INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Lung resection is a potentially curative treatment for many patients with lung cancer, trauma, or chronic localized infection. Major lung resection has also been utilized as a model to study the physiologic consequences of restrictive lung disease (1). There are normally large functional reserves of lung diffusing capacity, and theoretically, > 50% of the lung tissue could be lost without significant exercise limitation caused by gas-exchange derangement (2). In reality, however, patients who have undergone pneumonectomy (PNX) demonstrate a markedly reduced exercise capacity, roughly in proportion to the fraction of lung removed (3, 4). The impairment could not be explained by abnormalities in gas exchange or pulmonary vascular resistance (3, 4). Rather, we previously showed that the most important sources of limitation are a proportionally reduced maximal stroke index (SI) (4), increased respiratory muscle energy requirements at a given ventilation, and reduced maximal ventilatory power generation (5). A low maximal cardiac index (CI) further aggravates the limitation imposed by respiratory muscle energetics, by intensifying the competition between respiratory and nonrespiratory muscles for the diminished total body O₂ delivery following PNX. Potential sources of an impaired maximal SI after PNX include: (1) a greater pulmonary vascular resistance; (2) restricted cardiac filling due to displacement of the mediastinum and anatomic distortion of the cardiac fossa (5); and (3) deconditioning due to a sedentary life-style after surgery. Cardiac impairment due to deconditioning should be reversible with intensive exercise training, whereas impairment caused by other factors should be irreversible. To address the reversibility of exercise impairment, we examined the cardiopulmonary response during exercise before and after physical training in patients who had undergone PNX, in comparison with that in average, age-matched, healthy nonsmoking subjects.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

The protocol for the study was approved by the Institutional Review Board for Human Research of the University of Texas Southwestern Medical Center. Informed consent was obtained from all subjects. Patients were recruited from the Chest Medicine and Thoracic Surgery Clinics at Parkland Memorial Hospital, as well as from outlying hospitals in Dallas, Texas. Seven patients were studied at 2 mo to 10 yr after left or right PNX for treatment of localized lung cancer (mean period after surgery: 58 mo). All patients were ex-smokers who had stopped smoking before PNX. They were considered surgically cured and were free of overt residual disease on physical examination and chest X-ray. None received adjuvant radiation or chemotherapy. Only patients with relatively normal remaining lungs, as judged by spirometric results > 70% of those predicted for one lung, were eligible for study. Patients experiencing significant postoperative complications or who had other overt cardiopulmonary diseases were excluded. Eight healthy, never-smoking subjects were recruited from the general public as controls.

Apparatus

Routine spirometry was conducted with a rolling-seal spirometer (Ohio Instruments, Madison, WI). Exercise testing was performed on a mechanically braked Monark bicycle ergometer. The mouthpiece of the spirometer was connected to a turbine flowmeter (Sensormedics, Anaheim, CA), a pneumatically operated three-way-sliding valve (No. 8500; Hans Rudolph, Kansas City, MO), and a two-way respiratory valve (No. 2700; Hans Rudolph). Gas concentrations at the mouth and in the expired air distal to a mixing chamber were measured continuously with a mass spectrometer (MGA 1100; Perkin-Elmer, Oak Brook, IL). Ventilation (E), O₂ uptake (O₂), CO₂ output (CO₂), respiratory rate, tidal volume (VT), and heart rate (HR) were measured and averaged every 10 breaths. All volumes were expressed at body temperature, ambient pressure water-saturation (BTPS) conditions. An electrocardiogram (ECG) was continuously monitored during exercise. All signals were digitized by computer at 50 Hz.

Protocol

Studies were done on separate days. Some of the subjects were admitted to the General Clinical Research Center at our institution for close monitoring during test days. On Day 1, routine spirometry was conducted. Maximal respiratory pressures were determined with a manometer. Ventilatory capacity was assessed from maximal voluntary ventilation over a period of 15 s and maximal sustained ventilation over a period of 4 min under isocapneic conditions. Maximal O₂ uptake (O_2max) was determined with a continuous incremental protocol, increasing the workload by 25W every 2 min until the respiratory quotient exceeded 1.0 and the subject voluntarily terminated exercise because of exhaustion, or until a plateau in O₂ was reached with respect to workload.

On Day 2, a 22-gauge radial artery catheter was inserted under local anesthesia and connected to a vascular transducer and carrier amplifier. O₂, CO₂, E, VT, respiratory rate, HR, arterial blood pressure, and ECG were continuously recorded. Subjects exercised for 4 min at each of two to four preselected constant workloads, up to the highest that could be sustained for 3 to 4 min (80% of measured maximum). Steady-state conditions, judged by the constancy of O₂ uptake, HR, and respiratory rate, were generally reached by the end of the third minute at each workload. During the last minute of exercise, 5 ml of arterial blood was drawn anaerobically for analysis of routine blood gases (ABL3; Radiometer; Copenhagen, Denmark) as well as O₂ saturation (OSM3; Radiometer), hemoglobin concentration (Model DB; Beckman Instruments, Fullerton, CA), hematocrit, and lactate concentration (Yellow Springs Instruments, Yellow Springs, OH). After blood was drawn, the subject inspired via the automatic pneumatic valve assembly, from a selected end-expiration up to total lung capacity (TLC). The inspired gas mixture contained 9% He, 0.3% C¹⁸O, and 0.6% C₂H₂ in either 30% O₂ with a balance of N₂, or in 90% O₂. The volume of gas inspired was measured independently, using a bag-in-a-box system with a rolling-seal spirometer. The subject rebreathed this gas mixture in and out of an anesthetic bag for 16 s before termination of exercise. Gas concentrations at the mouth were monitored continuously during rebreathing. Respiratory frequency during rebreathing at rest was synchronized at 30 breaths/min with a tape recorder. During exercise, the subjects were allowed to breathe at their spontaneously chosen frequencies. At least 20 min of rest was allowed between exercise periods, or until HR and respiratory rate had returned to baseline.

On Day 3, endurance for submaximal exercise was determined as the time to exhaustion at 80% of measured maximal power. Subjects were not connected to the mouthpiece during this measurement.

Measurements during Rebreathing

The rebreathing technique has been described previously (4). Time lag between computer activation of the automatic valve at end-expiration and complete switching was 20 ms. System volume during rebreathing was estimated from helium dilution. End-expiratory lung volume (EELV) was calculated from the system volume by subtracting rebreathing bag volume, apparatus dead space (90 ml), and the volume inspired during the computer-valve lag time. This latter volume was negligible at rest, but increased during exercise. Diffusing capacity of carbon monoxide (DL_CO) and cardiac output were estimated from the exponential disappearances of end-tidal C¹⁸O and C₂H₂ concentrations, respectively, relative to He. The time-zero correction obtained from the CO disappearance curve was applied to the C₂H₂ disappearance curve according to the method of Sackner and associates (6, 7). Septal tissue volume was estimated from the extrapolated intercept of the C₂H₂ disappearance curve to time zero. Only the linear portion of the semilogarithmic curve was analyzed; the first three breaths and breaths beyond 12 s during exercise were routinely discarded.

Exercise Training

After completion of the initial evaluation, a supervised exercise training program was begun, consisting of exercise in the laboratory on a bicycle ergometer at a workload equivalent to 65% of the subject's measured O_2max for 30 min per day, 5 d per week, over a period of 8 wk. Each training session was preceded by a 5 min warm-up period and followed by a 5 min cool-down period at minimal workload. HR was recorded by with an ECG and blood pressure was monitored every 5 min with a sphygmomanometer, respectively. Workload was individually adjusted every week to ensure maintenance of an HR during exercise of about 80% of maximal predicted HR. After the training period, all measurements were repeated with the subject at rest and during exercise, as described earlier.

Statistical Analysis

Data were normalized for body surface area (BSA) and expressed as mean ± SEM. Comparison of groups with one another before or after training was done with one-way analysis of variance (ANOVA). Results before and after training were compared through repeated-measures ANOVA followed by Bonferroni's correction for multiple comparisons and Wilcoxon's signed rank test, using a commercial software package (STATVIEW 4.5, Abacus Concepts, Berkeley, CA). A value of p < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pretraining Data at Rest

Demographic data are presented in Table 1. Patients who had undergone PNX were slightly older than control subjects. There were no significant differences in BSA between groups. Five patients had had right PNX and two had had left PNX. We assumed the partition of normal lung volume, ventilation, and perfusion to be 45% and 55% to the left and right lungs, respectively. Resting data are shown in Tables 2 and 3. The remaining lungs in patients were relatively normal as confirmed by the data for FEV₁ and FVC, which were within normal range for those expected for one lung. Compared with those of normal subjects, lung volumes were significantly lower in patients who had undergone PNX, and respiratory rate was significantly higher. Ventilatory capacities were lower by 54%. Cardiac and stroke indexes were lower by 28% and 31%, respectively. DL_CO was lower by 46%, predominantly because of a reduction in alveolar volume. Mean septal lung-tissue volume was lower by 41%. The difference in arterial and venous O₂ content was greater in patients than in normal subjects.

Pretraining Data during Exercise

At maximal exercise, patients achieved a lower HR than did normal subjects; however, the average HR in patients still exceeded 80% of their predicted maximal HR (Table 4). Patients also achieved a significantly lower maximal power and O_2max than did normal subjects at peak exercise (Table 4). The slopes of the relationship between CI and O₂ were similar for the two groups, but patients had a lower CI at a given level of O₂, and achieved a much lower maximal CI and maximal stroke volume (Figure 1). Arterial PO₂ and O₂ saturation were significantly lower in patients than in control subjects; alveolar-arterial O₂ tension gradient (AaDO₂) was higher in patients, but intersubject variability was large and hence the difference between groups did not achieve statistical significance. There was no significant difference between the groups in arterial hemoglobin concentration, arterial O₂ content, or arteriovenous difference in O₂ content. Mean arterial blood pressure was similar in the two groups. From rest to exercise, CI increased by 121% in patients and by 153% in normal subjects. Compared with normal subjects, maximal CI in patients was reduced by 37%, and SI was reduced by 23%. Limitation in both maximal HR and SI contributed to the lower CI in patients. DL_CO increased from rest to exercise by 29% and 43% in PNX and normal subjects, respectively. In patients who had undergone PNX, DL_CO at peak exercise was significantly lower than in control subjects, by 52%.

View larger version (19K): [in this window] [in a new window]   Figure 1.   Relationship between CI and O2 before training. Pooled data from all subjects. Open circles = normal subjects. CI = 6.36  O2 + 2.09, r = 0.87. Closed circles = patients after PNX. CI = 5.04  O2 + 1.32, r = 0.84.

Posttraining Resting Data

Ventilatory function and lung diffusing capacities did not change with training in either group (Tables 2 and 3). There were no changes in resting arterial blood gases or cardiac or stroke indexes from before to after exercise training in either group. Hemoglobin concentration was unchanged. After training, resting arterial PCO₂ was slightly but significantly higher in PNX patients than in control subjects.

Posttraining Maximal Exercise Data

After training, O_2max improved by 17% in patients and 24% in normal subjects as compared with pretraining values (Table 4). Maximal workload increased in both groups, but the increase was statistically significant in PNX patients only. Arterial PCO₂ was significantly higher and arterial PO₂ lower in PNX patients at peak exercise, indicating a ventilatory constraint to exercise as consistent with our previous findings (5). Arterial O₂ saturation was unchanged in patients but decreased slightly in control subjects. Blood lactate concentration at peak exercise did not change with training in either group. Maximal CI and maximal SI increased significantly in normal subjects (by 8% and 36%, respectively), but not in patients who had undergone PNX. Training increased endurance for submaximal exercise similarly in both groups (by 433% in patients and 547% in normal subjects). Arteriovenous difference in O₂ content at maximal exercise increased significantly after training in both groups.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Summary of Findings

Our present data again show that both ventilatory and cardiac constraints to exercise develop in patients after PNX. As compared with that of normal control subjects, maximal oxygen uptake in patients was reduced by approximately 50%. Ventilatory capacity and DL_CO were also reduced and arterial PCO₂ was higher during exercise. CI was impaired at rest and exercise. These ventilatory and cardiac constraints did not prevent the patients from achieving a beneficial response to exercise training, evidenced by the 36% increase in maximal power output and 17% increase in maximal oxygen uptake, and by a 23% increase in peripheral oxygen extraction after training. However, maximal CI and SI failed to increase after training in patients.

Combined Cardiopulmonary Limitations to Exercise after Pneumonectomy

DeGraff and colleagues (3) systematically examined the sources of exercise limitation in patients after PNX for tuberculosis, and concluded that the major limiting factor is a reduced maximal cardiac output. Ventilatory impairment could not explain the nearly 50% reduction of O_2max in their patients. Diffusion impairment was also not significant, except in those patients who had had resection of more than 67% of lung. Three weeks of endurance training on a treadmill failed to improve O_2max in three patients; however, all of their patients had significant residual disease involving the remaining lung, thus limiting the interpretation of these results. Subsequently, we found that patients who had undergone PNX for cancer and who had relatively normal remaining lungs had a preserved ability to augment DL_CO from rest to exercise (4); this mechanism of compensation protected against the development of a significant diffusion limitation during exercise. Indeed, most patients showed only a modest decrease in arterial O₂ saturation during peak exercise. However, CI was significantly lower at any level of O₂, and maximal CI was reduced by nearly 50% in patients who had undergone PNX as compared with normal control subjects (5). At the same time, mechanical work of breathing was increased at any level of ventilation; hence, respiratory-muscle energy demand was greater in the patients (5). The combination of diminished O₂ delivery and a greater energy requirement of ventilation leads to an intensified competition between respiratory and nonrespiratory locomotive muscles for available O₂ substrate. Oxygen supply to locomotive muscles must be curtailed if ventilation is to be sustained during heavy exercise. Greater anaerobic metabolism and production of lactate by peripheral muscles results in greater CO₂ production which further aggravates the ventilatory burden. Hence, exercise capacity becomes limited in PNX patients by a vicious cycle of O₂ supply-demand imbalance attributable to combined ventilatory and cardiovascular factors.

In accord with our present findings, we previously reported that respiratory muscle dysfunction in patients who had undergone PNX was irreversible (5). We found that the maximum power generated by respiratory muscles was lower in the PNX patients than in normal subjects. A 4-wk selective respiratory-muscle endurance-training program significantly improved maximal respiratory pressures in both patients and control subjects without improving O_2max or maximal cardiac output in either group. Training improved ventilatory endurance in normal subjects but not in patients, suggesting a fixed anatomic and mechanical restriction of the thoracic pump as the major source of ventilatory dysfunction. However, it is important to note that the various aspects of pulmonary dysfunction (greater respiratory-muscle power requirement and reduced DL_CO, lung volume, and ventilatory capacity) did not preclude patients from improving their aerobic capacity through enhancement of peripheral O₂ utilization in response to exercise training. Furthermore, the present data raise the possibility that an irreversible impairment in CI, by limiting O₂ delivery to respiratory muscles, may be partly responsible for the irreversibility of ventilatory dysfunction after PNX. If maximal CI could be increased after PNX, it might be possible to increase maximal respiratory-muscle blood flow and enhance maximal ventilatory power output during exercise.

Sources of Cardiac Impairment after PNX

Comparison of physiologic responses in dogs and humans studied in our laboratory (4, 5, 8) first suggested physical fitness as a potential determinant of aerobic capacity after PNX. In the dog model, animals underwent exercise training before and after right PNX, except in the immediate postoperative period (8, 10); in these animals O_2max was reduced by only 14% and maximal cardiac output by less than 30% after PNX. In contrast, most patients were not physically fit before PNX, and may have adopted an even more sedentary lifestyle after PNX. The present results indicate that patients were not more deconditioned than average normal subjects, and physical fitness cannot account for the cardiac impairment in patients who have undergone PNX. Further consideration suggests the following possible explanations for the fixed cardiac impairment in PNX patients:

1. An increased pulmonary vascular resistance may have limited SI. Mean pulmonary artery pressure () and pulmonary vascular resistance have been reported to be 20% to 27% higher in PNX patients than in normal subjects at rest and during submaximal exercise (15). However, the highest reached by patients never exceeded that achieved in normal control subjects during maximal exercise at sea level (about 45 mm Hg) (16). In contrast, normal subjects exercising in hypobaric hypoxia can continue to increase cardiac output up to a of nearly 60 mm Hg [17]. In dogs after right PNX, mean reached more than 70 mm Hg during exercise, nearly 50% higher than in control animals (1). Hence, an increased pulmonary vascular resistance alone cannot fully explain the marked reduction of stroke volume in PNX patients. These data suggest that in PNX patients the right ventricle is also unable to generate the high pressure necessary to overcome a higher ventricular afterload.
2. Pleuromediastinal adhesion and organized hemofibrothorax are largely absent in the dog model of PNX, but are present to variable degrees in PNX patients. In all of our patients the chest X-ray showed the typical finding of a displaced mediastinum, with the cardiac shadow abutting the rib cage. However, the cardiac silhouette and fibrous adhesions could not be distinguished on a plain X-ray. In another group of eight patients after PNX (including one patient from the present study), we obtained magnetic resonance imaging of the thorax and confirmed the marked mediastinal displacement in all patients. The ipsilateral hemithorax is partly collapsed, in association with an elevated and relatively immobile hemidiaphragm. The cardiac fossa, instead of being surrounded by the normal compliant lung, becomes bounded by relatively noncompliant structures (rib cage, variable degrees of pleuromediastinal adhesions, an immobile hemidiaphragm, and an overexpanded remaining lung), which surround both the left and right ventricles. A representative image is shown in Figure 2. Such anatomic distortions reduce the compliance of the cardiac fossa and may potentially restrict ventricular filling, preventing utilization of the Starling mechanism to overcome the increase in ventricular afterload. Such impairment would become accentuated during exercise and would not respond to exercise training. We suspect that anatomic restriction is a major determinant of cardiac function after PNX; however, the precise contribution to overall cardiac impairment of increased pulmonary vascular resistance and adhesion and hemofibrothorax as discussed here cannot be determined from the present study.

View larger version (168K): [in this window] [in a new window]   Figure 2.   Coronal magnetic resonance image of the thorax from a patient after left PNX, showing mediastinal displacement, distortion of the left hemithorax, and encasement of the left ventricle by fibrotic tissue and by relatively rigid and immobile structures.

Peripheral Oxygen Utilization after Training

Peripheral oxygen extraction is often impaired in patients with cardiopulmonary disease. Among functional, metabolic, and structural abnormalities observed in peripheral skeletal muscles that may be responsible for the impairment are a lower static and dynamic endurance, lower glycolytic enzyme and glycogen contents, lower aerobic oxidative enzyme activity, decrease in slow-twitch type 1 fibers, increase in fast-twitch type 2b fibers, and muscle atrophy (18). Decreased muscle blood flow cannot explain all of the muscle changes (19). Many of these abnormalities are consistent with the effects of deconditioning and are reversible after physical training (20). Belman and Kendregan (21) reported that patients with significant chronic obstructive lung disease failed to show the classic response in muscle enzymes after exercise training, despite a significant improvement in endurance for submaximal exercise in the trained muscle. In patients with chronic heart failure, skeletal-muscle metabolism and function improved after training independent of changes in muscle mass, muscle blood flow, or cardiac output (22, 23). Our data are consistent with these previously reported findings of dissociated central and peripheral responses; peripheral impairment in O₂ utilization is reversible with training even in the absence of a central response in O₂ delivery. In addition, endurance for submaximal exercise improved to the same extent in patients and control subjects. Hence, even if cardiac output is restricted by cardiopulmonary disease, considerable clinical benefit can still be derived from exercise training. The same principle should be applicable not only to patients who have had PNX, but also to those with other forms of restrictive lung disease.

In conclusion, maximal and submaximal exercise performance in patients who have undergone PNX can be improved through endurance exercise training. However, improvement is limited by a fixed maximal SI, in accord with an increased right ventricular afterload and/or anatomic and mechanical restriction of the cardiac fossa. Despite the irreversible cardiac limitation, physical training resulted in significant improvements in peripheral oxygen utilization and endurance of submaximal exercise in our PNX patients. Future efforts should be directed toward operative and postoperative measures that can prevent or limit the development of pleuromediastinal adhesions and maintain a normal compliance of the cardiac fossa.