INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Lung volume reduction surgery (LVRS) has shown promise in improving functional capacity in patients with emphysema. While several studies (1) have demonstrated improvements in pulmonary function as measured by spirometry, the effects of LVRS on pulmonary hemodynamics have not been as intensively studied. Echocardiographically determined fractional change in the right ventricular (RV) area was demonstrated to increase after LVRS (2), suggesting an improvement in RV systolic function, possibly due to a reduction in pulmonary vascular resistance (PVR), although this was not measured. Beneficial hemodynamic effects could stem from relief of hypoxia, re-expansion of compressed pulmonary vessels, allowing recruitment of capillaries and increased venous return, leading to augmentation of RV preload. However, the LVRS procedure could also result in deformation of vessels, removal of a significant amount of the pulmonary vascular bed, and an increase in pulmonary artery (PA) pressure due to increased venous return (4). These adverse effects could have an influence on the ultimate clinical outcome. In this study we report preliminary pulmonary hemodynamic findings in a group of patients studied both before and after LVRS. We hypothesized that PA pressures would be reduced after the LVRS procedure and that improvement in symptoms would correlate with improvement in pulmonary hemodynamics.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

We studied nine consecutive patients with emphysema before and after bilateral LVRS. All patients were studied within 2 wk prior to and 3 mo after LVRS. All subjects were severely dyspneic on a maximum medical regimen and had not smoked for at least 6 mo prior to study. Patients underwent preoperative hemodynamic evaluation after standard pulmonary function testing, a treadmill stress test with thallium-201 perfusion scanning, ventilation-perfusion lung scanning, and high-resolution computed tomography of the lungs. Selection criteria for undergoing surgery included FEV₁ less than 40% predicted, acute increase in FEV₁ of less than 12% following bronchodilators, residual volume more than 130% predicted, arterial PCO₂ less than 50 mm Hg, systolic PA pressure less than 40 mm Hg and mean PA pressure less than 35 mm Hg, lack of evidence of coronary ischemia on thallium-201 stress scan, lack of giant bullous disease, and lack of clinically important other comorbid conditions. Dyspnea preoperatively was assessed using the modified Mahler scale, which incorporates multidimensional ratings in three domainsmagnitude of task, magnitude of effort, and functional impairmentinto an overall score of chronic activity-related shortness of breath. Changes in symptom intensity over time in these domains were also rated. The dyspnea index and transitional dyspnea index were calculated as the sums of the scores in each domain (5). For the Mahler dyspnea index, the lower the score, the lower the perceived functional ability and the greater the dyspnea. Similarly, for the Mahler transitional dyspnea index, the lower the score, the less the rated improvement. All patients underwent preoperative and postoperative pulmonary rehabilitation for 6-wk periods. All patients had returned to daily life and had completed the postoperative rehabilitation program by the time the postoperative hemodynamic assessment was done. The protocol was approved by the Human Subjects Review Committee of Long Island Jewish Medical Center, and each patient provided written informed consent.

Surgical Approach

LVRS was performed for all patients through a median sternotomy, resecting 40-60% of the upper lobes of both lungs with use of a linear stapling device and application of bovine pericardium to seal leaks.

Hemodynamic Testing

All patients underwent a full right heart catheterization under fluoroscopic guidance in the cardiac catheterization laboratory. The catheter was inserted under local anesthesia through the right antecubital vein or the right internal jugular vein. Standard pressure measurements were made and oxygen saturations were measured in each chamber. All pressure measurements were made at end-expiration with the transducer carefully leveled to mid-right atrium. Normal values were taken as per Parmley and Talbot (6). Oxygen saturation was determined from the arterial side with a pulse oximeter (Nellcor, Inc., Hayward, CA). Resting arterial blood gas tensions and pH were measured with the patient breathing ambient air before and after surgery. Cardiac outputs were determined using thermodilution technique in triplicate (Baxter Healthcare Corp., Irvine, CA). In one patient, a postoperative cardiac output could not be obtained because of a hypotensive and bradycardic episode just prior to these measurements. In five of the nine patients left ventricular ejection fraction (LVEF) was obtained before and after surgery by multiple gated radionuclide angiography (MUGA). This was done immediately before undergoing catheterization. Since both the MUGA and catheterization were performed under identical conditions within 1 h of each other, we used the LVEF measured by MUGA to calculate left ventricular (LV) volumes. Using the stroke volume obtained from the catheterization, calculations of LV end-diastolic (LVEDV) and end-systolic (LVESV) volumes were made for the six patients who had both a cardiac output and MUGA EF determination before and after surgery. Stroke volume (SV) was calculated as cardiac output/heart rate, LVEDV was calculated as SV/EF, and LVESV was calculated as LVEDV-SV.

Statistical Measurements

Data were compiled as mean ± SD. Student's t test for paired variables was used to assess statistical significance. The null hypothesis was rejected at the 5% level.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The study group consisted of three males and six females (eight Caucasians and one African American) with a mean age of 64.4 ± 7.3 yr (range, 52 to 74 yr). Computed tomography of the chest demonstrated emphysema that was largely inhomogeneous and primarily apical in distribution. The results of the hemodynamic studies before and after surgery are shown in Table 1. Mean arterial pressure, heart rate, cardiac output, cardiac index, right ventricular end diastolic pressure (RVEDP), SV, systemic vascular resistance, PVR, and baseline arterial PO₂ were unchanged before and after surgery. FEV₁ rose significantly postoperatively (0.64 ± 0.15 to 0.99 ± 0.25 L, p < 0.006). Individual data for pulmonary hemodynamics are tabulated for each patient (Table 2).

Significant increases were noted in pulmonary artery systolic pressure (PA_s, 35.7 ± 6.6 preoperatively to 47.9 ± 12.4 postoperatively, p = 0.0025) and pulmonary artery mean pressure (PA_m, 26.5 ± 5.1 preoperatively to 31.8 ± 7.7 postoperatively, p = 0.033) (Figure 1). Preoperatively, no patient had a PA_s pressure higher than 42 mm Hg and seven patients had pressures that were 40 mm Hg or less. After LVRS, five patients had PA_s of 45 mm Hg or greater, of whom three had an increase to 60 mm Hg or greater (Figure 1, upper panel ). Six patients had an increase in PA_m (Figure 1, lower panel ) to an absolute value of 27 mm Hg or greater and of these, three had an increase to 35 mm Hg or greater. The mean increase in pulmonary artery diastolic pressure (PA_d) was significant (19.3 ± 3.4 preoperatively to 23.6 ± 6.7 postoperatively, p = 0.028) (Figure 2, upper panel ).

View larger version (18K): [in this window] [in a new window]   Figure 1.   The pulmonary artery (PA) systolic pressure (top panel ) and PA mean pressure (bottom panel ) were both significantly increased after LVRS.

View larger version (15K): [in this window] [in a new window]   Figure 2.   The pulmonary artery (PA) diastolic pressure (top panel ) was significantly increased after LVRS. There was no change in the pulmonary artery occlusion pressure after surgery (lower panel ).

There was no overall change in the pulmonary artery occlusion pressure (PAOP) (14.1 ± 4.8 preoperatively to 15.2 ± 6.8 postoperatively). Seven patients had preoperative PAOP of 12 mm Hg or greater (Figure 2, lower panel ). In one patient, PAOP declined to 8 mm Hg from a preoperative value of 12 mm Hg. In two other patients with preoperative values less than 12 mm Hg, the PAOP increased to 12 mm Hg postoperatively. Although in six of eight patients the PVR values increased, the mean change (208.6 ± 86.8 preoperatively to 251.8 ± 149.4 postoperatively) did not reach statistical significance (Figure 3). The person with the lowest postoperative PVR (Table 2, patient 2) had the highest PAOP. This is because of the way that PVR is calculated such that as PAOP increases relative to pulmonary artery pressure, PVR decreases. This particular individual developed clinical signs of left ventricular failure postoperatively, although postoperative LVEF was normal. This indicates some diastolic dysfunction postoperatively, possibly related to mechanical heart-lung interactions. Although we cannot rule out the development of cardiac ischemia, this appears less likely since all patients were screened for ischemic heart disease before surgery.

View larger version (16K): [in this window] [in a new window]   Figure 3.   Increase in pulmonary vascular resistance (PVR) after LVRS was not significant, although it rose in six of nine patients.

Paired LVEF and PAOP data were available for six patients (Figure 4). There was no indication that elevated PAOP was associated with depressed LV systolic function as measured by LVEF. Preoperatively, a negative correlation was noted between PA_s and the Mahler dyspnea index, demonstrating that the higher the PA_s, the greater the degree of dyspnea (Mahler dyspnea index = 6.6  0.1 PA_s, r = 0.81, p = 0.008). However, there was no correlation between the change in PA_s from the preoperative to the postoperative period and the transitional dyspnea index. There was no correlation between resting cardiac output and dyspnea preoperatively and changes in cardiac output and dyspnea with surgery.

View larger version (15K): [in this window] [in a new window]   Figure 4.   Relationship between left ventricular ejection fraction (LVEF) and pulmonary artery occlusion pressure (PAOP) for the six patients in whom paired LVEF and PAOP values were available. The direction of the arrows indicates the change from the preoperative to postoperative measurements.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We demonstrated an increase in PA pressures compared with preoperative baseline values 3 mo following surgery in our patients who underwent LVRS. This was not accompanied by a decrease in resting cardiac output or changes in other indices of LV function. The increase in PA pressures postoperatively did not correlate with change in symptomatic status after surgery. We observed that the patients with the highest initial PA pressures demonstrated the greatest increase in PA pressure 3 mo postoperatively, with some into the range that would have met our exclusion criteria for surgery.

LVRS may become an important option in the therapy of patients with emphysema. Several papers have confirmed that clinical improvement, at times quite marked, occurs in some patients who have undergone LVRS (1). An improvement in elastic recoil is one likely mechanism for the improvement (2, 4, 7), but improvement in the configuration of the inspiratory muscles and decreased airway resistance may also play a role in the decrease in symptoms after this surgery. Patients with the most functional limitation may be those who demonstrate the lowest mixed venous oxygen tension and resting cardiac output (8). Hence, the development of pulmonary hypertension in our patients might have been expected to worsen symptomatic status. This was not observed in our study, however.

Several mechanisms may contribute to the increase in PA pressures following LVRS. Since six of our nine patients demonstrated an increase in PVR, part of the reason for this effect may be a decrease in cross-sectional area. Thus, in these patients a postoperative increase in PA pressures could be at least partly attributed to an increase in PVR. Only 40-60% of the upper lobes was removed, and most of this was in avascular lung as judged by perfusion lung scan and at surgery. Hence, it does not appear likely that simple resection of lung tissue could account for the increase in PVR. Some of the increase in PVR could have been due to reactive or hypoxic vasoconstriction; however, arterial oxygen saturation levels during the preoperative and postoperative catheterizations were maintained at greater than 90%, and resting hypoxemia was not observed. A postoperative change in the balance of pulmonary vascular reactivity mediated by vasodilatory and vasoconstrictor substances may have altered PVR. Finally, since angiography was not performed, we cannot exclude kinking of the pulmonary vessels occurring as a result of surgical changes in pulmonary architecture. Of the two patients showing a decline in PVR, one developed congestive heart failure after surgery, probably due to diastolic dysfunction, and one demonstrated a large increase in cardiac output at rest postoperatively. In these two cases, the elevation in PA pressure was attributable to changes in the underlying hemodynamic milieu: diastolic dysfunction in the first case and possibly an increase in venous return in the second.

Originally, Nakhjavan and colleagues (9) and later, Dantzker and Scharf (4) suggested that flattening of the diaphragm and hyperinflation could attenuate venous return, thus leading to a decrease in resting cardiac output. Indeed, Mithoefer and coworkers (7, 8) suggested that decreased resting cardiac output was a major correlate of symptoms in chronic obstructive pulmonary disease (COPD). Our data show that, despite no change in resting cardiac output, substantial improvement in dyspnea occurred with surgery. Therefore, resting cardiac output is not predictive of changes in dyspnea. In this study, we did not assess the role of exercise-associated hemodynamic changes following LVRS. It is possible that beneficial hemodynamic effects may be demonstrated during exercise after surgery.

Although there was no change in PAOP, the PAOP was elevated in seven of our nine patients preoperatively and remained elevated in eight patients postoperatively. While in patients without COPD an elevated PAOP indicates elevated LV preload, patients with COPD may demonstrate elevated PAOP with no evidence of LV dysfunction. This is probably related to pulmonary hyperinflation and an increase in juxtacardiac or intrathoracic pressures (10), effectively decreasing LV diastolic compliance. In our patients, paramediastinal lung was not removed. Thus, although overall lung volume decreased, mechanical heart-lung interactions in the cardiac fossa appeared to have been unaffected.

In keeping with the above, there was no evidence that elevated PAOP represented systolic dysfunction in our patients, since LVEF was not depressed. Diastolic dysfunction was not routinely assessed and echocardiograms were not performed. As noted above, however, in one case diastolic dysfunction was suspected on clinical grounds.

We examined the relationship between dyspnea and the PA pressures. Preoperatively, the higher the PA pressure, the greater the degree of dyspnea, as evidenced by an inverse relationship with the Mahler dyspnea score. However, with surgery there was no correlation between the Mahler transitional dyspnea index and the rise in PA pressures. Therefore, PA pressures rose without a corresponding increase in dyspnea. The reason for the loss of correlation between change in PA pressure and symptoms in the postoperative period is not clear. Many factors contribute to dyspnea. Possibly the beneficial effects on respiratory muscle function and lung mechanics outweighed those due to elevated PA pressures over the short term.

LVRS seems to be a significant new therapeutic tool in the treatment of emphysema. Nonetheless, this study demonstrated that hemodynamics are changed in a direction that may have long-term detrimental effects. While more study is required to investigate the mechanisms of the short-term improvement in symptoms, the hemodynamics should be carefully followed as a potential source of deterioration following the initial improvement.