated patients was not associated with significant alterations in hemodynamic, respiratory, or metabolic variables [10]. Our study also did not observe any adverse changes in physiological variables leading to in-bed cycling premature interruption.

In consent with other studies [21], we demonstrated a significant favorable influence of early physiotherapy on muscle strength improvement. The isometric handgrip, elbow joint flexion, and knee joint extension methods using portable dynamometers were used to evaluate muscle strength. It has been demonstrated that dynamometry can provide reliable measurements in critically ill patients. However, it should be considered that moderate changes in strength may be required to overcome measurement error during the acute recovery period [22].

In actual clinical practice, various respiratory physiotherapy elements are used to activate weakened respiratory muscles and help with airway hygiene, thereby supporting ventilation and contributing to the patient’s early disconnection from mechanical ventilation. The long-term rehabilitation plan included trunk stabilization exercises; passive, active, and resistance exercises with a thera-band; cycling exercises; sitting and standing exercises; walking between parallel bars; and walking upstairs, all five times a week for 30–60 minutes. After rehabilitation, significant improvements were noted in increased muscular strength in the upper and lower extremities, improved locomotion, and walking up and down the stairs (and time to disconnect from mechanical ventilation) [23]. In a prospective, randomized, controlled study, Chih-Cheng et al. [24] investigated the effects of 6 weeks of respiratory exercise and upper and lower limb exercise in patients with long-term mechanical ventilation. The program was developed as a 5-week workout, focusing on respiratory muscle exercises, limb exercises, functional training, and walking. In the rehabilitated group, functional status and muscle strength improved significantly compared with the control group, in which an apparent decrease in both values was observed. A shorter time to disconnecting from mechanical ventilation was recorded in the intervention group [24]. Our data show that daily addition of in-bed cycling physiotherapy to standard physiotherapy increases muscle strength to a similar degree in both groups but in a shorter time than routine rehabilitation.

4.3 Study limitations

Our study’s limitations included the small number of subjects, which decreases the results’ generalizability and limits the chance to prove statistical significance. However, the addition of in-bed cycling exercise to standard physiotherapy showed an unequivocal trend to benefit in all followed end-points (length of mechanical ventilation, muscle strength, Barthel index for activities of daily living, ICU length of stay) without significant physiologic adverse effects. The loss of subjects to follow-up also belongs to study limitations, further decreasing sample size and potentially introducing bias.