Immediate effects of upper limb exercises with and without deep breathing on lung function after cardiac surgery – a randomized crossover trial

Abstract

Background

Open heart surgery, involving median sternotomy, may cause diminished chest wall motion and restrictive pulmonary function in the early postoperative period. Thoracic and upper extremity range of motion (ROM) exercises are often recommended after surgery but have not been evaluated regarding effect on lung volumes and oxygenation. The objective of this study was to evaluate the immediate effect of upper limb elevations, with or without simultaneous deep breathing, on lung function after cardiac surgery.

Methods

In a randomized 2 × 2 crossover trial, 22 adult patients (> 18 years old) were assessed during one of the first days after surgery in the spring of 2022 at Örebro University Hospital, Sweden. Exercises involving five bilateral upper limb elevations, performed either with simultaneous deep breathing (ROM-DB) or without (ROM), while sitting in an upright position at the edge of the bed, were evaluated. Peripheral oxygen saturation (Rad-5v; Masimo, Irvine, USA) was the primary outcome. Tidal volume and respiratory rate were recorded continuously during the exercises (Spiropalm; Cosmed, Rome, Italy). Heart rate, pain, exertion and dyspnoea were evaluated before and after the exercises.

Results

Both ROM-DB and ROM momentarily increased peripheral oxygen saturation (+ 1% ± 1, p = 0.004 and + 1% ± 1, p < 0.001, respectively), with no significant differences between these exercises (p = 0.525). ROM-DB significantly increased the V_T compared with ROM (798 ± 316 vs. 602 mL ± 176, p = 0.004). However, ROM-DB induced more pronounced pain (p = 0.012), exertion (p = 0.035) and dyspnoea (p = 0.013) than ROM.

Conclusions

Upper limb elevations improved oxygenation momentarily, both performed with and without simultaneous deep breathing, with no significant differences between these exercises. The additive deep breathing improved tidal volume compared with upper limb elevations alone, but induced more pain, exertion and dyspnoea during the performance of exercise.

Trial registration

ClinicalTrials.gov (NCT05278819).

Background

Cardiac surgical procedures are performed with the aim of prolonging life and improving quality of life. In the early postoperative period, side effects such as pain [1,2,3], fear of movement [1] and impaired lung function [4,5,6,7] are common. Atelectasis, pulmonary oedema and hypoxemia may arise, and impairment of lung function can persist for a long period postoperatively [8,9,10,11]. Due to the median sternotomy, impaired thoracic mobility may be present and range of motion exercises (ROM) are therefore commonly prescribed after the surgery [11, 12]. Data on the use of upper-body exercises protocols following cardiac surgery are sparse [13].

Postoperative physiotherapy, including early mobilization and breathing exercises, is part of standard care during the first days after cardiac surgery, mainly aimed at preventing postoperative complications, improving lung function and postoperative physical recovery [12, 14,15,16]. Deep breathing exercises are recommended during the first postoperative period to enhance recovery of pulmonary function [17,18,19,20]. Thoracic and upper extremity ROM exercises, usually initiated during the in-hospital stay and proceeding through the outpatient period of rehabilitation, are part of standard treatment as well [12, 14,15,16]. Data are scarce on the effect of upper extremity exercises on lung function after open heart surgery, however one study of a Pilates program with a follow-up on the day of hospital discharge, has shown improved maximal inspiratory pressure but no significant effect on maximal expiratory pressure, vital capacity or peak expiratory flow [21].

No previous study has been found investigating the effects on lung function of combining ROM exercises and deep breathing after surgery. The aim of this study was to evaluate the immediate effect of upper limb elevations, with and without deep breathing, on lung function after cardiac surgery. The hypotheses were that upper limb elevations increase oxygenation and tidal volume (V_T), and that simultaneous deep breathing would further improve lung function.

Methods

Study design and study population

This study was conducted as a single center, randomized study with a 2 × 2 crossover design. Patients ≥ 18 years scheduled for coronary artery bypass grafting (CABG), aortic (AVR) or mitral valve replacement/repair (MVR) or combined CABG and valve surgery, at Örebro University Hospital in Örebro, Sweden, from March to May 2022 were eligible for inclusion. The patients were consecutively included and assessed on the earliest day possible between postoperative day two and four during weekdays when the principal investigator (MR) was on duty. Only patients requiring oxygen supplementation with a fraction of inspired oxygen (FiO₂) less than 40% to maintain peripheral oxygen saturation (SpO₂) at or above 92% were included. Patients who experienced complications necessitating prolonged immobilization or intubation in the intensive care unit were not included in this study. These complications could include severe infections, respiratory failure, or other critical conditions that required extended ICU care, making them ineligible to participate in the research.

Other exclusion criteria were re-do surgeries, cognitive inability to perform exercises and/or measurements (arbitrarily assessed by the physiotherapist at the ward) or inability to communicate in Swedish.

Surgery and postoperative care

All patients received general anesthesia and cardiopulmonary bypass during surgery. Median sternotomy was used as surgical approach. For patients undergoing CABG, saphenous vein grafts were mostly used, and the internal mammary artery was used whenever required. During anesthesia and following surgery, oxygen at 40–80% was provided. Drainage was applied to the pericardium and mediastinum and occasionally one or both pleurae, usually for less than 24 h after surgery. In accordance with standard routines for postoperative pain management at the thoracic surgery ward, patients received paracetamol (1 g four times daily) and oxycodone (5–10 mg twice daily and supplementary if needed).

All study patients received the standard postoperative physical therapy treatment provided in the thoracic surgery ward, which included routine deep breathing exercises and early postoperative mobilization. The routine deep breathing exercises were instructed to be performed every hour postoperatively, with three sets of ten deep breaths, using a positive expiratory pressure (PEP) device, aiming at a pressure of 10–15 cmH₂O [18]. The postoperative mobilization was typically initiated on the day of surgery or the 1st postoperative day. Instructions for daily ROM exercises for the thoracic cage and upper limbs were given on the 1st or 2nd postoperative day.

Interventions

All patients performed both interventions, which consisted of five bilateral, anterior upper limb elevations with (ROM-DB) and without (ROM) simultaneous deep breathing. These exercises were performed in a sitting upright position on the edge of the bed. The patients were instructed to keep their hands together during the elevations and to raise their arms as high as possible. To perform ROM-DB, the patients were instructed to take a slow deep breath during elevation of the arms, and to exhale while lowering the arms, at a pace of their own preference. The immediate effect of ROM-DB was compared with the immediate effect of ROM alone.

Randomization

In a 2 × 2 crossover manner, the patients were randomized either to perform ROM-DB first, followed by ROM, or vice versa. A computerized randomization list was administered by an independent statistician. Sealed, opaque envelopes were prepared by an assistant not otherwise involved in the study. Group assignment was revealed to the assessor (principal investigator, MR) on opening the randomization envelope after baseline measurements.

Outcomes

The primary outcome was SpO₂. Secondary outcomes were VT, respiratory rate, heart rate, pain, exertion, and dyspnea. The primary outcome (SpO₂) and heart rate were measured with a pulse oximeter device (Rad-5v; Masimo, Irvine, USA) on the index finger of the patient’s right hand. Values were recorded immediately before and after the exercises as well as the highest value during four minutes of rest following the exercises. V_T and respiratory rate were recorded continuously throughout the whole study procedure (see section below) using a portable spirometer (Spiropalm 6MWT; Cosmed, Rome, Italy) connected to an oronasal silicone mask (7350 series; Hans Rudolph Inc, Shawnee, KS, USA). For the analysis of V_T and respiratory rate, the mean values of the last minute before and the first minute after the exercises was used. Values during exercise are based on mean values representing the total time frame used for performing the exercise. Endpoint values are based on mean values in the fourth minute after performing the last exercise of the test session.

The patients rated pain, exertion and dyspnea by pointing at a numeric rating scale (NRS) ranging from 0 (none at all) to 10 (worst imaginable). After completion of the endpoint measurement, patients were asked to localize the pain experienced. Baseline characteristics of patients and surgical data were retrieved from medical records.

Study procedure

Baseline measurements of SpO₂, V_T, respiratory rate and heart rate were carried out after resting for at least 10 min lying in bed in a supine position, with approximately 30 degrees head of bed elevation. After the supine rest, the patients were mobilized to a sitting upright position on the edge of the bed and rested for four minutes. Immediately before the first series of elevations, assessments of SpO₂, V_T, respiratory rate, heart rate, pain, exertion and dyspnea were conducted and repeated immediately after the exercises. V_T was measured continuously during the interventions. After a wash-out period of four minutes with the patient remaining in a sitting upright position on the edge of the bed, the measurements were repeated the same way for the second set of elevations. After another four minutes’ wash-out period, an endpoint measurement was performed. During resting periods, the patient was instructed to sit and rest without talking. The study procedure took maximally 30 min to complete.

Statistical analysis

The number of patients in this study was determined based on a study by Fjerbaek et al. [22], which demonstrated that mobilization from supine to sitting led to a significant improvement in SpO2, and a study on the immediate effects of deep breathing exercises on oxygenation after cardiac surge [17]. With regard to the power calculations for crossover designs aiming at 80% power (α = 5%), at least 18 patients were required for this study considering a SD of ± 1.4 and a mean difference of 1% in SpO₂ between interventions. To compensate for potential dropouts, another four patients were included.

The Shapiro-Wilk test was used to test normality distribution. Results are reported in mean values and SD or median and interquartile range [IQR]. Comparisons between ROM-DB and ROM were analyzed with paired Student’s t-test or Wilcoxon signed rank test, depending on levels of measurement and data distribution, and a 95% confidence interval (CI) was used. Carry-over effects were analyzed with Wilcoxon rank sum test for each outcome [23]. All results refer to two-sided tests, and p values < 0.05 were considered significant. Version 27 of the SPSS software package (SPSS Inc) was used for statistical analysis.

Results

A total of 22 patients, age 65 ± 9 years, were included in the final analyses (Fig. 1). Another eight patients declined participation. The included patients underwent CABG (n = 11), CABG with AVR (n = 1), isolated AVR (n = 6) or isolated MVR (n = 4). The study was conducted on postoperative day two (n = 13), day three (n = 5) or day four (n = 4).

Fig. 1

Flowchart for the recruitment

Baseline characteristics for the sample are presented in Table 1. Two patients required supplementary oxygen via nasal cannula (2 and 4 L/min) during the intervention to ensure a SpO₂ level of ≥ 92%. Three patients experienced discomfort with the face mask. Statistical analyses showed no carry-over effect (i.e., a complete wash-out) between ROM-DB and ROM for any of the variables measured. Change of body position from supine to sitting before the intervention started increased mean SpO₂ significantly by + 1% ± 2 (p = 0.010) without any significant changes in mean V_T (-25 mL ± 110, p = 0.297).

Table 1 Patient characteristics

Peripheral oxygen saturation

Both ROM-DB and ROM induced an immediate decrease in SpO₂, as presented in Table 2 (values after vs. before the exercises; -2% ± 2, p < 0.001, and − 1% ± 1, p = 0.032, respectively). The decrease in SpO₂ was more pronounced for ROM-DB than for ROM (-1%: 95% CI, -2% to -1%, p = 0.036). Shortly after this decrease, with a mean time for ROM-DB of 57 ± 53 s and for ROM of 38 ± 34 s, SpO₂ peaked to a level exceeding the values before the elevations, with mean + 1% ± 1, both after ROM-DB (p = 0.001) and after ROM (p = 0.004), and with no significant difference between the exercises (p = 1.000). The SpO₂ returned to baseline values within the four-minute resting period, with no significant difference between ROM-DB and ROM (p = 1.000).

Table 2 Outcome measurements before and after upper limb elevations with and without deep breathing exercises (n = 22)

Tidal volume, respiratory rate and heart rate

Tidal volume was significantly larger during ROM-DB compared with ROM (798 mL ± 316 vs. 602 mL ± 176, p = 0.004) (Fig. 2). The V_T remained significantly increased after the exercises compared with pre-exercise values, but without any significant difference between ROM-DB and ROM (p = 0.491) (Table 2). At the end of the four-minute resting period, V_T had returned to baseline values, with no statistically significant difference between ROM-DB and ROM (p = 0.144). There were no statistically significant differences between the exercises regarding changes in heart rate or respiratory rate (Table 2).

Fig. 2

Tidal volume at different time points in relation to the exercises^{a, b}

^a Box plot with the median value as a horizontal bar and the edges of the box representing the first and third quartiles. The width of the box is the interquartile range (IQR). The whiskers extend to the smallest or largest value within 1.5 times the interquartile quartile range.

^b Values before and after exercise are based on mean values of the last minute before and the 1st minute after the exercises. Values during exercise are based on mean values representing the total time frame used for performing the exercise. Endpoint values are based on mean values in the 4th minute after performing the last exercise of the test session.

Pain, exertion, dyspnea

Before the first upper limb exercises, NRS pain was 1.5 [1.0, 3.0]. The patients located the pain to the sternum (n = 15), right shoulder (n = 2), thoracic spine (n = 1) and right side of the neck due to central venous catheter (n = 1). ROM-DB caused a significant increase in NRS pain from 1.0 [0.3, 3.0] to 3.0 [1.0, 5.0], p = 0.001, whereas ROM did not induce any change in NRS pain (p = 0.109). This difference between ROM-DB and ROM was statistically significant (p = 0.012).

ROM-DB caused a significant increase in exertion from 3.0 [1.0, 4.3] to 4.0 [2.0, 6.0], p < 0.001, while ROM alone did not induce any change in exertion (p = 0.092). This difference between ROM-DB and ROM was statistically significant (p = 0.035).

ROM-DB further caused a significant increase in dyspnea from 3.0 [2.0, 4.0] to 4.0 [3.0, 6.0], p < 0.001, whereas ROM did not cause any change in dyspnea (p = 0.581). All of these increases reverted back to baseline values within the four-minute resting period.

Discussion

During the initial postoperative days following cardiac surgery, engaging in a single session of upper limb exercises led to an initial decrease in SpO₂, followed by an increase that surpassed the pre-exercise values, with no notable distinctions observed between performance with simultaneous deep breathing (ROM-DB) and those without (ROM). Performance of ROM-DB induced more pain, exertion and dyspnea during the performance of exercise than during ROM alone, However, there was a notable increase in V_T during ROM-DB in comparison to ROM. All changes disappeared within four minutes of rest after the exercise session.

The initial decrease and the following increase in SpO₂ can be explained by physiological principles related to energy turnover and oxygen demand during and after muscle exertion [24]. The increase in SpO₂ was minor, but could still be important for patients in the early postoperative period. On the other hand, the importance and suitability of active upper limb elevations should be considered especially for patients with more severe impairments in oxygenation, with respect to the observed initial decrease in SpO₂ in this study and in previous observations of a cardiovascular response of exercise after surgery, causing a larger reduction in mixed venous oxygen saturation [25].

In a study on cardiac surgery patients by Westerdahl et al. [16], deep breathing exercises also led to small changes in oxygenation (PaO₂ 0.2 kPa) while a significant increase of 5% of aerated lung area on computed tomography could be demonstrated. Enlargement of V_T, as observed in this study, should be seen as desirable with the postoperative course of open heart surgery in mind, since that results in greater alveolar ventilation by recruiting more lung tissue and thus counteracts postoperative impairment of lung function. The increased V_T could possibly cover the increased oxygen demand during muscle activity [24]. This is similar to observations made previously by Van den Bosch et al. [26] in patients examined on the 1st day after abdominal surgery, suggesting that adaptations of alveolar ventilation to metabolic needs may be predominantly achieved by variations in V_T. This phenomenon can be explained by the fact that due to the constant dead space; an increased breathing rate does not seem to be as effective as increasing the V_T in order to achieve a larger alveolar minute ventilation [27].

Fjerbaek et al. [22], showed a significant positive effect of mobilization from supine to a sitting position in a chair (1% SpO₂, 100 mL V_T) on both SpO₂ and V_T for cardiac surgery patients on postoperative day two or three. In the present study, change of body position from supine to sitting before the intervention start significantly increased SpO₂ but did not alter V_T. It could be speculated that a greater muscular and metabolic task than a change in body position alone is required to further increase V_T, for example ROM exercises or a standing/ambulating transfer to a chair.

Nevertheless, the short-term effects on SpO₂ and V_T shown may contribute to the treatment as a whole with the aim of restoring the postoperative lung function. The results of the present study suggest that upper limb elevations may be considered as a part of the postoperative treatment in order to improve lung function. The benefit of deep breathing as an additive to upper limb elevations at this early postoperative stage should be investigated more closely, especially considering the increase in pain, dyspnea and exertion level observed in this study.

In rehabilitation after cardiac surgery, there are currently no clear guidelines regarding the optimal duration and performance of range of motion (ROM) exercises. The appropriate length of time for these exercises remains uncertain and warrants further evaluation to determine their efficacy and value in the recovery process.

The interventions in our trial should be studied over a longer period of time, covering the whole period of physical rehabilitation, to be able to evaluate all potential benefits.

Strengths and limitations

Because of the complete wash-out and absence of a carry-over effect, analysis and interpretation of the results could be done regardless of alignment.

The patients were examined on different postoperative days. It may be speculated that differences in outcome could be dependent on whether examinations are conducted on the 2nd, 3rd or 4th day after the operation. While the interventions in this study are standardized, the authors recognize that individual patient responses can vary due to a range of factors including pre-existing health conditions, the type and extent of cardiac surgery, patient age, and overall physical condition, as well as the physiotherapists expertise. The design of this study, with each patient acting as their own control and with examination of both exercises on the same postoperative day, minimizes the potential factors that could influence the efficacy of the intervention.

It cannot be ruled out that ongoing oxygen treatment during the test session (two patients) and discomfort with the face mask (three patients) may have affected the outcome for these individuals. V_T can even be influenced by the measurement itself, according to Gilbert et al. [28], resulting in increased volume. It is possible to speculate whether such an influence increased pre- and post-exercise V_T and by that left less room for further enlargement by the exercises, implicating the possibility that the exercises examined could lead to a larger change in V_T and/or SpO₂ in a clinical setting without ongoing V_T measurement than demonstrated in this study.

Furthermore, it is possible to speculate whether the exercises studied influenced mucus clearance and thereby potentially altered outcomes for V_T and/or respiratory rate. It would have to be assumed that a reduced V_T is compensated by an increased respiratory rate in order to maintain the required minute volume. Since the results didn´t show any major alternations in respiratory rate during and the statistical analysis didn´t show any carry-over effects, this possibility can be ruled out to some extent.

Conclusions

In cardiac surgery patients, upper limb exercises resulted in significant short-term improvements in peripheral oxygenation and V_T, irrespective of whether performed with or without simultaneous deep breathing. While simultaneous deep breathing induced a greater increase in V_T during exercises, it also triggered a transient rise in pain, exertion, and dyspnea compared to exercises performed without deep breathing. Repetitions of upper limb exercises over an extended follow-up period may potentially amplify these outcomes; however, this needs to be further studied.