Dyspnea Response Following Bilateral Thoracoscopic Staple Lung Volume Reduction Surgery*

Purpose: Lung volume reduction surgery (LVRS) has shown promise for treating patients with severe emphysema in recent clinical trials. However, response following surgery is difficult to assess due to frequent discrepancies between subjective and objective outcomes. We evaluated the relationship between improvement in dyspnea and pulmonary function response in 145 consecutive patients with inhomogeneous emphysema enrolled in a bilateral thoracoscopic lung volume reduction protocol in order to assess predictors of improved dyspnea outcome and correlation of subjective and objective improvement measures. Materials and methods: Baseline complete pulmonary function testing, spirometry, gas exchange, plethysmography, gas dilution lung volumes, along with resting dyspnea index determinations were performed preoperatively, and repeated short term (mean, 33 days; n=129) and long term (>6 months; mean, 276 days; n=84) following surgery. Results: Improvement in FEV 1 percent predicted was significantly associated with improvement in dyspnea scores, though considerable variability exists (r=0.04, p<O.Ol, short term; r=0.4, p=O.l, long term). In this preselected patient group, those with the extreme degrees of hyperinflation may have less improvement in dyspnea following L VRS than those with milder preoperative hyperinflation. Greater improvement in dyspnea short term and long term was seen in patients with lower presenting residual volume/total lung capacity ratios (r=0.4, p=0.02, short term; r=0.4, p<0.05, long term). Conclusions: Bilateral thoracoscopic staple L VRS results in significant objective and subjective improvement in patients with severe emphysema and hyperinflation. There was considerable variability between improvement in dyspnea and improvement in spirometry, and preoperative predictors of response may differ between these outcome variables. Further studies are needed to define the long-term implications of these findings.

s urgical lung volume reduction procedures to improve pulmonary status are being actively investigated at many centers for patients with severe emphysema.l-10 Optimization of operative proce-dures and development of selection criteria for lung volume reduction surgery (LVRS) are evolving based on assessment of postoperative outcomes. Except for a few reports, most initial studies have justifiably focused analysis primarily on objective pulmonary function outcome measures in order to avoid influence of placebo effects that could potentially bias subjective response measures.l· 2 · 4 -7 · Il -19 However, there are well-documented discrepancies between objective measures of pulmonary function and dyspnea response in patients with emphysema.l. 7,13,[19][20][21][22] Since LVRS is currently aimed primarily at providing symptomatic relief to severely dyspneic patients,5· 7 variables associated with subjective dyspnea improvement must be determined.
This study focuses attention on semiquantitative subjective dyspnea response in order to determine factors that predict improved postoperative dyspnea outcome. In this study, we evaluated the relationship between improvement in dyspnea and pulmonary function response in 145 consecutive patients enrolled in a study of bilateral thoracoscopic LVRS to determine which patients had optimal dyspnea outcome, and to uncover any correlations between subjective and objective improvement measures.
While these analyses are limited by the subjective nature and simplicity of the modified dyspnea scale scoring system, they provide potential initial insights into factors associated with beneficial LVRS outcome that may be further investigated with more specifically directed studies.

MATERIALS AND METHODS
One hundred forty-five consecutive patients enrolled in a prospective bilateral thoracoscopic staple lung volume reduction protocol were studied. Selection criteria and operative procedures have been described previously. 5 • 11 · 23 Informed consent was obtained from all patients. Despite maximal medical management, all patients were markedly symptomatic. Chest radiographs showed hyperexpansion of the thorax \vith flattening or inversion of the diaphragm.
Contraindications to surgery included current cigarette smoking, age older than 80 years, severe cardiac disease, history of cancer \vithin the last 5 years, ventilator dependency, or prior thoracic surgery. Relative contraindications included age older than 75 years, severe anxiety or depression, or C0 2 retention (resting PaC0 2 >55 mm Hg). 5 To be accepted for the procedure, the pattern of emphysema on CT scan had to be severe and heterogeneous. Radionuclide lung perfusion scans were also used to confirm the het erogeneous pattern of emphysema.s.u. 23 All patients underwent complete baseline pulmonary function testing, including the following: spirometry, gas exchange measures (room air arterial blood gas measurement, diffusion of carbon monoxide [Dco ]), plethysmography, and gas dilution lung volumes. Maximum inspiratory and expiratory flow volume curves, thoracic gas volume, and airway resistance were measured in a plethysmograph (Collins/Cybermedic Classic TCI and Body Plethysmograph; Warren E. Collins Inc; Braintree, Mass), and compared to predicted values. Resting dyspnea index determinations were performed concurrently using the Modified Medical Research Council (MMRC) dyspnea scale. 24 The modified dyspnea scale sco ring used in this study was as follows: grade  O=not troubled \vith breathlessness except during strenuous exercise; grade 1 =troubled by shortness of breath when hurrying on the level or walking up a slight hill; grade 2=walks slower than people of the same age on the l evel because of breathlessness or has to stop for breath when walking at own pace on the level; grade 3=stops for breath after walking about 100 yards or after a few minutes on the l evel; and grade 4=too breathless to leave the house or breathless when dressing or undressing. All MMRC scale assessments were performed by one nurse working directly with the patients at the time in a preoperative on-site evaluation and again at the first follow-up visit. Subsequent long-term follow-up MMRC evaluations were fllled out by mail. The procedure was explained in detail by the nurse in a uniform manner to all patients. All patients who filled out the MMRC questionnaire by mail were familiar with the process from their prior evaluations at the medical center. Pulmona1y function testing (spirometry) and dyspnea index measurements were repeated short term postoperatively ( < 120 days, mean [±SD] 33±28 days, n=l30), and longer t erm postoperatively (defined as >6 months, mean [±SD] 276±90 days, n=84).
The relationships between improvement in dyspnea postsurgery and pulmonary function at presentation (short and long term) as well as improvement in dyspnea and improvement in pulmonary function were examined. Change in dyspnea was compared to baseline  Three additional 1to 2-cm incisions were made f or standard instruments. Patients were turned to the contralateral decubitus position for separate sterile preparation and draping after completion of surgery on the initial side. The preoperative l ung CT scans and ventilation/perfusion scans were used to identifY areas of dys function al or degenerated lung tar geted for resection viith the staples.s· 11 • 23 Ring forceps manipulated the l ung into a 60-mm endoscopic s tapler ( ELC 60; Ethicon; Cincinnati) with bovine pericardium 4, 5 ( Peristrips; Biovascular; St. Paul, Minn ) or Instat (Johnson and Johnson; New Brunswick, NJ) to buttress the s taples. The staples w ere fired an average of 15 times for bilateral operations. Typically, approximately half of the upper l obe was resected. The c hest was drained with two apically p l aced chest tubes following completion of the procedure. No suction was used on the c hest drainage syste m. Heimlich valves were used for prolonged air leaks (5 days ) to facilitate earlier discharge from the hospital . Patients we re routinely transferred to the ICU for overnight observation after extubation in the operating room .

Rehabilitation
Patients did not receive preoperative rehabilitation at the Medical Center prior to LVRS. All patients underwent a similar regimen of pulmonary rehabil itation at Chapman Medical Center beginning immediately f ollowing hospital discharge. The rehab ilitation consisted of a 10-day outpatient regimen involving a multidisciplinary approach with nursing, respiratory, dietary, nutritional, psychosocial, occupational, and physical therapy. Patient education, physical exercise (walking, flexibility, and strength ening), self-monitoring, breathing retraining, and bronchial hygeine instruction were included.

RESULTS
The mean length of stay for the 139 survrvmg patients was 8.8±5.6 days (mean±SD) (range=3 to 49 days). There were six deaths (4.2% operative mortality). Causes of death are summarized in Table  1. Clinical follow-up is available for all patients. Short-term follow-up pulmonary function tests are available for 130 of 139 (94%) surviving patients and showed a mean increase in postoperative FEV 1 of 50% from baseline ( Table 2) . Spirometry and lung volumes improved significantly at the time of shortterm follow-up as summarized in Table 2. Dco did not change at short-term follow-up. Long-term follow up pulmonary function tests were available for 84 (65 %) patients and continued to show similar improvement in objective lung function results ( Ta-ble 3). Though only modest changes in Dco were seen, improvement was statistically significant at the time of long-term follow-up.
Baseline dyspnea scores averaged 3.0±0.7 (mean±SD ) and improved to 1.7±0.8 short term (p < O.OOOl ) and 1.3±0.9 long term (p<O.OOOl compared to baseline). Distribution of dyspnea score improvement short term and long term is shown in Figure 1.
Baseline FEV 1 was weakly associated with resting dyspnea score (r=0.27, p=0.06). Improvement in FEV 1 was associated with improvement in the dyspnea score (r=0.3, p=0.3, both short and long term) though considerable individual variability was seen between FEV 1 and dyspnea response (Fig 2). When improvement in FEV 1 was measured as change in FEV 1 percent predicted, the correlation with im-provement in dyspnea scores was closer (r=0.4, p<O.Ol, short term; r=0.4, p=O.l, long term) (Fig   2).
Baseline FEV 1 did correlate weakly with improvement in dyspnea score short term (r=0.3, p<0.05, short term) , with a tendency toward greater improvement in those with lower initial FEV 1 . However, this relationship was not seen at long-term follow-up (r=O.l, p=0.9, long term) (Fig 3). A similar pattern was seen when change in FEV 1 was examined as the outcome variable (Fig 3).
When the relationship between presenting measures of hyperinflation and dyspnea response was examined, there were definite trends toward decreased dyspnea response in patients with the greatest degrees of hyperexpansion. At long-term follow-up, baseline residual volume (RV) was neg-atively associated with improvement in dyspnea score (r=0.4, p<0.05), with a tendency toward greater improvement in those with lower initial RVs (Fig 4). A trend toward greater improvement in FEV 1 was also seen in patients with lower RVs (Fig 4). Greater improvement in dyspnea short and long term was seen in patients with lower presenting RV/total lung capacity (TLC ) ratios (r=0.4, p=0.02, short term; r=0.4, p<0.05, long term ) (Fig 5).
A similar but not statistically significant trend toward decreased dyspnea response was seen in patients with higher preoperative TLC and trapped gas volumes (Fig 6).
Despite the negative association between severity of hyperinflation and dyspnea response, some patients with severe hyperinflation improved substantially. There were 36 patients with preoperative  Thirty-seven patients (28% of operative patients) had minimal or no improvement in FEV 1 postoperatively (defined as < 10% increase in FEV 1 ). Yet, 10 of those 37 patients (27%) improved by two or more dyspnea scores.

DISCUSSION
This study was undertaken to investigate semiquantitative dyspnea response to LVRS . Preoperatively, dyspnea scores were significantly associated with baseline FEV 1 . However, as expected, there was conside rable variability between individual resting dyspnea scores and baseline FEV 1 . Evaluated as a group, improvement in dyspnea scores postoperatively was also loosely associated with improvement in FEV 1> and more closely with improvement in FEV 1 percent predicted. Again considerable individual variability was seen between dyspnea and FEV 1 improvement.
A substantial fraction (27%) of patients with minimal improvement in FEV 1 had marked improvement (>2 U) in dyspnea scores. This may be the result of placebo effect or from physiologic response not reflected in FEV 1 changes. This subgroup of patients with limited FEV 1 response and marked dyspnea improvement illustrates potential limitations of using objective improvement in FEV 1 as the sole outcome measure, and difficulties in developing absolute selection criteria cutoffs.
-..J u ..J .... In this preselected patient group (required to have hyperinflation), those with the greatest degrees of preoperative hyperinflation (assessed as RV, TLC, or RV trLC ratios) appeared to have less improvement in dyspnea following LVRS than those with lesser degrees of hyperinflation. This somewhat surprising result contrasts with a more general lack of correlation between presenting physiologic function and improvement following L VRS when change in FEV 1 is used as the primary outcome variable in this patient series. The negative correlation between hyperinflation and dyspnea response in this study could be artifactual, given the relatively weak associations seen. However, in the extremes of hyperinflation , both improvement in dyspnea and FEV 1 appear slightly reduced. These findings could reflect the degree of underlying emphysema and the amount of functional lung available postoperatively. Such findings could suggest that there may be limits on the severity of preoperative disease for optimal response.
Despite the trend toward reduced dyspnea response with greater degrees of hyperinflation, >25% of patients with the highest degrees of hyperinflation did improve significantly in dyspnea scores and objective pulmonary function measures. Thus, it remains difficult to set firm limits on hyperinflation for selection criteria from these data.
The findings of this study are limited by the fact that the MMRC dyspnea scale is subjective and relatively simplistic. Other dyspnea indexes with more specific applicability toward LVRS patients may yield different results. O' Donnell et al 19 investigated associations between relief of dyspnea follow-922 ing LVRS and bullectomy in eight patients and pulmonary function using transitional dyspnea index measurements. They found very close correlation between improvement in transitional dyspnea index and change in FVC (r=0.94, p<0.05) or change in FEV 1 (r=0.77, p<0.05) in independent analyses. It is uncertain why they found much closer association between subjective and objective response than was seen in our study. Use of the transitional dyspnea index rather than the modified dyspnea index may explain some of the differences. Additionally, there were a very small number of patients described by O'Donnell et al 19 with some undergoing decompressive bullectomy that may have increased the observed correlations. Future studies using multiple subjective outcome measurement tools may be needed to clarify these issues .
The preselection of patients based on specific radiographic and pulmonary function criteria for inclusion in our study protocol 5 · 11 narrows the range of disease presentation of patients enrolled. This could predispose to misleading apparent relationships between presenting variables and outcomes due to the limited parameter ranges for the study variables. Finally, the proportion of patients missing long-term follow-up in this study may also have biased results in our study. Patients in this study did not undergo rehabilitation until the postoperative period. Thus, effects of postoperative rehabilitation would be included in the overall protocol subjective response assessment of these patients. Different results may be seen in programs involving preoperative rehabilitation.
Despite these limitations, some trends appear to emerge from this study. Overall, dyspnea improvement was associated with improvement in FEV 1 following LVRS. However, the correlation was weak, with considerable individual variability between dyspnea response and spirometry improvement. Therefore, careful evaluation of subjective and objective outcomes will be needed to assess LVRS response. If these findings are confirmed in future studies, selection criteria may need to incorporate these factors into the decision-making process regarding optimal surgical candates.