STATE OF THE ART – THORACIC
Interactive CardioVascular and Thoracic Surgery 19 (2014) 269–281 doi:10.1093/icvts/ivu126 Advance Access publication 12 May 2014
Perioperative physiotherapy in patients undergoing lung cancer resection Ana Rodriguez-Larrada, Ion Lascurain-Aguirrebenaa, Luis Carlos Abecia-Inchaurreguib and Jesús Secoc,d,* b c d
Department of Physiology, University of the Basque Country, Leioa, Spain Department of Preventive Medicine and Public Health, Faculty of Pharmacy, University of the Basque Country, Vitoria-Gasteiz, Spain Institute of Biomedicine (IBIOMED), University of León, León, Spain Present address: Visiting Researcher, University of the Basque Country, Leioa, Spain
* Corresponding author. Institute of Biomedicine (IBIOMED), Campus de Vegazana s/n, C.P. 24071 León, Spain. Tel: +34-987-293127; fax: +34-987-442070; e-mail: [email protected] ( J. Seco). Received 14 January 2014; received in revised form 20 March 2014; accepted 27 March 2014
Abstract Physiotherapy is considered an important component of the perioperative period of lung resection surgery. A systematic review was conducted to assess evidence for the effectiveness of different physiotherapy interventions in patients undergoing lung cancer resection surgery. Online literature databases [Medline, the Cochrane Central Register of Controlled Trials (CENTRAL), EMBASE, SCOPUS, PEDro and CINAHL] were searched up until June 2013. Studies were included if they were randomized controlled trials, compared 2 or more perioperative physiotherapy interventions or compared one intervention with no intervention, included only patients undergoing pulmonary resection for lung cancer and assessed at least 2 or more of the following variables: functional capacity parameters, postoperative pulmonary complications or length of hospital stay. Reviews and meta-analyses were excluded. Eight studies were selected for inclusion in this review. They included a total of 599 patients. Seven of the studies were identified as having a low risk of bias. Two studies assessed preoperative interventions, 4 postoperative interventions and the remaining 2 investigated the efficacy of interventions that were started preoperatively and then continued after surgery. The substantial heterogeneity in the interventions across the studies meant that it was not possible to conduct a meta-analysis. The most important finding of this systematic review is that presurgical interventions based on moderate-intense aerobic exercise in patients undergoing lung resection for lung cancer improve functional capacity and reduce postoperative morbidity, whereas interventions performed only during the postoperative period do not seem to reduce postoperative pulmonary complications or length of hospital stay. Nevertheless, no firm conclusions can be drawn because of the heterogeneity of the studies included. Further research into the efficacy and effectiveness of perioperative respiratory physiotherapy in this patient population is needed. Keywords: Pulmonary rehabilitation • Physiotherapy • Perioperative care • Lung cancer
INTRODUCTION Lung cancer is responsible for the largest proportion of cancerrelated deaths around the world [1]. Complete surgical resection with curative intent is the most effective treatment for localized non-small-cell lung cancers. The fact that the disease has usually spread by the time it is discovered means that it is sometimes appropriate to use radiation therapy and chemotherapy in combination with surgery. Unfortunately, only 15% of lung cancers are diagnosed at a localized stage and, thus, are candidates for lung resection, the 5-year survival rate in these cases being 52% [2]. For patients who are candidates for surgery, there is a high risk of postoperative pulmonary complications (PPCs) [3]. PPCs are common after abdominal, cardiac or thoracic surgery and are associated with high rates of mortality, high hospital costs and prolonged length of hospital stay (LOS) [4–7]. Currently, there is no standardized definition for PPCs and complications included in the different studies vary substantially. This explains why PPC rates in the literature range from 2 to 40% [8–11], problems usually considered PPC
being: pneumonia, atelectasis, acute respiratory failure, need for reintubation, pulmonary oedema, bronchospasm, pneumothorax and prolonged air leaks. Several strategies and interventions have been developed in an attempt to reduce the incidence of PPCs: screening for and modification of risk factors, optimization of preoperative status, patient education, intraoperative management and postoperative pulmonary care [12]. Physiotherapy has been regularly utilized in both pre- and postoperative care with the aim of preventing or reducing complications [13, 14] and has recently been recommended by the European Respiratory Society, the European Society of Thoracic Surgeons and the American College of Chest Physicians for providing functional benefits [15, 16]. Several high-quality studies and systematic reviews have assessed the efficacy of different physiotherapy interventions in cardiac and upper abdominal surgery [17–19]. To date, however, there have been few studies investigating the efficacy of physiotherapy interventions in lung cancer resection procedures and, thus, there is limited evidence on which to base treatment recommendations [20].
© The Author 2014. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
STATE OF THE ART
a
270
A. Rodriguez-Larrad et al. / Interactive CardioVascular and Thoracic Surgery
Therefore, the objective of our study was to review systematically the evidence for the efficacy of perioperative respiratory physiotherapy in patients undergoing pulmonary resection for lung cancer in terms of recovery of pulmonary function, recovery of exercise capacity, incidence of PPCs and LOS.
MATERIALS AND METHODS Search and study selection An electronic literature search was performed in MEDLINE, The Cochrane Central Register of Controlled Trials (CENTRAL), EMBASE, SCOPUS, PEDro and CINAHL to find potentially relevant randomized controlled trials (RCTs) published until 30 June 2013. The search strategy was designed to ensure maximum sensitivity, and no language restrictions were applied. The search strategy is shown in Table 1. Studies were included if they were RCTs, compared 2 or more perioperative physiotherapy interventions or compared 1 intervention with no intervention, included only operable patients undergoing lung cancer resection and assessed at least 2 or more of the following variables: functional capacity parameters, PPCs or LOS. Reviews and meta-analyses were excluded. Thereafter, a backward search was performed, reviewing reference lists of included articles in search of further relevant citations. The titles and abstracts of all references obtained in the search were screened. The full text of those that were eligible was assessed against the inclusion criteria independently by two authors (A.R-L., J.S.). Disagreements were resolved by consensus with a third author (I.L-A.).
Risk of bias assessment and data analysis Following the recommendations of the Cochrane Handbook for Systematic Reviews [21], the risk of bias of the studies included was independently assessed by three reviewers (A.R.-L., J.S. and I.L-A.). Unpublished data were requested when necessary from the authors of the original studies. Disagreements were resolved by consensus. In our analysis, the maximum possible score was 10, given that it was not possible to blind the patient and the therapist administering the intervention and, therefore, results must be considered cautiously. These 2 items were marked as ‘not applicable’. Studies that met ≥5 of the 10 proposed criteria were considered to have a low risk of bias. The variability in the outcome measures and the heterogeneity of the interventions across the studies meant that it was not possible to conduct a meta-analysis. Therefore, a qualitative analysis was carried out, assessing the methodological quality of the included trials and the consistency of their findings.
RESULTS Study selection and characteristics The electronic search yielded 470 references. Studies were classified into 3 different sub-groups according to the stage at which the interventions were carried out: studies in which patients received the intervention preoperatively, those in which the intervention was carried out postoperatively and those that included a
Table 1:
Search strategy used in this review
MEDLINE (Pubmed) #1 ‘Lung neoplasms’ [Mesh] #2 Lung resection surgery #3 ‘Rehabilitation’ [Mesh] #4 ‘Physical therapy modalities’ [Mesh] #5 ‘Exercise’ [Mesh] #6 Exercise training #7 ‘Postoperative complications’ [Mesh] #8 ‘Length of stay’ [Mesh] #9 (#1 OR #2) AND #3 AND #7 #10 (#1 OR #2) AND #3 AND #8 #11 (#1 OR #2) AND #4 AND (#7 OR #8) #12 (#1 OR #2) AND #5 AND (#7 OR #8) #13 (#1 OR #2) AND #6 AND (#7 OR #8) CENTRAL (The Cochrane Library) #1 MeSH descriptor Lung Neoplasms explode all trees #2 Lung Cancer #3 (#1 OR #2) #4 MeSH descriptor Rehabilitation explode all trees #5 Physical therapy #6 MeSH descriptor Exercise explode all trees #7 Exercise training #8 MeSH descriptor Postoperative Complications explode all trees #9 MeSH descriptor Length of Stay explode all trees #10 (#3 AND #4) #11 (#3 AND #5) #12 (#3 AND #6) #13 (#3 AND #7) #12 (#10) AND (#8 OR #9) #13 (#11) AND (#8 OR #9) #14 (#12) AND (#8 OR #9) #15 (#13) AND (#8 OR #9) EMBASE (Ovid) #1 lung cancer #2 lung neoplasms #3 rehabilitation #4 physical therapy #5 exercise training #6 postoperative complications #7 length of stay #8 (#1 OR #2) AND (#3 OR #4 OR #5) #9 (#8) AND (#6 OR #7) PEDro #1 lung cancer #2 lung neoplasms #3 rehabilitation #4 physical therapy #5 exercise training #6 postoperative complications #7 length of stay #8 (#1 OR #2) AND (#3 OR #4 OR #5) #9 (#8) AND (#6 OR #7)
combined intervention starting preoperatively and continuing after surgery. Assessment of the title and abstract resulted in 19 studies initially being selected for their potential relevance and significance for inclusion in this review [22–40]. Five assessed a preoperative intervention [22, 23, 30, 32, 34], 8 a postoperative intervention [24–27, 33, 37, 38, 40] and the remaining 6 included a preoperative intervention that continued after surgery [28, 29, 31, 35, 36, 39]. After their full text had been read, 11 studies were excluded: 7 because they were not randomized studies [30–36] and 4 for not providing data on at least 2 of the outcome measures of interest [37–40 ]. A flow diagram of the selection of studies is presented in Fig. 1.
271
STATE OF THE ART
A. Rodriguez-Larrad et al. / Interactive CardioVascular and Thoracic Surgery
Figure 1: Flow diagram. RTC: randomized clinical trial.
The remaining 8 studies [22–29] were selected for inclusion in this systematic review. They included a total of 599 patients who underwent anatomical resection with curative intent for lung cancer. Details of the studies and of the interventions are summarized in Tables 2 and 3, respectively.
Assessment of risk of bias The assessment of the risk of bias is shown in Table 4. Six articles described the randomization and allocation concealment procedure [22, 24–27, 29], whereas Benzo et al. [23] and Turna and co-workers [28] provided these data on request. Seven studies [22–25, 27–29] were single blinded. It was not possible to establish whether the remaining study was blinded, because it was not stated in the article and we failed to contact the authors [26]. The losses to follow-up reported in the original articles were 12.5% [22], 10.5% [23], 2.17% [24], 3.77% [25], 3.94% [27] and 17.95% [29], and Turna and co-workers [28] provided these data
on request (20%). The dropout rate in the remaining study was not stated in the article and, as noted above, we were unsuccessful in contacting the authors [26]. Seven studies achieved a score of 8, 9 or 10 out of 10 [22–25, 27–29], whereas one study scored 6 [26] and was considered to have several methodological flaws. There was no disagreement between the reviewers regarding the data extraction or assessment of the risk of bias.
Outcome measures The findings of the studies included are summarized in Table 5. All studies measured PPCs and LOS, although the criteria used to measure PPCs varied across them. Other outcome measures used were functional parameters, such as the forced expiratory volume in 1 s (FEV1), percentage of predicted FEV1, forced vital capacity (FVC), percentage of predicted FVC, maximal inspiratory pressure, maximal expiratory pressure, 6-min walk test (6MWT), shuttle
272
Table 2: Characteristics of the studies included Study
Methods
Benzo et al. (USA) [23]
RCT Used central randomization for allocation concealment Single-blinded study (outcome assessor)
Postoperative intervention Agostini et al. RCT (UK) [24] A randomization in blocks of 4 on postoperative day 1 was performed Single-blinded study (outcome assessor)
Arbane et al. (UK) [25]
RCT The randomization was performed using computer-generated tables Single-blinded study (outcome assessor)
Ludwig et al. (Germany) [26]
RCT The randomization was performed preoperatively according to year of birth
Interventions
Outcomes
24 patients with: (i) non-small-cell lung cancer resection by open thoracotomy or by VATS; and (ii) previous pulmonary disease, interstitial lung disease or obstructive airway disease, with impaired respiratory function by spirometry 12 patients received pulmonary rehabilitation and 12 received chest physical therapy Final analysis of the postoperative outcome was based on data from 21 patients: 12 from pulmonary rehabilitation arm, 9 from the chest physical therapy arm (3 patients being excluded because of inoperable cancer) 20 patients underwent resection by open thoracotomy and 1 by VATS No significant baseline differences between groups Randomized patients with lung cancer resection by open thoracotomy or by VATS (at least lobe) and moderate-to-severe COPD 10 patients received preoperative pulmonary rehabilitation using a customized protocol versus 9 patients in the control group No differences in frequency of open thoracotomies, pneumonectomies or VATS between groups No significant baseline differences between groups Final analysis of postoperative outcome was based on data from 17 patients: 9 from the pulmonary rehabilitation group and 8 from the control group (2 patients being excluded because of inoperable cancer)
Treatments administered over a period of 4 weeks, 5 sessions per week Pulmonary rehabilitation group: focused on incremental strength and endurance training through aerobic exercise Chest physical therapy group: included instructions about and practice of techniques for lung expansion
Follow-up: 4 weeks Functional parameters: PPCs LOS
Treatments administered over a period of 1 week, 2 sessions a day Control group received no presurgical treatment Intervention group completed a customized protocol based on self-efficacy-based exercise prescription
Follow-up: 1 week PPCs LOS
184 randomized patients over 18 years old. Patients were excluded if they underwent emergency thoracotomy, procedures involving the mediastinum and chest wall, had planned lung resection via VATS were immobile preoperatively or were unable to perform preoperative spirometry or allocated breathing exercise 88 patients were allocated to the control group and 92 to the intervention group (3 patients being excluded) Significant differences were found in age and American Society of Anaesthesiologists scores between groups (higher in the control group). No differences were observed in other baseline characteristics 53 patients undergoing lung resection for non-small-cell lung carcinoma were randomized. Patients were excluded if there was no lung resection, if they underwent pneumonectomy and if stay in intensive care unit post-surgery was longer than 48 h Final analysis of the postoperative outcome was based on data from 51 patients: 25 from control group and 26 from the intervention group 49 patients underwent resection by open thoracotomy and 2 by VATS No significant baseline differences between groups
Treatments administered from day 1 post-surgery until hospital discharge Control group: thoracic expansion exercises Intervention group: same as controls but with the use of the Coach 2 incentive spirometer
Follow-up: from day 1 post-surgery until hospital discharge Functional parameters: PPCs LOS
Treatments administered from day 1 post-surgery to day 5 post-surgery, and there was a further 12-week programme of home support Control group: standard medical and/or nursing care Intervention group: same as controls plus a programme that incorporated daily strength and mobility training
135 patients: 80 from control group 55 from intervention group The surgical approach was a muscle-sparing thoracotomy in all cases No significant baseline differences between groups
Control group: standard postoperative treatment Intervention: same as controls plus IPPB
Follow-up: from day 1 post-surgery to week 12 after discharge Functional parameters: Quadriceps muscle strength by magnetic stimulation Quality of life PPCs LOS Follow-up: from day 1 post-surgery until hospital discharge Functional parameters: PPCs LOS
A. Rodriguez-Larrad et al. / Interactive CardioVascular and Thoracic Surgery
Preoperative intervention Morano et al. RCT (Brazil) [22] The randomization was done in ‘blocks’ of 4, and individual allocations were placed in sealed envelopes Single-blinded study (outcome assessor)
Participants
Reeve et al. (New Zealand) [27]
RCT A randomization table was generated by a computer Single-blinded study (outcome assessor)
Perrin et al. (France) [29]
RCT The randomization was performed using computer-generated numbers and individual allocations were placed in sealed envelopes Single-blinded study (outcome assessor)
Control group: standard medical and/or nursing Intervention group: same as controls plus daily chest physiotherapy
Follow-up: until discharge PPCs LOS
60 lung cancer operable patients (stage IA to IIIB) without major cardiac morbidity (American Society of Anaesthesiologists II or better) were randomized 30 patients allocated to each group No dropouts reported No significant baseline differences between groups except in mean peak expiratory flow and diffusion lung capacity for carbon monoxide, which were significantly lower in the intervention group 39 patients with a preoperative forced expiratory volume in 1 s of proportion of individuals were obese in the treatment group Yes
No: age and ASA 9. Were the groups similar at score in the baseline regarding the most control group important prognostic indicators? 10. Were co-interventions Yes Yes Yes avoided or comparable? 11. Was the compliance Yes Yes Yes acceptable in all groups? Yes Yes Yes 12. Was the timing of the outcome assessment in all groups similar? Total 10/10 10/10 9/10
ASA: American Society of Anaesthesiologists; PEF: peak expiratory flow; DLCO: diffusion lung capacity for carbon monoxide.
A. Rodriguez-Larrad et al. / Interactive CardioVascular and Thoracic Surgery
A
Preoperative intervention
Table 5: Summary of results Study
Functional parameters
Postoperative intervention Agostini et al. Intervention group vs control group: (UK) [24] (i) FEV1 on postoperative day 4: 40 ± 16 vs 41 ± 14%; P = 0.817 (ii) %ppFEV1 achieved on postoperative day 4: 72 ± 19 vs 71 ± 21%; P = 0.744 Arbane et al. Intervention group vs control group: (UK) [25] (i) 6MWD: both groups experienced significant deterioration 5 days postoperatively compared with preoperatively, performance then returning to preoperative levels by postoperative week 12, with no differences between groups: 131.6 (101.8) m vs 128.0 (90.7) m; P = 0.89) (ii) Quadriceps strength over 5-day in-patient period: 4.0 (21.2) kg vs −8.3 (11.3) kg; P = 0.04 between groups Ludwig et al. Intervention group vs control group: (Germany) [26] (i) Differences in pre- and postoperative lung function test results between groups were not statistically significant Reeve et al. Not assessed (New Zealand) [27] Preoperative intervention that continued postoperatively Pehlivan et al. Pulmonary function test parameters did not differ (Turkey) [28] between two groups after IPT period before operation (i) IPT significantly increased preoperative PaO2 (before IPT: 76.75 ± 9.97 mmHg vs after IPT: 79.01 ± 9.44 mmHg; P 0.05
Intervention group vs control group: (i) PPCs (pneumonia, respiratory complications requiring additional ventilatory support and/or return to high-dependency care) 2 vs 3; P >0.05
Intervention group vs control group: (i) PPCs: 27 vs 19%; no significant differences between the 2 groups
Intervention group vs control group: (i) 11 (6–37) vs 11 (5–41) days; differences not statistically significant
Intervention group vs control group: (i) PPCs (not detailed): 4.8 vs 2.9%; P = 1.00
Intervention group vs control group: (i) 6.0 vs 6.0 days; P = 0.87
IPT group vs control group: (i) PPCs: 3.3 vs 16.7% (P = 0.04)
IPT group vs control group: (i) 5.40 ± 2.67 vs 9.66 ± 3.09 days (P
Interactive CardioVascular and Thoracic Surgery 19 (2014) 269–281 doi:10.1093/icvts/ivu126 Advance Access publication 12 May 2014
Perioperative physiotherapy in patients undergoing lung cancer resection Ana Rodriguez-Larrada, Ion Lascurain-Aguirrebenaa, Luis Carlos Abecia-Inchaurreguib and Jesús Secoc,d,* b c d
Department of Physiology, University of the Basque Country, Leioa, Spain Department of Preventive Medicine and Public Health, Faculty of Pharmacy, University of the Basque Country, Vitoria-Gasteiz, Spain Institute of Biomedicine (IBIOMED), University of León, León, Spain Present address: Visiting Researcher, University of the Basque Country, Leioa, Spain
* Corresponding author. Institute of Biomedicine (IBIOMED), Campus de Vegazana s/n, C.P. 24071 León, Spain. Tel: +34-987-293127; fax: +34-987-442070; e-mail: [email protected] ( J. Seco). Received 14 January 2014; received in revised form 20 March 2014; accepted 27 March 2014
Abstract Physiotherapy is considered an important component of the perioperative period of lung resection surgery. A systematic review was conducted to assess evidence for the effectiveness of different physiotherapy interventions in patients undergoing lung cancer resection surgery. Online literature databases [Medline, the Cochrane Central Register of Controlled Trials (CENTRAL), EMBASE, SCOPUS, PEDro and CINAHL] were searched up until June 2013. Studies were included if they were randomized controlled trials, compared 2 or more perioperative physiotherapy interventions or compared one intervention with no intervention, included only patients undergoing pulmonary resection for lung cancer and assessed at least 2 or more of the following variables: functional capacity parameters, postoperative pulmonary complications or length of hospital stay. Reviews and meta-analyses were excluded. Eight studies were selected for inclusion in this review. They included a total of 599 patients. Seven of the studies were identified as having a low risk of bias. Two studies assessed preoperative interventions, 4 postoperative interventions and the remaining 2 investigated the efficacy of interventions that were started preoperatively and then continued after surgery. The substantial heterogeneity in the interventions across the studies meant that it was not possible to conduct a meta-analysis. The most important finding of this systematic review is that presurgical interventions based on moderate-intense aerobic exercise in patients undergoing lung resection for lung cancer improve functional capacity and reduce postoperative morbidity, whereas interventions performed only during the postoperative period do not seem to reduce postoperative pulmonary complications or length of hospital stay. Nevertheless, no firm conclusions can be drawn because of the heterogeneity of the studies included. Further research into the efficacy and effectiveness of perioperative respiratory physiotherapy in this patient population is needed. Keywords: Pulmonary rehabilitation • Physiotherapy • Perioperative care • Lung cancer
INTRODUCTION Lung cancer is responsible for the largest proportion of cancerrelated deaths around the world [1]. Complete surgical resection with curative intent is the most effective treatment for localized non-small-cell lung cancers. The fact that the disease has usually spread by the time it is discovered means that it is sometimes appropriate to use radiation therapy and chemotherapy in combination with surgery. Unfortunately, only 15% of lung cancers are diagnosed at a localized stage and, thus, are candidates for lung resection, the 5-year survival rate in these cases being 52% [2]. For patients who are candidates for surgery, there is a high risk of postoperative pulmonary complications (PPCs) [3]. PPCs are common after abdominal, cardiac or thoracic surgery and are associated with high rates of mortality, high hospital costs and prolonged length of hospital stay (LOS) [4–7]. Currently, there is no standardized definition for PPCs and complications included in the different studies vary substantially. This explains why PPC rates in the literature range from 2 to 40% [8–11], problems usually considered PPC
being: pneumonia, atelectasis, acute respiratory failure, need for reintubation, pulmonary oedema, bronchospasm, pneumothorax and prolonged air leaks. Several strategies and interventions have been developed in an attempt to reduce the incidence of PPCs: screening for and modification of risk factors, optimization of preoperative status, patient education, intraoperative management and postoperative pulmonary care [12]. Physiotherapy has been regularly utilized in both pre- and postoperative care with the aim of preventing or reducing complications [13, 14] and has recently been recommended by the European Respiratory Society, the European Society of Thoracic Surgeons and the American College of Chest Physicians for providing functional benefits [15, 16]. Several high-quality studies and systematic reviews have assessed the efficacy of different physiotherapy interventions in cardiac and upper abdominal surgery [17–19]. To date, however, there have been few studies investigating the efficacy of physiotherapy interventions in lung cancer resection procedures and, thus, there is limited evidence on which to base treatment recommendations [20].
© The Author 2014. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
STATE OF THE ART
a
270
A. Rodriguez-Larrad et al. / Interactive CardioVascular and Thoracic Surgery
Therefore, the objective of our study was to review systematically the evidence for the efficacy of perioperative respiratory physiotherapy in patients undergoing pulmonary resection for lung cancer in terms of recovery of pulmonary function, recovery of exercise capacity, incidence of PPCs and LOS.
MATERIALS AND METHODS Search and study selection An electronic literature search was performed in MEDLINE, The Cochrane Central Register of Controlled Trials (CENTRAL), EMBASE, SCOPUS, PEDro and CINAHL to find potentially relevant randomized controlled trials (RCTs) published until 30 June 2013. The search strategy was designed to ensure maximum sensitivity, and no language restrictions were applied. The search strategy is shown in Table 1. Studies were included if they were RCTs, compared 2 or more perioperative physiotherapy interventions or compared 1 intervention with no intervention, included only operable patients undergoing lung cancer resection and assessed at least 2 or more of the following variables: functional capacity parameters, PPCs or LOS. Reviews and meta-analyses were excluded. Thereafter, a backward search was performed, reviewing reference lists of included articles in search of further relevant citations. The titles and abstracts of all references obtained in the search were screened. The full text of those that were eligible was assessed against the inclusion criteria independently by two authors (A.R-L., J.S.). Disagreements were resolved by consensus with a third author (I.L-A.).
Risk of bias assessment and data analysis Following the recommendations of the Cochrane Handbook for Systematic Reviews [21], the risk of bias of the studies included was independently assessed by three reviewers (A.R.-L., J.S. and I.L-A.). Unpublished data were requested when necessary from the authors of the original studies. Disagreements were resolved by consensus. In our analysis, the maximum possible score was 10, given that it was not possible to blind the patient and the therapist administering the intervention and, therefore, results must be considered cautiously. These 2 items were marked as ‘not applicable’. Studies that met ≥5 of the 10 proposed criteria were considered to have a low risk of bias. The variability in the outcome measures and the heterogeneity of the interventions across the studies meant that it was not possible to conduct a meta-analysis. Therefore, a qualitative analysis was carried out, assessing the methodological quality of the included trials and the consistency of their findings.
RESULTS Study selection and characteristics The electronic search yielded 470 references. Studies were classified into 3 different sub-groups according to the stage at which the interventions were carried out: studies in which patients received the intervention preoperatively, those in which the intervention was carried out postoperatively and those that included a
Table 1:
Search strategy used in this review
MEDLINE (Pubmed) #1 ‘Lung neoplasms’ [Mesh] #2 Lung resection surgery #3 ‘Rehabilitation’ [Mesh] #4 ‘Physical therapy modalities’ [Mesh] #5 ‘Exercise’ [Mesh] #6 Exercise training #7 ‘Postoperative complications’ [Mesh] #8 ‘Length of stay’ [Mesh] #9 (#1 OR #2) AND #3 AND #7 #10 (#1 OR #2) AND #3 AND #8 #11 (#1 OR #2) AND #4 AND (#7 OR #8) #12 (#1 OR #2) AND #5 AND (#7 OR #8) #13 (#1 OR #2) AND #6 AND (#7 OR #8) CENTRAL (The Cochrane Library) #1 MeSH descriptor Lung Neoplasms explode all trees #2 Lung Cancer #3 (#1 OR #2) #4 MeSH descriptor Rehabilitation explode all trees #5 Physical therapy #6 MeSH descriptor Exercise explode all trees #7 Exercise training #8 MeSH descriptor Postoperative Complications explode all trees #9 MeSH descriptor Length of Stay explode all trees #10 (#3 AND #4) #11 (#3 AND #5) #12 (#3 AND #6) #13 (#3 AND #7) #12 (#10) AND (#8 OR #9) #13 (#11) AND (#8 OR #9) #14 (#12) AND (#8 OR #9) #15 (#13) AND (#8 OR #9) EMBASE (Ovid) #1 lung cancer #2 lung neoplasms #3 rehabilitation #4 physical therapy #5 exercise training #6 postoperative complications #7 length of stay #8 (#1 OR #2) AND (#3 OR #4 OR #5) #9 (#8) AND (#6 OR #7) PEDro #1 lung cancer #2 lung neoplasms #3 rehabilitation #4 physical therapy #5 exercise training #6 postoperative complications #7 length of stay #8 (#1 OR #2) AND (#3 OR #4 OR #5) #9 (#8) AND (#6 OR #7)
combined intervention starting preoperatively and continuing after surgery. Assessment of the title and abstract resulted in 19 studies initially being selected for their potential relevance and significance for inclusion in this review [22–40]. Five assessed a preoperative intervention [22, 23, 30, 32, 34], 8 a postoperative intervention [24–27, 33, 37, 38, 40] and the remaining 6 included a preoperative intervention that continued after surgery [28, 29, 31, 35, 36, 39]. After their full text had been read, 11 studies were excluded: 7 because they were not randomized studies [30–36] and 4 for not providing data on at least 2 of the outcome measures of interest [37–40 ]. A flow diagram of the selection of studies is presented in Fig. 1.
271
STATE OF THE ART
A. Rodriguez-Larrad et al. / Interactive CardioVascular and Thoracic Surgery
Figure 1: Flow diagram. RTC: randomized clinical trial.
The remaining 8 studies [22–29] were selected for inclusion in this systematic review. They included a total of 599 patients who underwent anatomical resection with curative intent for lung cancer. Details of the studies and of the interventions are summarized in Tables 2 and 3, respectively.
Assessment of risk of bias The assessment of the risk of bias is shown in Table 4. Six articles described the randomization and allocation concealment procedure [22, 24–27, 29], whereas Benzo et al. [23] and Turna and co-workers [28] provided these data on request. Seven studies [22–25, 27–29] were single blinded. It was not possible to establish whether the remaining study was blinded, because it was not stated in the article and we failed to contact the authors [26]. The losses to follow-up reported in the original articles were 12.5% [22], 10.5% [23], 2.17% [24], 3.77% [25], 3.94% [27] and 17.95% [29], and Turna and co-workers [28] provided these data
on request (20%). The dropout rate in the remaining study was not stated in the article and, as noted above, we were unsuccessful in contacting the authors [26]. Seven studies achieved a score of 8, 9 or 10 out of 10 [22–25, 27–29], whereas one study scored 6 [26] and was considered to have several methodological flaws. There was no disagreement between the reviewers regarding the data extraction or assessment of the risk of bias.
Outcome measures The findings of the studies included are summarized in Table 5. All studies measured PPCs and LOS, although the criteria used to measure PPCs varied across them. Other outcome measures used were functional parameters, such as the forced expiratory volume in 1 s (FEV1), percentage of predicted FEV1, forced vital capacity (FVC), percentage of predicted FVC, maximal inspiratory pressure, maximal expiratory pressure, 6-min walk test (6MWT), shuttle
272
Table 2: Characteristics of the studies included Study
Methods
Benzo et al. (USA) [23]
RCT Used central randomization for allocation concealment Single-blinded study (outcome assessor)
Postoperative intervention Agostini et al. RCT (UK) [24] A randomization in blocks of 4 on postoperative day 1 was performed Single-blinded study (outcome assessor)
Arbane et al. (UK) [25]
RCT The randomization was performed using computer-generated tables Single-blinded study (outcome assessor)
Ludwig et al. (Germany) [26]
RCT The randomization was performed preoperatively according to year of birth
Interventions
Outcomes
24 patients with: (i) non-small-cell lung cancer resection by open thoracotomy or by VATS; and (ii) previous pulmonary disease, interstitial lung disease or obstructive airway disease, with impaired respiratory function by spirometry 12 patients received pulmonary rehabilitation and 12 received chest physical therapy Final analysis of the postoperative outcome was based on data from 21 patients: 12 from pulmonary rehabilitation arm, 9 from the chest physical therapy arm (3 patients being excluded because of inoperable cancer) 20 patients underwent resection by open thoracotomy and 1 by VATS No significant baseline differences between groups Randomized patients with lung cancer resection by open thoracotomy or by VATS (at least lobe) and moderate-to-severe COPD 10 patients received preoperative pulmonary rehabilitation using a customized protocol versus 9 patients in the control group No differences in frequency of open thoracotomies, pneumonectomies or VATS between groups No significant baseline differences between groups Final analysis of postoperative outcome was based on data from 17 patients: 9 from the pulmonary rehabilitation group and 8 from the control group (2 patients being excluded because of inoperable cancer)
Treatments administered over a period of 4 weeks, 5 sessions per week Pulmonary rehabilitation group: focused on incremental strength and endurance training through aerobic exercise Chest physical therapy group: included instructions about and practice of techniques for lung expansion
Follow-up: 4 weeks Functional parameters: PPCs LOS
Treatments administered over a period of 1 week, 2 sessions a day Control group received no presurgical treatment Intervention group completed a customized protocol based on self-efficacy-based exercise prescription
Follow-up: 1 week PPCs LOS
184 randomized patients over 18 years old. Patients were excluded if they underwent emergency thoracotomy, procedures involving the mediastinum and chest wall, had planned lung resection via VATS were immobile preoperatively or were unable to perform preoperative spirometry or allocated breathing exercise 88 patients were allocated to the control group and 92 to the intervention group (3 patients being excluded) Significant differences were found in age and American Society of Anaesthesiologists scores between groups (higher in the control group). No differences were observed in other baseline characteristics 53 patients undergoing lung resection for non-small-cell lung carcinoma were randomized. Patients were excluded if there was no lung resection, if they underwent pneumonectomy and if stay in intensive care unit post-surgery was longer than 48 h Final analysis of the postoperative outcome was based on data from 51 patients: 25 from control group and 26 from the intervention group 49 patients underwent resection by open thoracotomy and 2 by VATS No significant baseline differences between groups
Treatments administered from day 1 post-surgery until hospital discharge Control group: thoracic expansion exercises Intervention group: same as controls but with the use of the Coach 2 incentive spirometer
Follow-up: from day 1 post-surgery until hospital discharge Functional parameters: PPCs LOS
Treatments administered from day 1 post-surgery to day 5 post-surgery, and there was a further 12-week programme of home support Control group: standard medical and/or nursing care Intervention group: same as controls plus a programme that incorporated daily strength and mobility training
135 patients: 80 from control group 55 from intervention group The surgical approach was a muscle-sparing thoracotomy in all cases No significant baseline differences between groups
Control group: standard postoperative treatment Intervention: same as controls plus IPPB
Follow-up: from day 1 post-surgery to week 12 after discharge Functional parameters: Quadriceps muscle strength by magnetic stimulation Quality of life PPCs LOS Follow-up: from day 1 post-surgery until hospital discharge Functional parameters: PPCs LOS
A. Rodriguez-Larrad et al. / Interactive CardioVascular and Thoracic Surgery
Preoperative intervention Morano et al. RCT (Brazil) [22] The randomization was done in ‘blocks’ of 4, and individual allocations were placed in sealed envelopes Single-blinded study (outcome assessor)
Participants
Reeve et al. (New Zealand) [27]
RCT A randomization table was generated by a computer Single-blinded study (outcome assessor)
Perrin et al. (France) [29]
RCT The randomization was performed using computer-generated numbers and individual allocations were placed in sealed envelopes Single-blinded study (outcome assessor)
Control group: standard medical and/or nursing Intervention group: same as controls plus daily chest physiotherapy
Follow-up: until discharge PPCs LOS
60 lung cancer operable patients (stage IA to IIIB) without major cardiac morbidity (American Society of Anaesthesiologists II or better) were randomized 30 patients allocated to each group No dropouts reported No significant baseline differences between groups except in mean peak expiratory flow and diffusion lung capacity for carbon monoxide, which were significantly lower in the intervention group 39 patients with a preoperative forced expiratory volume in 1 s of proportion of individuals were obese in the treatment group Yes
No: age and ASA 9. Were the groups similar at score in the baseline regarding the most control group important prognostic indicators? 10. Were co-interventions Yes Yes Yes avoided or comparable? 11. Was the compliance Yes Yes Yes acceptable in all groups? Yes Yes Yes 12. Was the timing of the outcome assessment in all groups similar? Total 10/10 10/10 9/10
ASA: American Society of Anaesthesiologists; PEF: peak expiratory flow; DLCO: diffusion lung capacity for carbon monoxide.
A. Rodriguez-Larrad et al. / Interactive CardioVascular and Thoracic Surgery
A
Preoperative intervention
Table 5: Summary of results Study
Functional parameters
Postoperative intervention Agostini et al. Intervention group vs control group: (UK) [24] (i) FEV1 on postoperative day 4: 40 ± 16 vs 41 ± 14%; P = 0.817 (ii) %ppFEV1 achieved on postoperative day 4: 72 ± 19 vs 71 ± 21%; P = 0.744 Arbane et al. Intervention group vs control group: (UK) [25] (i) 6MWD: both groups experienced significant deterioration 5 days postoperatively compared with preoperatively, performance then returning to preoperative levels by postoperative week 12, with no differences between groups: 131.6 (101.8) m vs 128.0 (90.7) m; P = 0.89) (ii) Quadriceps strength over 5-day in-patient period: 4.0 (21.2) kg vs −8.3 (11.3) kg; P = 0.04 between groups Ludwig et al. Intervention group vs control group: (Germany) [26] (i) Differences in pre- and postoperative lung function test results between groups were not statistically significant Reeve et al. Not assessed (New Zealand) [27] Preoperative intervention that continued postoperatively Pehlivan et al. Pulmonary function test parameters did not differ (Turkey) [28] between two groups after IPT period before operation (i) IPT significantly increased preoperative PaO2 (before IPT: 76.75 ± 9.97 mmHg vs after IPT: 79.01 ± 9.44 mmHg; P 0.05
Intervention group vs control group: (i) PPCs (pneumonia, respiratory complications requiring additional ventilatory support and/or return to high-dependency care) 2 vs 3; P >0.05
Intervention group vs control group: (i) PPCs: 27 vs 19%; no significant differences between the 2 groups
Intervention group vs control group: (i) 11 (6–37) vs 11 (5–41) days; differences not statistically significant
Intervention group vs control group: (i) PPCs (not detailed): 4.8 vs 2.9%; P = 1.00
Intervention group vs control group: (i) 6.0 vs 6.0 days; P = 0.87
IPT group vs control group: (i) PPCs: 3.3 vs 16.7% (P = 0.04)
IPT group vs control group: (i) 5.40 ± 2.67 vs 9.66 ± 3.09 days (P
