Effectiveness of Different Physiotherapy Protocols in Children in the Intensive Care Unit: A Randomized Clinical Trial : Pediatric Physical Therapy

Secondary Logo

Journal Logo


Effectiveness of Different Physiotherapy Protocols in Children in the Intensive Care Unit: A Randomized Clinical Trial

Souza, Gabrielle Sousa Barros PT; Novais, Mariana Furtado Marques PT; Lemes, Guilherme Euzébio PT; de Mello, Mary Lucy Ferraz Maia Fiuza MD; de Sales, Susan Carolina Diniz MD; Cunha, Katiane da Costa PhD; Rocha, Larissa Salgado de Oliveira PT, PhD; Avila, Paulo Eduardo Santos PhD; Rocha, Rodrigo Santiago Barbosa PT, PhD

Author Information
Pediatric Physical Therapy 34(1):p 10-15, January 2022. | DOI: 10.1097/PEP.0000000000000848
  • Free
  • Beyond the Read


One of the main causes of admission of infants and children in intensive care units (ICUs) is acute respiratory failure, characterized by the inability to perform gas exchanges properly.1 When the oxygenation index classifies the failure as severe,2 it is indicated that the management of this condition is done through invasive mechanical ventilation (IMV), which is accompanied by a series of complications.3 These complications include tracheal lesion or edema caused by the endotracheal tube or tracheostomy, pneumothorax, or prolonged ventilatory dependence caused by the positive pressure offered by the ventilator, and the production of oxygen free radicals caused by the use of oxygen.4 Immobilization associated with these complications is evident since the first week of admission to the ICU.5

There is evidence about early rehabilitation in critically ill adults, demonstrating safety, efficacy, and cost-benefit ratio. However, there are only a few studies that use this type of practice in children, mainly referring to mobilization and exercises at the right time, according to the characteristics of the child and the approaches to mobilize children in the hospital environment.6

Physical therapy based on physical exercise is a strategy used to prevent complications from immobilization, which includes acute muscle weakness, polyneuropathy, and myopathy of critical illness, all attributable to a prolonged stay in bed and affecting between 50% and 80% of children in the ICU.7,8 Physical therapy protocols are safe and viable, and tend to provide improvement in functional capacity, and reduce mortality associated to acute respiratory failure, time on IMV, and hospital stay,9,10 in addition to improving cardiac autonomic modulation.11

Heart rate variability (HRV) consists of fluctuation in the intervals between one heartbeat and another, and reflects the regulation of autonomic balance and functions such as maintaining blood pressure and vascular tone. Pathological conditions can produce autonomic cardiac dysregulation, while ideal HRV levels are linked to adaptability.12 HRV analysis can be performed noninvasively and consists of a mechanism used to assess the regulation of the autonomic nervous system in various clinical situations.13 In this context, the assessment of HRV shows promise in predicting important clinical outcomes in the ICU14 since continuous HRV screening can serve as a recovery marker in critically ill children.15 Thus, the present study investigated the effectiveness of different physical therapy protocols in the autonomic modulation of heart rate (HR), time on IMV, and length of stay in children in the ICU.


This was a randomized controlled prospective clinical trial. It was conducted in a pediatric ICU from July 2018 to September 2019 and included 20 children of both sexes on IMV. The study was approved by the Research and Ethics Committee by Santa Casa de Misericórdia do Pará (2.695.300) and the clinical trial was registered at http://clinicaltrials.gov (NCT03343717). All participants involved in the study were authorized to participate in the study by a legal guardian and informed consent was obtained.


Children aged 2 to 11 years, within the first 48 hours on IMV, participated. Children with decompensated heart failure, severe respiratory failure (Pao2/Fio2 <100), active bleeding, hemodynamic instability, acute brain disorder (intracranial pressure [ICP] >10 mm Hg in babies and ICP >15 mmHg in others), orthopedic changes (fractures, dislocations, and subluxations), neurological impairment with minimal function, and neuromuscular disease were excluded, because these criteria were contraindications for the physical exercise proposed in the protocol of this study.

The sample calculation was performed using the GraphPad StatMate application (version 1.01), using a significance of 5% and a test power of 95% for the determination of the sample. This determined that a total of 20 participants were required.


To ensure safety during data collection, vital signs such as HR, blood pressure, and peripheral oxygen saturation of the study participants were continuously assessed using the multiparametric monitor OMNI 600 (OMNIMED LTDA, Belo Horizonte, Brazil). We implemented the algorithm proposed by Mendez-Tellez et al,16 tolerating HR variation during application of the protocol in up to 20% of the resting HR, pulse saturation variation in up to 5%, and respiratory rate with maximum variation in up to 10 breaths. The time of IMV and the length of hospital stay were also collected, through the medical records available in the ICU.

The HRV data collection was performed once in the preprotocol period, on the fourth day of application and 1 week after the completion of the protocol by the same researcher. This was performed using a POLAR brand HR monitor, model RS800CX (Polar Electro TM, Kempele, Finland), in which the HR signal was captured by a strap with the signal receiver, which was placed on the chest of the participants at the location of the xiphoid process of the sternum. A frequency meter was transferred to a masked researcher who received a table with only the times and days of data collection, the data captured by the frequency meter were transferred to the Polar ProTrainer Software (Polar Electro OY, Kempele, Finland), and were stored and subsequently exported to the .txt format so that they could later be analyzed by a mathematical routine in the Kubios HRV 2.2 program (MATLAB, Kuopio, Finland) according to the guidelines of the Task Force of the European Society of Cardiology and by the North American Society of Electrophysiology.17

HRV measurement data were distributed to 2 masked evaluators. In case of disagreement, a third evaluator was consulted. All evaluators had extensive experience in physical therapy for children.

The analysis of HRV by the non-linear method, used in the study, was performed through a graph that shows the relationship of the R-R intervals with quantitative and qualitative patterns in the shape of an ellipse. Standard deviations SD1 and SD2 represent frequent high and rapid changes and long-term changes, respectively. The variables SD1, SD2, and the SD1/SD2 ratio were used. The 5-minute stretch with the greatest signal stability was selected, discarding the initial 30 seconds and the final 30 seconds of the collection.


The researchers who applied the protocols had more than 5 years of experience in the treatment of children admitted to ICUs. The protocols were applied by 2 masked researchers, one of them applied the protocol to the control group (CG) and the other to the experimental group (EG). The protocols applied to the CG and the EG were applied twice a day, for a duration of 30 minutes for each of the protocols, for 5 consecutive days, totaling 10 sessions, controlled through a daily record made by the professionals.

The participants were randomized by lot into 2 groups by a masked researcher, using the site http://randomizer.org. Participants in the CG received the hospital's rehabilitation protocol, which consisted of breathing exercises and techniques for bronchial hygiene maneuvers, including postural drainage, acceleration of expiratory flow, manual hyperinflation with Ambu and tracheal aspiration. The purpose of this protocol is to unblock the airways, pulmonary expansion and reverse atelectasis, and to reduce deformities and preserve joint mobility through passive mobilization techniques at the shoulder, elbow, wrist, hip, knee, and ankle joint.

Participants in the EG received the hospital protocol, in addition to the protocol proposed by the authors. This protocol aimed to perform physical exercises based on the degree of muscle strength. If children presented with muscle strength greater than 3/5 (assuming manual muscle testing), then they performed active exercise and children with strength grade of 3/5 or less performed active assisted exercise. For children who presented with muscle strength less than 3/5, they performed stretching and joint mobilization, in addition to sitting at the bedside. Participants with muscle strength greater than 3/5 performed active flexion and extension exercises in 10 repetitions for the hip, knee, ankle, shoulder, and elbow, in addition to sitting at the bedside and standing for 10 minutes.

Data Analysis

Data analysis was performed using the IBM SPSS 25 program. Normality of data was analyzed using the Shapiro-Wilk test. The analysis of variance of the data was performed using the multivariate analysis of variance test with Tukey HSD (honestly significant difference) post hoc. Demographic data were analyzed with the χ2 test. A significance level of P ≤ .05 was used and data were expressed as mean, median, and 95% confidence interval.

The possible influence of treatments was tested using an effect size to compare the CG with the EG. For this, the Cohen's d pooled method was used. This analysis was performed by the “Effect Size Generator” application, version 2.3 (Swinburne University of Technology, Center for Neuropsychology, Melbourne, Australia). The results were interpreted according to the method proposed by Cohen.18 A value below 0.49 is considered to have a small effect, a value between 0.5 and 0.79 would have a medium effect, and a value of greater than 0.8 is considered to have a large effect. Values below 0.19 are considered insignificant.


Fifty-seven children were evaluated for participation in the study, but after applying inclusion and exclusion criteria, 27 children were excluded. The 30 remaining participants were randomized into 2 groups (CG, n = 15; EG, n = 15) (Figure). Only 10 participants completed the study as 2 had complications related to the underlying disease and death, 2 had seizures, 2 had sepsis, and 1 had arrhythmia, all unrelated to the application of the protocols (Table 1).

Flowchart of volunteer participation in the study.
TABLE 1 - Clinical and Demographic Characteristics of Patients Involved in the Study
Characteristics CG (n = 10) EG (n = 10) P
Age, y 5.7 ± 3.2 7.8 ± 3.6 .4
Female, n 7 7 1.0
Male, n 3 3 1.0
Baseline disease
Renal failure, n 3 5 .8
Subdural hematoma drainage, n 1 1 1.0
Pneumonia, n 6 4 .6
Dobutamine, n 12 13 .9
Dosage, μg/kg/min 10.4 ± 5.1 8.3 ± 6.4 .7
Time of administration, d 1 ± 2 1 ± 2 1.0
Adrenaline, n 10 10 1.0
Dosage, μg/kg/min 0.1 ± 0.05 0.1 ± 0.05 1.0
Time of administration, d 1 ± 2 1 ± 2 1.0
Rocuronium bromide, n 6 6 1.0
Dosage, mg/kg 0.8 ± 0.3 0.9 ± 0.4 .8
Time of administration, d 1 ± 2 1 ± 2 1.0
Dormonid, n 9 10 .9
Dosage, mg/kg 0.1 ± 0.05 0.1 ± 0.05 1.0
Time of administration, d 1 ± 2 1 ± 2 1.0
Midazolam, n 6 8 .7
Dosage, mg/kg 5.8 ± 4.9 6.8 ± 7.1 .7
Time of administration, d 1 ± 2 1 ± 2 1.0
Ventilation time until beginning of the protocol, h 28 ± 15.7 32 ± 20.2 .2
Time until extubation, h 53 ± 25.2 29 ± 27.3 .03
Length of ICU stay, d 17 ± 15.5 10 ± 10.2 .04
Hospital length of stay, d 60 ± 30.2 55 ± 31.2 .4
Abbreviations: CG, control group; EG, experimental group; ICU, intensive care unit.

Table 1 shows the time in IMV in hours, with the extubation time of the CG being longer than the EG. In addition, the ICU stay in the EG was significantly shorter than the CG and the length of hospital stay did not differ between groups.

HRV in the CG showed no significant difference between preintervention, postintervention, and follow-up for the variables SD1 (P = .15), SD2 (P = .4), and SD1/SD2 (P = .3) (Table 2). In the EG, HRV had a significant increase both in SD1 (P = .001) and SD2 (P = .001), and in the SD1/SD2 ratio before and after intervention, and this was even greater at the follow-up, 1 week after the end of the protocol (P = .03) (Table 3).

TABLE 2 - Analysis of the Heart Rate Variability Indexes for the Control Groupa
SD1, ms 15.23 ± 9.80 (14.05)
8.25 ± 8.10 (8.45)
12.30 ± 8.21 (12.40)
SD2, ms 20.2 ± 9.1 (20.5)
15.34 ± 12.05 (14.1)
17.23 ± 15.21 (16.1)
SD1/SD2 0.75 ± 1.07 (0.6)
0.53 ± 0.67 (0.5)
0.71 ± 0.53 (0.65)
Abbreviations: F-UP, follow-up, 7 days after the end of the protocol; POST, postintervention period; PRE, preintervention period; SD1, dispersion of perpendicular points to the identity line; SD2, measure of the standard deviations of the dispersion of the Poincaré plot points along the RR intervals.
aMean ± SD (median) [95% CI]. P < .05.

TABLE 3 - Analysis of the Heart Rate Variability Indexes for the Experimental Groupa
SD1, ms 6.90 ± 4.32 (6.8)
15.43 ± 10.12 (15.2)
53.75 ± 17.15 (52.5)
SD2, ms 16.8 ± 8.0 (16.4)
28.23 ± 22.45 (27.8)
61.3 ± 19.09 (61.1)
SD1/SD2 0.41 ± 0.54 (0.43)
0.54 ± 0.45 (0.52)
0.87 ± 0.89 (0.88)
Abbreviations: F-UP, follow-up, 7 days after the end of the protocol; POST, postintervention period; PRE, preintervention period; SD1, dispersion of perpendicular points to the identity line; SD2, measurement of the standard deviations of the dispersion of the points of the Poincaré graph along the RR intervals.
aMean ± SD (median) [95% CI]. P < .05.

The comparison of variables before the beginning of the protocols was not statistically significant for the variables SD1 (P = .3), SD2 (P =. 15) and SD1 SD2 (P = .08). The values after application of the protocol were higher in the EG in relation to the CG for the variables SD1 (P = .05) and SD2 (P = .05); however, they did not differ for the SD1/SD2 ratio (P = .1). The analysis of HRV in the follow-up period had higher values for SD1 (P = .001) and SD2 (P = .001) in the EG in relation to the CG; however, for the SD1/SD2 ratio, no statistical differences were found (P = .4). Furthermore, the effect size (ES) values of the sample show large effect to SD1 (ES = 1.1) and SD2 (ES = 1.2), except in the SD1/SD2 ratio with small effect (ES = 0.4) (Table 4).

TABLE 4 - Analysis of Intergroup Heart Rate Variability Indexesa
Variables CG Preintervention EG Preintervention P1 CG Postintervention EG Postintervention P2 CG Follow-up EG Follow-up P3 ES
SD1, ms 15.23 ± 9.80 6.90 ± 4.32 .05 8.25 ± 8.10 15.43 ± 10.12 .05 12.30 ± 8.21 53.75 ± 17.15 .001 1.1
SD2, ms 20.2 ± 9.1 16.8 ± 8 .15 15.34 ± 12.05 28.23 ± 22.45 .05 17.23 ± 15.21 61.3 ± 19.09 .001 1.2
SD1/SD2 0.75 ± 1.07 0.41 ± 0.54 .08 0.53 ± 0.67 0.54 ± 0.45 .1 0.71 ± 0.53 0.87 ± 0.89 .08 0.4
Abbreviations: CG, control group; EG, experimental group; ES, effect size values; P1, P value between CG and EG preprotocol; P2, P value between CG and EG after the end of the protocol; P3, P value between CG and EG 7 days after the protocol; SD1, dispersion of perpendicular points to the identity line; SD2, measure of the standard deviations of the dispersion of the Poincaré plot points along the RR intervals.
aResults expressed as mean ± standard deviation. P < .05.


The present study sought to investigate the effectiveness of different physical therapy protocols in children admitted to an ICU and submitted to IMV. Although the study sample is small, the results are promising, indicating that the proposed protocol positively impacted those in the EG, by significantly increasing HRV, resulting in less time spent on IMV and in the ICU setting.

In the preprotocol period, it was identified that the SD1 variable had significant differences even before the application of the protocol. This fact is related to the 2 children who had values above the mean of the CG, even after performing randomization and having been included in the same group. The event did not affect the homogeneity of the sample, and the values that were lower in the EG were shown to be higher after the application of the protocol compared with the CG.

HRV has been studied and its decrease represents reduced adaptability, and this finding is a predictor of mortality in sepsis,19,20 cardiovascular problems,21 risk of developing stroke,22 and possibly related to the development of delirium between23 relationships with other systems. The data obtained appear as indicators of vagal activity and cardiac vagal control, associated with both physiology and psychological stress.24 The ratio between these 2 variables measures the unpredictability of the time series R-R, measuring the autonomic balance of the individual who tends to deteriorate with the worsening of health.12

Regarding the relationship between HRV and mechanical ventilation, a multicenter study showed that, in adults, lower HRV during the spontaneous breathing test that precedes extubation implies failure in extubation and consequent prolongation in IMV.25 Low HRV values are associated with longer IMV time in children, and higher values may be able to predict success in removing IMV.26 In the present study, the lowest HRV values belong to children who spent more time undergoing IMV.

Physical exercise has been shown to be increasingly important in improving autonomic function. Physical exercise therapy in children with heart failure improved HRV in the short term, reflecting on parasympathetic activity. Short-term combined exercise training can reverse impairments in cardiorespiratory fitness.27 Although the participants in that study were adults, physical exercise in people with reduced HRV improves autonomic modulation. The findings of the present study corroborate these results, since the improvements in HRV were greater in the EG, with the participants who exercised more. Exercise is a promising tool in improving HRV, but no studies were found during the literature search that indicated such a benefit in the children hospitalized in the ICU. This study appears to be a pioneer in this regard.

Thus, it is possible to infer that children submitted to a physical therapy protocol based on physical exercise, which has one of the benefits of increased HRV, both in the post-protocol and in the follow-up, have less susceptibility to events resulting from poor adaptability, in addition to having the benefit of less time in IMV and less time in the ICU.28 The same results of improvement in HRV were found in a clinical trial in children with pneumonia associated with mechanical ventilation.11

Other authors have also shown improvements in time on IMV and hospital stay, in addition to functional capacity, increased muscle strength, and decreased late complications29 in children undergoing early physical exercise. In children, this therapeutic practice is viable and safe,7 information confirmed by this study since there were no complications such as drops in oxygen saturation or variations in blood pressure that would cause instability to the child.


The limitations presented in the study were the nonstandardization of drugs used by children and the scarcity of publications on physical therapy based on physical exercise with children in the ICU, especially when related to HRV. In view of this, the authors recommend that it be continuously studied, especially in children, in which the relationship between early physical exercise and HRV is still poorly understood.

In contrast to the limitations, the benefit of the protocol allows cost reduction in the ICU, at least for length of stay, reducing the financial effect. In addition, the protocol applied in this study was able to reduce the time of IMV.


The effects that a physical therapy protocol based on physical exercise has on the autonomic modulation of HR, time on IMV, and length of stay in the ICU in children undergoing mechanical ventilation were positive in the present study. It means to say that, when comparing the 2 groups of children, there was an improvement in HRV and a decrease in time on IMV, as well as the length of ICU stay in children submitted to the protocol.


1. Friedman M, Nitu M. Acute respiratory failure in children. Pediatr Ann. 2018;47(1):268–273.
2. DesPrez K, McNeil J, Wang C, Bastarache J, Shaver C, Ware L. Oxygenation saturation index predicts clinical outcomes in ARDS. Chest. 2017;152(6):1151–1158.
3. Alibrahim O, Slain K. Noninvasive respiratory support in pediatric acute respiratory distress syndrome. In: Shein S, Rotta A, eds. Pediatric Acute Respiratory Distress Syndrome. Cham, Switzerland: Springer; 2020.
4. Haroon Hamid M. Basics of mechanical ventilation in children—a review. Pak Pediatr J. 2018;42:3–15.
5. Parry S, Puthucheary Z. The impact of extended bed rest on the musculoskeletal system in the critical care environment. Extrem Physiol Med. 2015;4:16.
6. Cuello-Garcia C, Mai S, Simpson R, Al-Harbi S, Choong K. Early mobilization in critically ill children: a systematic review. J Pediatr. 2018;203:25–33.e6.
7. Taylor C. Intensive care unit acquired weakness. Anaesth Intensive Care Med. 2018;19:79–82.
8. Senger D, Erbguth F. Critical-illness-myopathie und-polyneuropathie. Med Klin Intensivmed Notfmed, 2017;112:589–596.
9. Green M, Marzano V, Leditschke I, Mitchell I, Bissett B. Mobilization of intensive care children: a multidisciplinary practical guide for clinicians. J Multidiscip Health. 2016;25:247–256.
10. Dirkes S, Kozlowski C. Early mobility in the intensive care unit: evidence, barriers, and future directions. Crit Care Nurs. 2019;39(3):33–42.
11. Rocha G, Santos J, Oliveira M, Avila P, Rocha R. Efeitos da Mobilização Precoce Em Crianças com Pneumonia Associada a Ventilação Mecânica: Efeitos sobre variáveis não Lineares na Variabilidade Da Frequência Cardíaca. Rev Bra Ci Mov. 2019;27:93–98.
12. Shaffer F, Ginsberg J. An overview of heart rate variability metrics and norms. Front Public Health. 2017;5:258.
13. Charlier P, Cabon M, Herman C, et al. Comparison of multiple cardiac signal acquisition technologies for heart rate variability analysis. J Clin Monit Comput. 2020;34(4):743–752.
14. Bishop D, Wise R, Lee C, von Rahden R, Rodseth R. Heart rate variability predicts 30-day all-cause mortality in intensive care units. South Afr J Anaesth Analg. 2016;22:125–128.
15. Marsillio L, Manghi T, Carroll M, Balmert L, Wainwright M. Heart rate variability as a marker of recovery from critical illness in children. PLoS One. 2019;14(5):e0215930.
16. Mendez-Tellez PA, Nusr R, Feldman D, et al. Early Physical rehabilitation in the ICU: a review for the neurohospitalist. Neurohospitalist. 2012;2(3):96–105.
17. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Heart rate variability. Standards of measurement, physiological interpretation, and clinical use. Eur Heart J. 1996;17(3):354–381.
18. Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. New Jersey, NY: Lawrence Earlbaum Associates; 1988.
19. de Castilho FM, Ribeiro ALP, da Silva JLP, Nobre V, de Sousa MR. Heart rate variability as predictor of mortality in sepsis: a prospective cohort study. PLoS One. 2017;12(6):e0180060.
20. Bonjorno JC Junior, Caruso FR, Mendes RG, et al. Noninvasive measurements of hemodynamic, autonomic and endothelial function as predictors of mortality in sepsis: a prospective cohort study. PLoS One. 2019;14(3):e0213239.
21. Karimi Moridani M, Setarehdan S, Motie Nasrabadi A, Hajinasrollah E. Non-linear feature extraction from HRV signal for mortality prediction of ICU cardiovascular patient. J Med Eng Technol. 2016;40(3):87–98.
22. Chen C, Huang P, Tang S, et al. Complexity of heart rate variability can predict stroke-in-evolution in acute ischemic stroke patients. Sci Rep. 2015;5:17552.
23. Neerland B, Wyller T, Wyller V. Autonomic cardiovascular control in older children with acute infection and delirium: a pilot study of orthostatic stress responses. BMC Geriatr. 2019;19(1):23–30.
24. Laborde S, Mosley E, Thayer J. Heart rate variability and cardiac vagal tone in psychophysiological research – recommendations for experiment planning, data analysis, and data reporting. Front Psychol. 2017;8:213.
25. Seely A, Bravi A, Herry C, et al. Do heart and respiratory rate variability improve prediction of extubation outcomes in critically ill children? Crit Care. 2014;18:R65.
26. Johnston C, Dias C, de Cavalho W, Delgado A. Abstract PCCLB-61. Pediatr Crit Care Med. 2018;19:263.
27. Mostarda C, Castro-Filha J, Reis A, et al. Short-term combined exercise training improves cardiorespiratory fitness and autonomic modulation in cancer children receiving adjuvant therapy. J Exer Rehabil. 2017;13(5):599–607.
28. Lai C, Chou W, Chan K, et al. Early Mobilization reduces duration of mechanical ventilation and intensive care unit stay in children with acute respiratory failure. Arch Phys Med Rehabili. 2017;98:931–939.
29. Yayla A, Özer N. Effects of early mobilization protocol performed after cardiac surgery on patient care outcomes. Int J Nurs Practi. 2019;25(6):e12784.

cardiovascular system; early ambulation; intensive care units; pediatric

© 2021 Academy of Pediatric Physical Therapy of the American Physical Therapy Association