To compare the safety of conventional physiotherapy alone versus its combination with cycloergometry by analysing session interruptions and physiological tolerance in critically ill patients. Secondarily, efficacy was assessed through strength and functional related outcomes.
DesignSingle-centre, parallel, two-arm, randomized clinical trial.
SettingIntensive Care Department.
ParticipantsMechanically ventilated patients.
InterventionControl group received 30-min of conventional physiotherapy; intervention group received 15-min of cycloergometry and 15-min of conventional physiotherapy.
Main variables of interestSafety was evaluated by recording session interruptions and changes in blood pressure, heart rate, respiratory rate, SpO2, FiO2 and tidal volume before and after sessions. Muscle strength (modified Medical Research Council score, quadriceps and handgrip strength) was evaluated at first cooperation of participants, ICU discharge, 28-day and 6-month follow-up; Activities of Daily Living score and mobility scale at ICU discharge, 28 days and 6 months; and six-minute walking test and Short Form-36 at 28 days and 6 months.
Results46 participants completed 732 sessions. Both interventions produced significant but comparable physiological changes. Cycloergometry sessions were longer (30 vs. 25 min, p < 0.001) and had more interruptions (13% vs. 7%, p = 0.008), mainly due to fatigue and lack of cooperation. With the applied methodology no significant differences were observed in muscle strength or functional outcomes at any time point.
ConclusionPartially replacing conventional physiotherapy with cycloergometry was safe and well tolerated in critically ill patients. However, due to methodological limitations and the small sample size, no firm conclusions regarding efficacy can be drawn.
Comparar la seguridad de la fisioterapia convencional frente a su combinación con cicloergometría analizando las interrupciones de sesiones y la tolerancia fisiológica en pacientes críticos. Adicionalmente se evaluó la eficacia mediante variables de fuerza muscular y resultados funcionales.
DiseñoEnsayo clínico unicéntrico aleatorizado con dos ramas.
ÁmbitoServicio de Medicina Intensiva.
ParticipantesPacientes con ventilación mecánica.
IntervenciónGrupo control: 30-min de fisioterapia convencional. Grupo intervención: 15-min de cicloergometría y 15-min de fisioterapia convencional.
Variables principalesLa seguridad se evaluó registrando las interrupciones de las sesiones y los cambios en presión arterial, frecuencia cardíaca, frecuencia respiratoria, SpO₂, FiO₂ y volumen corriente antes y después de cada sesión. La fuerza muscular (MRCm, cuádriceps y handgrip) se valoró el primer día evaluable, al alta de UCI, a los 28-días y a los 6-meses. Las escalas de Actividades de la Vida Diaria y de movilidad se evaluaron al alta de UCI, a los 28-días y a los 6-meses; la prueba de la marcha de 6-min y el Short Form-36 a los 28-días y a los 6-meses.
Resultados46 participantes completaron 732 sesiones. Ambas intervenciones produjeron cambios fisiológicos significativos pero comparables. Las sesiones con cicloergometría fueron más largas (30 vs. 25 minutos; p < 0,001) y presentaron más interrupciones (13% vs. 7%; p = 0,008), principalmente por fatiga o falta de cooperación. Con la metodología aplicada no se detectaron diferencias significativas en fuerza muscular ni en resultados funcionales.
ConclusiónLa sustitución parcial de la fisioterapia convencional por cicloergometría fue segura y bien tolerada. Debido a limitaciones metodológicas no pueden extraerse conclusiones firmes sobre su eficacia.
Intensive Care Unit (ICU) acquired weakness is a complication affecting approximately 25–50% of patients.1,2 This condition significantly prolongs the duration of mechanical ventilation and hospital stays,1 as well as leading to a diminished quality of life, incomplete functional recovery, and notable sequelae upon hospital discharge.3–5 This entity is included within the so-called post-ICU syndrome (PICS). Physiotherapy plays a pivotal role in mitigating these issues and facilitating rehabilitation. The initiation of physiotherapy in the early phase of admission to the ICU is widely used6–9 and it is associated with better functional recovery at hospital discharge.10–12
In recent years, there has been growing interest in the incorporation of cycloergometry as a therapeutic approach within critical care physiotherapy. This intervention is based on the use of a stationary bicycle while subjects are in supine position and can be adapted to suit the subject's level of fitness and functional capacity, even in sedated patients. This device provides a controlled environment for aerobic training, facilitating a range of passive, active, and resisted in-bed exercises, thereby contributing to the preservation of muscle architecture,13,14 and it can be easily integrated into conventional physiotherapy protocols in ICU. Patient acceptance of in-bed cycling sessions has been positive15 and it encourages subjects to engage in physical activity.16
Published meta-analyses suggest that cycloergometry could be safe in critically ill patients, though the certainty of evidence remains low.17,18 Given the physiological instability of these subjects, careful subject selection and close monitoring are essential. While the utilization of these devices for ICU mobilization has been suggested to lead to improved functional recovery,19,20 there is ongoing debate due to controversial findings in published studies.14,21–26 For this reason, ensuring the safety of cycloergometry in critically ill patients remains essential. Further high-quality studies are needed to assess safety and effectiveness.
The aim of this study was to evaluate the safety of adding early cycloergometry to conventional physiotherapy as a rehabilitation strategy in critically ill patients, compared with early conventional physiotherapy alone, by examining session interruptions, potential associated risks and changes in physiological variables. In addition, participants’ functional status and muscle strength were assessed as secondary objective at different time points from the ICU stay throughout a 6-month follow-up for both treatment groups as measures of efficacy.
Patients and methodsStudy designThis single-centre, parallel, open-label, two-arm, randomized-controlled trial with critically ill patients with invasive mechanical ventilation was conducted in the Intensive Care Medicine Department in collaboration with the Physical Medicine and Rehabilitation (RHB) Department of the University Hospital Son Llatzer in Palma, Spain.
Participants and settingsThe study was conducted in a 16-bed, mixed medical and surgical ICU at a university hospital between June 2015 and February 2020. Patients of both sex were eligible if they were older than 18 years and were mechanically ventilated more than 48 h, regardless of sedation status, had functional independence prior to admission, and subjects or relatives had signed the informed consent. Patients were excluded if they met any of the following criteria: unable to adhere to follow-up (e.g., tourist), pregnancy, neuromuscular disease, an estimated fatal outcome within 48 h, inability to perform pedalling movements, active haemorrhage or a platelet count <50 × 109/L, psychiatric disorders, severe agitation, hemodynamic instability despite requiring noradrenaline > 0.5 mcg/kg/min, or admission due to cardiorespiratory arrest.
Randomization and treatment groupsWithin the first 3 days after ICU admission subjects were randomly allocated in a 1:1 ratio in two groups based on the type of physiotherapy session administered by means of a computer-generated randomization list in block sizes of 6 subjects. Treatment allocation was concealed using opaque sealed and numbered envelopes. All sessions were delivered by licensed physiotherapists who had experience with ICU environment and cycloergometric devices and were not blinded to the randomization group.
Control Group consisted of the application of conventional physiotherapy, as the usual care in our unit. Physiotherapists administered 30-minute session of physiotherapy (manual motor physiotherapy for the extremities and respiratory physiotherapy, if needed) on weekdays (5 sessions per week) that involved daily assessment of session intensity, which included phases of passive, passive/active, active, or active/resisted physiotherapy.
Intervention Group consisted of the application of 15-minute session of cycloergometric mobilization by physiotherapists with a motor-assisted bed-cycle (MOTOmed® Letto 2, RECK-Technik GmbH & Co. KG, Betzenweiler, Germany) with individualized intensity adjustments, following the methodology used in randomized clinical trials.26 This device allowed for active or passive leg pedalling exercises at six progressively challenging levels. Patients were positioned comfortably in a supine or reclining posture. Sedated patients engaged in passive leg pedalling for 15 min at a rate of 20 revolutions per minute.20,25 The remaining 15 min consisted of our standard care session involving arms and respiratory system, if needed.
Variables and measurementsDescriptive variables collected in the study included age, sex, weight, height, body mass index, comorbidities, Activities of Daily Living (ADL) score27 one month before hospital admission, main diagnosis, type of patient (medical or surgical), SAPS3 score28 and treatment received in ICU. For each session, a qualitative assessment of the intensity applied was recorded, classifying sessions as passive, passive/active, active, or active/resisted.
For the safety analysis, we documented the real-time duration of each session and prospectively recorded physiological variables immediately before and at the end of each session, regardless of whether it was completed or interrupted, without allowing for recovery time. Throughout all mobilization sessions, we continuously monitored blood pressure (BP), heart rate (HR), respiratory rate (RR) and pulse oximetry oxygen saturation (SpO2). For the analysis, we defined poorly tolerated sessions as those exhibiting any of the following abnormal physiological responses: HR exceeding 140 bpm or falling below 40 bpm, systolic BP exceeding 180 mm Hg or decreasing by 20% from baseline, SpO2 falling below 90% despite adjustment of FiO2 or RR exceeding 35 bpm. Additionally, we noted any instances of session interruption and, when applicable, any disconnection of ventilator or devices dislodgement (catheters or tubes). A session was promptly halted if the patient had the presence of malignant arrhythmias or signs of myocardial ischemia. Physiotherapists also interrupted sessions when clinical derangement, agitation or at the patient’s request due to poor cooperation or fatigue, in accordance with procedures used in other clinical trials.29
Treatment effects on muscle strength were assessed with the Medical Research Council (MRC) score.30 In addition, the maximum isometric strength of both quadriceps and the maximum handgrip force from the dominant hand were assessed using dynamometers, based on three repeated measurements per limb. These measurements were performed as soon as patients were capable of active cooperation, at ICU discharge and at 28 days and 6 months after hospital discharge.
Regarding functional status, the ADL score27 and the ICU Mobility Scale31 were assessed at ICU discharge and at 28 days and 6 months after hospital discharge. The 6-minute walk test (6MWT)32 and the SF-3633 questionnaire were assessed at 28 days and 6 months after hospital discharge.
Sample size calculationThe primary objective of this study was to evaluate safety. For the safety analysis, the sample size calculation assumed a 2% prevalence of interrupted physiotherapy sessions,29 considering a prevalence above 4% as relevant increase. Assuming a type I error of 5% and a statistical power of 80%, and a loss percentage of 5%, a total of 687 physiotherapy sessions were needed.
For the secondary objective based on efficacy of cycloergometry, 36 subjects were required in each treatment group to demonstrate a minimally clinically significant difference in the 6MWT (50 m),20,34 with a statistical power of 80% and a type I error of 5%.
Statistical analysisCategorical variables are expressed as frequency and percentages. Continuous variables are expressed as mean and standard deviation (SD), or as median and interquartile range (IQR), depending on whether they present normal distribution or not. Normality was assessed using the Shapiro–Wilk test. Categorical variables were compared using the χ2 test. For continuous variables, the independent samples were analyzed using the Student’s t-test or the non-parametric Mann–Whitney U-test, according to normality. For paired samples, comparisons were performed using the paired Student’s t-test or the Friedman test, depending on data distribution. A difference with a significance level equal to or less than 0.05 was be considered significant.
EthicsThe study has been approved by the institutional research ethics committee prior to its initiation with the protocol number IB 2565/15 of the Balearic Ethical Committee and is registered on clinicaltrials.gov under the identifier NCT02478411. The study has been conducted in accordance with the principles of the Declaration of Helsinki. The signed informed consents from all participants or participant relatives were obtained.
ResultsParticipants characteristicsThe flowchart of the participants enrolment is shown in Fig. 1a. Among the 634 patients screened for eligibility, 388 did not meet the inclusion criteria, 18 declined participation and 157 were not included for unavailability of resources or recruitment oversight. In total, 71 participants were randomized, and 68 (68% women) received the allocated treatment, 34 participants in control group and 34 participants in intervention group, with a median age of 66 (56–77) years. Ultimately, 24 subjects in the control group and 22 in the intervention group were included in the efficacy analysis.
Eight-hundred six physiotherapy sessions were scheduled for the 68 randomized subjects, 394 (49%) allocated in the control group and 412 (51%) in the intervention group. Thirty-one sessions (8%) in the control group and 43 (10%) in the cycloergometric group were not initiated (p = 0.22) due to clinical reasons before the start of the session as prone position, fever, hemodynamic instability, need for medical procedures (e.g., pleural drainage, tracheostomy, CT scan…), agitation, or physiotherapist workload, as shown in Fig. 1b. Subjects received a median of 7 (5–16) sessions during the ICU stay, with a median duration of sessions of 27 (20–30) min. Sessions were started at a median of 3 (2−5) days from ICU admission. Baseline characteristics according to treatment group are summarized in Table 1.
Demographic and clinical characteristics of subjects according to treatment group allocation.
| Control group (n = 34 patients) | Intervention group (n = 34 patients) | p | |
|---|---|---|---|
| At ICU admission: | |||
| Age (years), median (IQR) | 71 (55−79) | 64 (57−76) | 0.67 |
| Female sex, n (%) | 24 (71) | 22 (65) | 0.80 |
| BMI (kg/m2), mean (SD) | 29.5 (4.5) | 28.4 (6.7) | 0.41 |
| SAPS 3 score (points), mean (SD) | 61 (13) | 61 (13) | 0.83 |
| ADL score (points), median (IQR) | 6 (6−6) | 6 (6−6) | 0.78 |
| Type of patient: | |||
| Medical, n (%) | 31 (91) | 29 (85) | 0.71 |
| Surgical, n (%) | 3 (9) | 5 (15) | |
| Main diagnosis: | |||
| Pneumonia, n (%) | 22 (65) | 16 (47) | |
| Septic shock, n (%) | 7 (21) | 6 (18) | |
| Encephalopathy, n (%) | 3 (9) | 2 (6) | |
| Abdominal sepsis, n (%) | 2 (6) | 4 (12) | |
| COPD exacerbation, n (%) | 0 | 4 (12) | |
| Polytrauma, n (%) | 0 | 1 (3) | |
| Retropharyngeal hematoma, n (%) | 0 | 1 (3) | |
| During hospital admission: | |||
| Tracheostomy, n (%) | 11 (32) | 13 (38) | 0.80 |
| Number of sessions per patient, median (IQR) | 8 (4−13) | 7 (6−17) | 0.87 |
| Days between ICU admission and 1st session, days, median (IQR) | 3 (2−4) | 3 (2−5) | 0.52 |
| Days of IMV, median (IQR) | 10 (7−20) | 11 (7−23) | 0.94 |
| ICU LOS (days), median (IQR) | 17 (12−37) | 16(11−37) | 0.97 |
| Hospital LOS (days), median (IQR) | 36 (18−50) | 26(18−51) | 0.65 |
| Hospital mortality, n (%) | 10 (31) | 11 (32) | 0.92 |
Regarding the clinical condition of participants before the start of each session, no significant differences were observed between the two groups in FiO2 levels, the need for vasopressors, or the requirement for invasive mechanical ventilation. Continuous renal replacement therapy was more frequently used in the control group sessions (p < 0.001), whereas tracheostomy was more common in the intervention group sessions (p = 0.009). There were no significant differences between groups in the qualitative classification of session intensity. Session duration was significantly shorter in the control group compared to the cycloergometry group [25 (20–28) min vs. 30 (25–33) min, p < 0.001] (Table 2).
Clinical characteristics of physiotherapy sessions based on treatment group.
| Control group (n = 363 sessions) | Intervention group (n = 369 sessions) | p | |
|---|---|---|---|
| FiO2 at start of session, median (IQR) | 28 (24−35) | 30 (24−35) | 0.45 |
| Need of vasopressors, n (%) | 149 (41) | 127 (34) | 0.07 |
| Connected to CRRT, n (%) | 38 (11) | 7 (2) | <0.001 |
| Invasive mechanical ventilation, n (%) | 186 (65) | 214 (59) | 0.19 |
| Tracheostomy, n (%) | 108 (52) | 154 (65) | 0.009 |
| Session Intensity,a median (IQR) | 2 (1−2) | 1 (1−2) | 0.175 |
| Duration of sessions, minutes, median (IQR) | 25 (20−28) | 30 (25−33) | <0.001 |
| Presence of at least one poor tolerance criteria,b n (%) | 37 (10) | 53 (14) | 0.07 |
CRRT: continuous renal replacement therapy; IQR: Interquartile range; FiO2: fraction of inspired oxygen.
Overall, at least one sign of poor physiological tolerance was detected in 37 sessions (10%) in the control group and 53 sessions (14%) in the cycloergometry group (p = 0.07) (Table 2), affecting 16 subjects (47%) in the control group and 19 subjects (56%) in the cycloergometry group (p = 0.47). However, the presence of poor tolerance signs did not necessarily lead to session discontinuation. Of the 90 sessions in which at least one sign of poor tolerance was recorded, only 8 sessions (22%) in the control group and 21 sessions (40%) in the cycloergometry group were interrupted (p = 0.07).
Globally, session interruptions were more frequent in the intervention group, occurring in 13% of cases (47 sessions) compared to 7% (25 sessions) in the control group (p = 0.008) (Table 3). The primary reasons for session interruptions, detailed in Table 3, were patient fatigue and lack of cooperation.
Characteristics of interrupted sessions.
| Control group (n = 363 sessions) | Intervention group (n = 369 sessions) | p | |
|---|---|---|---|
| Number of interrupted sessions, n (%) | 25 (7) | 47 (13) | 0.008 |
| Duration of interrupted sessions, median (IQR) | 15 (12−17) | 15 (12−20) | 0.61 |
| Main reason of session interruption: | |||
| Poor-cooperation, n (%) | 11 (44) | 12 (26) | |
| Fatigue, n (%) | 7 (28) | 16 (34) | |
| SpO2 < 90%, n (%) | 3 (12) | 7 (15) | |
| Pain, n (%) | 1 (4) | 7 (15) | |
| Hypotension, n (%) | 3 (12) | 3 (6) | |
| Hypertension, n (%) | 0 | 1 (2) | |
| Removal rectal tube, n (%) | 0 | 1 (2) |
IQR: interquartile range.
In terms of the physiological impact of physiotherapy sessions, both groups resulted in significant changes in heart rate, respiratory rate, and tidal volume during sessions in each studied group (Table 4). Moreover, significant variations in SpO2 were observed during conventional physiotherapy (p = 0.01), whereas intervention group sessions resulted in changes in mean blood pressure (p = 0.007). Despite these differences, the overall percentage change in physiological parameters at the end of the sessions was similar between groups (Table 4). No episodes of malignant arrhythmias or signs of myocardial ischemia were recorded.
Physiologic variables at the start and at the end of all physiotherapy session according to treatment group.
| Control group (n = 363 sessions) | Intervention group (n = 369 sessions) | |||||
|---|---|---|---|---|---|---|
| Start of session | End of session | p | Start of session | End of session | p | |
| MBP (mm Hg), median, (IQR) | 92 (79−101) | 91 (78−102) | 0.44 | 86 (77−97) | 88 (77−99) | 0.007 |
| HR (bpm), median, (IQR) | 89 (78−105) | 92 (79−103) | 0.002 | 91 (81−102) | 94 (82−105) | 0.001 |
| RR (rpm), median (IQR) | 22 (17−28) | 23 (18−29) | 0.01 | 25 (19−30) | 25 (20−31) | 0.006 |
| SpO2 (%), median (IQR) | 97 (95−99) | 97(95−99) | 0.01 | 97 (95−99) | 97 (95−98) | 0.06 |
| FiO2 (%), median (IQR) | 28 (24−35) | 28 (24−35) | 0.38 | 30 (24−35) | 30 (24−35) | 0.80 |
| Tidal volume (mL), median (IQR) | 489 (395−624) | 513 (411−625) | 0.003 | 495 (402−609) | 507 (405−650) | 0.01 |
| Percentage of change at end of session: | Control group | Intervention group | p |
|---|---|---|---|
| ΔMBP (%), median, (IQR) | 0 (-5.5–5.9) | 1.9 (-5.9–9.4) | 0.06 |
| ΔHR (%), median, (IQR) | 1 (-2.4–4.4) | 1.2 (-3.4–6.4) | 0.50 |
| ΔRR (%), median (IQR) | 0 (-8.5–15.8) | 0 (-7.6–16.7) | 0.97 |
| ΔSpO2 (%), median (IQR) | 0 (-1.1 – 1.0) | 0 (-1.0–1.0) | 0.52 |
| ΔFiO2 (%), median (IQR) | 0 (0 – 0) | 0 (0 – 0) | 0.45 |
| ΔTidal volume (%), median (IQR) | 1.5 (-4.3–11.5) | 1.3 (-6.3–13.7) | 0.88 |
MBP: mean blood pressure; HR: heart rate; RR: Respiratory rate; SpO2: O2 saturation by pulsioximetry; FiO2: oxygen inspired fraction; Δ: change of variables from the start to the end of sessions measured as: (end value − initial value)/initial value) × 100. IQR: Interquartile range.
With the methodology employed in our study no significant differences between the groups were observed with respect to weakness (as assessed by the MRC scoring system, quadriceps strength and handgrip strength) and functional recovery (as measured by the ADL score, 6MWT, mobility scale and SF-36 score) in each follow-up point (Table 5).
Effects on muscle strength and functional recovery over time according to treatment group.
| At first patient cooperation | p | UCI discharge | p | 28 days after hospital discharge | p | 6 months after hospital discharge | p | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Control (n = 16) | Intervention (n = 18) | Control (n = 12) | Intervention (n = 16) | Control (n = 18) | Intervention (n = 17) | Control (n = 12) | Intervention (n = 8) | |||||
| MRC score, median (IQR) | 40.5 (33−51) | 43.5 (39.0−53.3) | 0.30 | 48 (46−56) | 53.0 (46.5−56) | 0.64 | 58 (53−60) | 56 (48−60) | 0.77 | 60 (58−60) | 55 (53−60) | 0.06 |
| Quadriceps strength (Kg), median (IQR) | 9.8 (6.0−14.9) | 11.0 (8.1−17.1) | 0.52 | 10.1 (8.2−12.6) | 9.3 (7.7−13.1) | 0.87 | 15.5 (12.7−20.9) | 16.6 (14.9−19.3) | 0.66 | 16.3 (11.8−19.4) | 18.5 (15.4−21.3) | 0.57 |
| Handgrip strength (Kg) median (IQR) | 11.6 (4.8−22.2) | 13.9 (8.0−23.1) | 0.30 | 13.3 (12.7−19.3) | 13.8 (8.9−23.9) | 0.77 | 27.9 (15.6−35.3) | 24.1 (15.8−31.8) | 0.48 | 29.9 (17.5−42.6) | 33.1 (23.9−44.0) | 0.43 |
| ADL, median (IQR) | – | – | 1 (0−4) | 2 (1−4) | 0.08 | 6 (5−6) | 5 (4−6) | 0.22 | 6 (6−6) | 6 (6−6) | 0.65 | |
| Mobility scale, median (IQR) | -- | -- | 5 (4.25−7) | 5 (4−8) | 0.92 | 10 (10−10) | 10 (9.3−10) | 0.46 | 10 (10−10) | 10 (9−10) | 0.22 | |
| 6MWT (minutes), median (IQR) | – | -- | -- | -- | 296.5 (227.5−419.5) | 330.0 (132.5−428.3) | 0.75 | 416.5 (344.3−470.25) | 360.0 (195−437.5) | 0.25 | ||
| SF-36, median (IQR) | -- | -- | -- | -- | 53.4 (26.4−73.3) | 38.64 (15.9−75.0) | 0.19 | 86.4 (78.4−93.2) | 72.7 (45.5−86.4) | 0.06 | ||
Analysing the evolution within each group using repeated-measures analyses, it was found that control and intervention groups showed a progressive improvement over time in ADL (p < 0.001 and p < 0.0001, respectively), mobility scale (p < 0.0001 and p = 0.015, respectively), SF-36 score (p = 0.035 and p = 0.034, respectively), and handgrip strength (p = 0.017 and p = 0.006, respectively). However, for MRC score, 6MWT, and quadriceps strength, although improving over time in both groups, only the control group exhibited significant gains (p = 0.002, p = 0.007, p = 0.047, respectively) (Table 5).
DiscussionThis study evaluated the safety and physiological tolerance of early rehabilitation using either conventional physiotherapy alone or a combination of conventional physiotherapy and cycloergometry in critically ill patients. When delivered for similar total duration, both interventions showed comparable safety profiles, with no serious adverse events and similar physiological changes after sessions, although sessions in the cycloergometry group were interrupted more frequently. Although no statistically significant differences in efficacy outcomes were identified, these findings must be interpreted with caution given the methodological constraints of the trial.
Safety remains a central concern in early rehabilitation given the vulnerability of participants as the TEAM study35 reported a higher incidence of adverse events in participants receiving early physiotherapy compared with usual care physiotherapy in critically ill patients. In that trial, longer sessions duration and sedation minimization in the intervention group may have contributed to these results, acting as potential confounders. Additionally, although cycloergometry has been associated with high subject satisfaction and increased motivation,15,16 its safety profile must be further evaluated and closely monitored given the frailty of this population.
In our study, the systematic assessment of physiological variables immediately before and after every session in both study groups enabled a rigorous and detailed characterization of safety events. This approach likely explains the higher overall incidence of safety events observed compared to previous reports that have used broader definitions or less intensive monitoring approaches.15,18,21,24,25,29 Importantly, most events were mild and did not require session interruptions.
In this context, the addition of cycloergometry to conventional physiotherapy was associated with a higher rate of session interruptions than conventional physiotherapy alone, resulting in an overall discontinuation frequency greater than previously described.29 However, no serious adverse events occurred, supporting the feasibility and safety of both interventions when implemented with strict physiological monitoring, as fatigue or lack of cooperation were main reasons for session discontinuation, in line with previous reports.29
Regarding efficacy, although previous randomized trials have reported potential benefits of cycloergometry such as reduced muscle autophagy14 or improvements in physical function in terms of better walking capacity,20,36 greater quadriceps strength20 or higher MRC scores,36,37 these findings have not been consistently replicated,21,24–26 leaving the overall evidence inconclusive. A recent meta-analysis18 including 33 randomized controlled trials and 3274 critically ill patients concluded that cycloergometry may improve physical function and reduce ICU and hospital length of stay, but it showed no significant effects on other clinically relevant outcomes, including mortality. Despite these apparently beneficial findings, the overall quality of evidence remains low or very low limiting the strength and generalizability of these findings.
Although our study aimed to strengthen the existing evidence regarding efficacy with cycloergometry and included a relatively high number of sessions per patient compared with previous reports,21,24–26 methodological aspects may have limited the ability to detect clinically meaningful differences. Consequently, the absence of statistically significant effects should be interpreted with caution.
First, the modest sample size represents an important limitation, as it may have resulted in insufficient statistical power to identify clinically relevant differences in efficacy between groups. This limitation arises from several practical and methodological challenges encountered during the study. Subject recruitment was slow due to the limited availability of a single cycloergometer in our department and the small number of physiotherapists trained in cycloergometry. Additionally, loss of subjects during follow-up primarily due to mortality, but also to non-attendance at outpatient assessments, further limits the robustness of our long-term findings regarding the efficacy of physiotherapy. Second, cycloergometry sessions were shorter than in comparable studies20,24,37 due to feasibility constraints. We cannot rule out that 15 min may represent a suboptimal dose of physiotherapy to induce measurable functional or strength enhancements, and that this limitation restricts the conclusions that can be drawn regarding efficacy. In many randomized trials, physiotherapy intensity has been imbalanced between groups, as patients allocated to cycloergometry often also receive conventional physiotherapy,14,21,24–26,37 increasing total intervention time and potentially confounding the results. In contrast, our trial maintained comparable session durations between groups, allowing for a clearer assessment of the independent effect of cycloergometry but the duration of the intervention was likely insufficient. Third, both interventions were delivered at low intensity, with median levels ranging from passive to passive–active. Although trials comparing different intensities of the same physiotherapy modality have not found significant differences in long-term functional outcomes,38,39 given that low workloads in our study may be insufficient to stimulate functional recovery or prevent muscle atrophy, the intervention dose may have been inadequate. Moreover, assessing session intensity is inherently challenging, and the lack of detailed workload metrics in our study may have introduced some heterogeneity between groups, further limiting the interpretation of the negative efficacy results. Finally, frequent session interruptions in the cycloergometry group reduced the effective intervention time and may have attenuated any potential benefit. The relatively brief stimulus,21,24–26,40 combined with frequent session interruptions, may have limited the intervention’s impact on muscle preservation and functional outcomes. Furthermore, subjects with limited tolerance may not have completed enough effective sessions to experience the theoretical advantages of cycle ergometry. Together, these factors restrict conclusions about efficacy and highlight the need for caution when interpreting the absence of significant differences.
In summary, both conventional and cycloergometric early physiotherapy were safe and physiologically well tolerated. Because of the uncertainty introduced by the methodological limitations of this trial, including restricted intervention duration, concerns on intensity quantification and insufficient statistical power, the reliability of the observed efficacy results cannot be ensured.
ConclusionsPartially replacing conventional physiotherapy with cycle ergometry proved safe and physiologically well tolerated in critically ill patients. Given the methodological limitations related to intervention dose, intensity quantification, and the restricted sample size, efficacy outcomes cannot be interpreted with confidence, and no reliable conclusions about the effects on treatment efficacy can be drawn.
Author’s contributionsGemma Rialp and Catalina Forteza have made a substantial contribution to the study design, data acquisition, analysis and interpretation, as well as to the drafting of the manuscript.
Isabel Gil contributed to patient management, data acquisition, and the revising of the manuscript.
Maria Romero contributed to data acquisition, data analysis and interpretation, and the revising of the manuscript
Catalina Morey has made a substantial contribution to the study design and the revising of the manuscript.
Fiorella Sarubbo has made a substantial contribution to data analysis and interpretation, as well as to the drafting of the manuscript.
All authors read and approved the final version of the manuscript.
Declaration of Generative AI and AI-assisted technologies in the writing processArtificial intelligence tools were employed exclusively to refine the language and improve the clarity of expression in the manuscript. No AI-generated content was used for data analysis, interpretation, or substantive scientific contributions.
Funding sourceThe authors received a Research Grant for Pilot Projects from Son Llatzer University Hospital in 2019. This grant had no implications in the study design, data collection, analysis or interpretation, in the writing of the report, or in the decision to submit the article for publication.
The authors certify that there is no conflict of interest with any financial organization regarding the material discussed in the manuscript. The authors declare that RECK-Technik provided an unconditional loan of the cycle ergometer. The company had no role in the design, conduct, analysis, or reporting of the study.
The authors sincerely thank Carmen Buen and Petra Vidal, Chief and Lead Physiotherapist of the Department of Physical Medicine and Rehabilitation of the Son Llatzer University Hospital in Palma, respectively, for their valuable contribution to the implementation of the study. Authors gratefully acknowledge RECK-Technik for their unconditional loan of the cycloergometer used in this study.
