Perioperative physiotherapy has regularly been used to prevent or reduce pulmonary morbidity after lung resection1,2 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.3,4 In 2006, Varela and co-workers5 published a cross-sectional study showing that implementing an intensive perioperative respiratory physiotherapy program for patients undergoing lobectomy, in addition to incentive spirometry and standard nursing care, decreased the rate of postoperative atelectasis and produced costs savings associated with a shorter hospital stay. The experimental program started the day before surgery and was based on aerobic exercise up to discharge.

To date, no single-blind randomized clinical trial evaluating the benefits of respiratory physiotherapy or an added specific intervention, focusing only on the immediate perioperative period, has shown significant differences between groups in the frequency of postoperative pulmonary complications (PPCs) or length of stay (LOS).6–10 There is a need to identify exactly which interventions are most beneficial in the immediate perioperative period in terms of reducing PPCs and in which patients, particularly now that guidelines are recommending surgery in increasingly higher risk patients.3 This is of clinical and economic importance as PPCs are directly associated with higher mortality and costs.11

Hence, the aim of this study was to evaluate the effects of a postoperative physiotherapy program based on respiratory exercises in reducing PPCs and LOS in patients undergoing lobectomy.

Our results indicate that the implementation of a postoperative intensive physiotherapy program based on respiratory exercises is effective in reducing the risk in patients undergoing lobectomy of PPCs (P=.017; 95% CI: 0.081–0.780) and LOS (P=.017; median [IQR] in control group 14 [7] and 12 [6] in experimental group; Huber M-estimator 14.21 in control group and 12.81 in experimental group). To the best of our knowledge, this is the first study showing the benefits derived from a physiotherapy program added to standard nursing care and incentive spirometry that it is not based on cardiopulmonary exercise.

The importance of controlling factors that could affect the internal validity of a study, particularly incision type, is well established. In this regard, we emphasize that case selection criteria remained substantially unchanged during the study period; both series were operated on by the same team, following similar operative and resectability selection criteria and anesthetic management. The approach for all patients was open thoracotomy, some with posterior muscle-sparing or axillary thoracotomy, at the surgeon's choice, but we did not collect this information because we considered, as others have done before,5,14–16 that both surgical approaches are homogeneous. Furthermore, prior evidence is available showing that these different types of incision have similar postoperative pulmonary function outcomes, complications, and resulting pain.17–19 Given all this, we believe that all relevant clinical strategies have been included. Moreover, we are aware that surgical practices in recent years have changed and that video-assisted thoracoscopic surgery (VATS) lung resections are increasingly common, whereas all patients in the present study were operated on using open thoracotomy. Nevertheless, data published by the European Society of Thoracic Surgeons20 report open lobectomy as still the most widely used surgical procedure in lung resection, so the results of this study remain useful for current clinical practice and meaningful for a wide population.

We observed a significantly higher number of epidural catheters for postoperative analgesia in the experimental group (78.3%) compared with the control group (60.8%) (P=.006). Postoperative analgesia by epidural catheter consisted of bupivacaine, fentanyl and adrenalin, while the intravenous route consisted of morphine, metamizol and droperidol; both were administered using patient-controlled systems. Epidural analgesia has a recognized role in the control of postoperative pain,13 so it may be that patients experience less pain, thus influencing outcomes. In this study, however, analgesia by PCEA provided no significant benefit in the bivariate (P=.818) or multivariate analysis (P=.153, 95% CI: 0.706–9.257) of PPCs. These results are consistent with the latest systematic reviews conducted by Romero et al.21 and Joshi et al.22 in concluding that although epidural analgesia is more effective than intravenous in reducing post-thoracotomy pain, it has no proven efficacy in reducing the incidence of PPCs.

Moreover, if we compare our results with recent studies of physiotherapy interventions focusing only on the immediate perioperative period, our PPC frequency (6.6% in experimental group) was similar to that reported by Novoa et al.14 (4.7%) and Reeve et al.9 (4.8%) but considerably lower than that of Agostini et al.6 (12%), Arbane et al.23 (16%), and Ludwig et al.8 (19%). Variability in the reported PPC incidences between studies has usually been explained in part by the differences in the definitions included under the term “PCC”. In this study, PPCs included were limited to pulmonary complications more amenable to physiotherapy intervention,24–26 to wit, atelectasis, pneumonia and respiratory failure. Moreover, any medical report of atelectasis and pneumonia was always confirmed by radiological images and respiratory failure was confirmed by arterial blood gases.

Our lung function results indicate that preoperative %FEV1 played a role in the development of PPCs in the study population (P=.042; 95% CI: 0.941–0.999). These results are consistent with the recommendations published in the clinical guidelines of the European Respiratory Society and the European Society of Thoracic Surgeons,3 which suggest that the threshold for increased risk of postoperative mortality and morbidity in patients undergoing lung resection is a ppoFEV1 of 30%. The preoperative %FEV1 in our study was 86.9% (SD 19.2) in the control group and 85.4% (SD 20.7) in the experimental group, explaining how this variable was close to statistical significance in the bivariate analysis of PPCs (P=.078). Thus, the influence of %FEV1 on the risk of developing PPCs in our study should be interpreted with some caution.

We found that LOS (median [IQR] 14 [7] days; Huber M-estimator 14.21 in the control group and 12 [6] days; Huber M-estimator 12.81 in the experimental group) was significantly reduced by the implementation of the physiotherapy program (P=.017). Even though a decrease in LOS is demonstrated, hospitalization in our series was longer than that currently reported by other studies: Reeve et al.9 reported a median of 6 days, Agostini et al.6 5 days, and Arbane et al.23 7.7 days. The difference in LOS may be explained by diversity in medical care and discharge criteria. The criteria followed in our study required a drain of less than 200ml/day for chest tube removal, and no patient was discharged with chest tubes. It would be interesting to know how chest tube removal was managed in the above-mentioned studies, and if this was associated with LOS. While Agostini et al.6 and Reeve et al.9 included different operative procedures in their studies (lobectomies, pneumonectomies, wedge resections, sleeve resections, etc.) and the approach in all patients was open thoracotomy, the series of Arbane et al.23 included both lobectomies and pneumonectomies, approached by open thoracotomy or VATS. This heterogeneity in the study population could influence LOS, as minor lung resection procedures and VATS have better perioperative outcomes than other procedures or open thoracotomy, including duration of hospitalization.27 It would also be interesting to study the influence of LOS on post-hospitalization outcomes, such as emergency visits and/or hospital readmission frequency for any surgery-related cause. This topic was studied in 2004 by Varela et al., 28 who reported a readmission rate of 5% during the first 30 days after discharge in patients operated on by lobectomy; the authors found no influence of LOS on post-hospitalization outcomes (median LOS 7 days). We subsequently re-reviewed the medical records of our patients and found a decrease in readmission incidence for any surgery-related cause from 5.9% (6 patients) in the control group to 2.8% (3 patients) in the experimental group, though the reduction was not statistically significant (P=.268). In our opinion, cost-effectiveness studies that also take into account post-hospitalization outcomes would be necessary to assess the benefits of physiotherapy interventions from a broader perspective; this has important clinical bearing and is essential for taking evidence-based decisions.

We considered introducing propensity-matched methodology,29 but we finally decided not to, due to the similarity between the control and experimental groups in baseline and perioperative characteristics, and since this procedure would reduce the final sample size.

The limitations of our study are derived mainly from its retrospective design, which made it difficult to recover data. The fact that all perioperative practices remained unchanged through the study period could minimize design deficiencies. No diffusing capacity of the lung for carbon monoxide (DLCO) or exercise capacity variables were recorded, as these parameters were not determined at the time of data collection.

Finally, the significant clinical and economic impact of PPCs in patients undergoing lung resection (increased mortality and costs)11 generates a need to determine the role of postoperative physiotherapy in reducing such events. Future research should explore the optimum interventions for different sub-groups of patients, not only in the immediate perioperative period but also before and after thoracic surgery. This knowledge would help in the development of guidelines for clinicians and policymakers, and would also increase health-related quality of life in this patient population.

In conclusion, a postoperative intensive physiotherapy program focused on respiratory exercises has demonstrated efficacy in reducing the risk of PPCs and LOS in patients undergoing lobectomy.

The authors declare no conflict of interest.

This study was approved by the local institutional review board.