Methods

Top Methods Results Discussion Author & Article Info References

Patients

Patients who were admitted to our institution because of symptomatic malignant pleural effusion and who required therapeutic thoracentesis were candidates for this 3-year prospective study. Malignant effusion was diagnosed only if malignant cells were found in a pleural biopsy specimen or effusion fluid. Exclusion criteria were 1) previous effusion of chemotherapy-sensitive cell types [for breast cancer, ovarian cancer, small-cell cancer of the lung, or lymphoma, for example], except when the patients had undergone a course of chemotherapy that had not eliminated the effusion; 2) chemotherapy or radiotherapy previously administered or planned within 2 months; 3) life expectancy of 1 month or less; and 4) loculated pleural effusion.

Measurements and Interventions

Figure 1 is a diagram of the device that was used to measure the elastance of the pleural space. Measurements were done while patients were in a sitting position. Under ultrasonographic guidance, an appropriate intercostal space was selected. After a local anesthetic was administered, 20 mL of pleural effusion was drawn into a heparin-rinsed plastic syringe for the measurement of glucose level and pH and for cytologic examination. Then, a 16-gauge intracatheter needle was inserted into the pleural space. The needle was attached to a three-way stopcock with two extension tubes that led to a central venous pressure monitor and a 50-mL syringe. The whole system was filled with normal, heparinized saline. Using a leveling rod, we marked a zero point on the central venous pressure monitor at the level of the puncture site. The pleural pressure, defined as the mean of the pleural fluid pressures during inspiration and expiration, was measured immediately and after each 100 mL of effusion was withdrawn. Thoracentesis was terminated when the pressure was lower than –10 cm H₂O, when the patient developed symptoms potentially related to reexpansion pulmonary edema, or when a total of 500 mL of effusion had been drained. The elastance of the pleural space was calculated as the change in pressure of the pleural effusion (in cm H₂O) divided by the amount of fluid removed [11]; the standard measure was expressed as being equivalent to the decline in pressure of the pleural effusion after 500 mL of fluid had been removed. In five patients who had a decrease in pressure of at least 19 cm H₂O and trapped lung, thoracentesis was stopped before 500 mL of effusion had been drained; the elastance value in these patients was adjusted by linear extrapolation on the basis of the change in pressure that would have occurred if 500 mL had actually been removed [11].

View larger version (43K): [in this window] [in a new window]   Figure 1. Device used to measure the elastance of the pleural space.

After these measurements had been done, a 9.3-mm thoracostomy tube connected to a watersealed bottle was placed in the pleural space for further drainage of the effusion by gravity. Suction was applied to the chest tube only if the patient had radiographic evidence of trapped lung. Radiography was done to ensure proper positioning of the thoracostomy tube. Drainage was discontinued when less than 150 mL of effusion was drained per day [3, 7, 12] for 2 consecutive days and the lung had reexpanded, when less than 250 mL of effusion was drained per day for 4 consecutive days and the lung had reexpanded, or when drainage had continued for 10 days. When one of the three criteria was met, radiography was done to evaluate the reexpansion of the affected lung and the approximation of the pleurae. Figure 2 shows the degrees of reexpansion of the affected lung and the approximation of the pleurae. In category 1, the lung had completely reexpanded and the pleurae had approximated well. In category 2, the affected lung had reexpanded and the pleurae had approximated in most areas, but a small amount of residual pneumothorax, pleural effusion, or some heterogeneous white patches were evident on radiographs. In category 3, the lung did not reexpand and the visceral and parietal pleurae were separated by pneumothorax in most places. The radiographs were evaluated by at least two of the authors. If reexpansion of the lung and approximation of the pleurae were difficult to interpret, the patients were placed in category 2. Patients in category 1 or 2 were classified as having a "nontrapped lung."

View larger version (121K): [in this window] [in a new window]   Figure 2. Categories of reexpansion of the affected lung and approximation of the pleurae. Arrowheads mark the trapped lung.

The chest tube was then clamped, and we injected 60 mg of bleomycin diluted in 100 mL of normal saline (usual-dose group) [4, 5, 7, 8, 13-15] or, in the latter half of the study, 30 mg diluted in 50 mL of normal saline (low-dose group) [13, 14] into the pleural space for pleurodesis in all patients except two who had trapped lung. The body position of the patient was changed if the patient was not too ill [16-18]. The chest tube was reopened 2 hours later and removed the next day [5]. Chest radiography was done to ensure that the chest tube had been removed properly.

Patients were then followed closely in the outpatient clinic, and chest radiography was done to evaluate the outcome of pleurodesis. If the results were equivocal, we did either chest ultrasonography with or without diagnostic thoracentesis or computed tomography of the chest. The results of pleurodesis were evaluated 30 days after the chest tube was removed and then until the patient died or was lost to follow-up [3, 4, 12-15, 19]. Successful pleurodesis was defined as no recurrence of effusion, recurrence of only a small amount of effusion, or loculated effusion and elimination of the requirement for further therapeutic thoracentesis to alleviate symptoms [10, 12-15, 19]. Unsuccessful pleurodesis was defined as recurrence of the effusion in an amount similar to that seen before treatment or the requirement for further therapeutic thoracentesis to alleviate symptoms.

Statistical Analysis

The two-sample t-test was used to compare the mean elastance, pH, glucose level, and duration of follow-up for patients with trapped and nontrapped lungs. The Fisher exact test was used to evaluate the association between categorized variables. Confidence intervals for means and proportions were determined when appropriate.

Results

Top Methods Results Discussion Author & Article Info References

Sixty-five patients (38 men and 27 women) between 39 and 83 years of age (mean age, 62.7 years) were included in the study. Two patients had gastric cancer, 2 had breast cancer, 1 had renal cancer, 1 had rectal cancer, 45 had bronchogenic adenocarcinoma, and 14 had adenocarcinoma of uncertain origin. All patients had pleural effusion on chest radiography that reached the hilar level or higher in the sitting or standing position. Fourteen patients had trapped lung, and 51 did not have trapped lung.

As Table 1 and Figure 3 show, 11 of 14 patients with trapped lung (groups 1 and 4) had an elastance of 19 cm H₂O or more, although only 3 of the 51 patients without trapped lung (groups 2, 3, and 5) had such an elastance (P < 0.001). Mean elastance was 30.59 cm H₂O in the 14 patients with trapped lung and 8.31 cm H₂O in the 51 patients without trapped lung (difference, 22.3 cm H₂O [95% CI, 11.1 to 33.5]; P = 0.001).

View larger version (13K): [in this window] [in a new window]   Figure 3. Relation of pleural elastance, effusion pH, effusion glucose level, trapped lung, and outcome of pleurodesis. x = patients who were lost to follow-up; open circle = patients with successful pleurodesis; \#9679; = patients with unsuccessful pleurodesis; group 1 = patients with trapped lung lost to follow-up at 1 month; group 2 = patients without trapped lung lost to follow-up at 1 month; group 3 = patients without trapped lung with successful pleurodesis at 1 month; group 4 = patients with trapped lung with unsuccessful pleurodesis at 1 month; group 5 = patients without trapped lung with unsuccessful pleurodesis at 1 month.

Five patients who had trapped lung were lost to follow-up at 1 month (Figure 3; group 1); pleurodesis was noted to have been unsuccessful in two of these five patients before they were lost to follow-up. Another five patients with elastance less than 19 cm H₂O who did not have trapped lung (Figure 3; group 2) were also not evaluated at 1 month because of loss to follow-up (n = 3), death (n = 1), or postoperative cardiac tamponade (n = 1). Tight pleural approximation was noted during surgery in the patient with postoperative cardiac tamponade.

Fifty-five patients were evaluated 1 month after the thoracostomy tube was removed (Figure 3; groups 3, 4, and 5). Pleurodesis was unsuccessful in all 9 patients who had trapped lung (Figure 3; group 4) and in the 3 patients who had an elastance of 19 cm H₂O or more without trapped lung (Figure 3; group 5). In contrast, 42 of the 43 patients (98%) with an elastance less than 19 cm H₂O without trapped lung had successful pleurodesis (Figure 3; groups 3 and 5).

The 14 patients with trapped lung had a higher elastance (30.59 cm H₂O) than did the 51 patients without trapped lung (8.31 cm H₂O) (P = 0.001); the effusion from the former group also had a lower pH (7.133 compared with 7.308; P = 0.001) and a lower glucose level (3.16 mmol/L compared with 5.27 mmol/L; P = 0.011) than the effusion from the latter group. Duration of follow-up in the group with trapped lung (2.7 months) was similar to that in the group without trapped lung (4.1 months) (P = 0.161). In Table 2 and Figure 3, patients are stratified according to risk factors (glucose level < 3.33 mmol/L, pH < 7.2, or elastance 19 cm H₂O). Of these three possible predictors, elastance was most frequently associated with trapped lung (P < 0.001). The data in Table 2 show that sensitivity, specificity, and positive and negative predictive values for trapped lung were consistently higher for elastance than for pH, but the differences were not statistically significant. Discordance between effusion pH and elastance was noted in 14 patients. Of these 14 patients, 5 had an effusion pH of 7.2 or more and an elastance of 19 cm H₂O or more. Four of the 5 had trapped lung, and 1 did not have trapped lung. In contrast, of the 9 patients with an effusion pH less than 7.2 and an elastance less than 19 cm H₂O, 2 had trapped lung and 7 did not have trapped lung.

The relation between the three risk factors and the outcome of pleurodesis at 1 month in the 46 patients without trapped lung is shown in Table 3. Once again, the sensitivity, specificity, positive predictive value, and negative predictive value were higher for elastance than for pH and glucose level, but none of the differences was statistically significant. In five patients, the risk factors of effusion pH and elastance were discordant as predictors of successful pleurodesis. One of the five patients with effusion pH of 7.2 or more and an elastance of 19 cm H₂O or more had unsuccessful pleurodesis; pleurodesis was successful in three patients and unsuccessful in one patient with effusion pH less than 7.2 and elastance less than 19 cm H₂O.

When all 65 patients were included in the analysis, significant correlation was found between elastance and effusion pH (r = –0.499;P < 0.001), between effusion glucose level and pH (r = 0.385; P < 0.001), and between effusion glucose level and elastance (r = –0.314;P = 0.0107).

The usual-dose bleomycin group contained 31 patients, and the low-dose group contained 32 patients. Of the 3 patients without trapped lung whose elastance was 19 cm H₂O or more and who had unsuccessful pleurodesis (Figure 3; group 5), 1 was in the low-dose group and 2 were in the usual-dose group. Of the 43 patients whose elastance was less than 19 cm H₂O and who did not have trapped lung (Figure 3; groups 3 and 5), 22 were in the low-dose group and 21 were in the usual-dose group. The patient who had unsuccessful pleurodesis was in the usual-dose group (Figure 3; group 5).

Discussion

Top Methods Results Discussion Author & Article Info References

This study shows that patients with an elastance of 19 cm H₂O or more had a higher incidence of trapped lung than did those with a lower elastance. Elastance was the best predictor for trapped lung and outcome of pleurodesis, but a correlation was also found with pH and glucose levels of the effusion. The outcome of pleurodesis depends on 1) the approximation of visceral and parietal pleurae after drainage of the effusion and 2) an inflammatory reaction and, ultimately, fibrosis of the pleurae induced by intrapleural injection of sclerosing agents [2-4, 7-9]. Although trapped lung does not allow successful pleurodesis, trapped lung is difficult to detect before thoracoscopy, thoracostomy, and drainage of the effusion [5, 6]. Rodriguez-Panadero and colleagues [20] showed that a pleural effusion pH of less than 7.2 had an 86% positive predictive value for trapped lung. However, these investigators defined trapped lung as the gross thoracoscopic finding of a tumor burden score of 3 (that is, massive tumoral lesions) on the visceral pleura. Although such a lesion on the visceral pleura is frequently associated with trapped lung, it does not by itself indicate trapped lung. In our study, as in many others, trapped lung was defined as poor approximation of the pleurae after drainage of effusion. Good approximation of the pleurae is also determined by factors other than local tumor burden, including lesions in the lung or airway, the ability of the mediastinum to shift to the side of the lesion, the ability of the diaphragm to move upward, and the inward movement of the rib cage. In fact, in a subsequent paper [6], the same group of investigators showed that pleurodesis was successful in 11 of 22 patients who had massive lesions on the visceral pleura. Because patients with trapped lung defined in the more conventional way have infrequently had successful pleurodesis [2-57, 9], it is possible that these 11 patients did not have trapped lung despite the presence of massive lesions on the affected visceral pleura. Pleurodesis was unsuccessful in all patients with trapped lung in our study.

Previous studies have also shown that pleurodesis is less successful in patients whose effusions have a glucose level less than 3.33 mmol/L or a pH less than 7.3 (or 7.2), even in the absence of trapped lung [6, 19, 21]. However, 57% of such patients still achieve good results after pleurodesis [6, 19]. In our study, the 98% success rate for pleurodesis in patients without trapped lung who have an elastance less than 19 cm H₂O further demonstrates that when approximation of the pleurae is excellent or fair, pleurodesis is likely to be successful. However, it is sometimes difficult to determine the presence of trapped lung radiographically; that is, it is difficult to accurately place a patient in category 2 or 3 using conventional radiography. For example, all three patients in our study who had an elastance of 19 cm H₂O or more who did not have trapped lung confirmed on radiographs and had unsuccessful pleurodesis (Figure 3; group 5) were placed in category 2, although computed tomography in two patients revealed poor approximation of the pleurae. These two patients, in fact, should have been placed in category 3.

Approximation of the pleurae is a mechanical rather than a chemical process. Therefore, it is not surprising that elastance is a better predictor of successful pleurodesis than are biochemical measurements. The significant correlation between elastance, pH, and glucose level is not surprising because these variables share a common mechanism. A thick layer of tumor or fibrin on the pleurae impairs the movement of glucose to the pleural fluid, the movement of acid out of the pleural fluid [19, 21, 22], and the reexpansion of the lung. The statistically insignificant difference between elastance and effusion pH in predicting a trapped lung and the outcome of pleurodesis in our study is probably the result of the small sample size.

The chest tube was removed 1 day after the intrapleural injection of bleomycin in all of our patients, regardless of the amount of effusion drainage on that day. The success rate of pleurodesis, however, was very high (98%) in patients without trapped lung who had an elastance less than 19 cm H₂O. These results agree with those of Villanueva and colleagues [12]. Thus, it is not necessary to wait until the drainage fluid declines to less than 150 mL/d before removing the chest tube [2, 3, 5, 7, 12]. The reported success rate of pleurodesis with intrapleural injection of bleomycin as the sclerosing agent in patients without trapped lung has ranged from 54% to 84% [4, 5, 8, 13-15], which contrasts with the 91% success rate (42 of 46 patients) that we found. We used bleomycin hydrochloride instead of bleomycin sulfate, which is used in the United States [7]. It is not certain whether this difference in pharmacologic formulation could result in a different outcome of pleurodesis, but it is unlikely.

The costs of the agents used for chemical pleurodesis in the United States vary from $0.50 for 10 g of talc to $1104 for 70 mg of bleomycin. The costs of doxycycline range from $86 to $403, depending on the number of instillations required [8]. The efficacy of bleomycin and agents of the tetracycline class, such as minocycline and doxycycline, are similar [8]. Bleomycin causes no pleuritic pain (a major problem when doxycycline is used) and thus needs no premedication with intrapleural injection of lidocaine (as is necessary with doxycycline). The major disadvantage of bleomycin is its cost. In our study, a low dose of bleomycin (30 mg) was as effective (100%) as the usual dose (60 mg; 95% effective) in patients who had an elastance less than 19 cm H₂O and nontrapped lung. Thus, cost-effectiveness is improved by using low doses of bleomycin for pleurodesis. The patients who have an elastance 19 cm H₂O should probably be managed with other methods, such as thoracoscopy with talc insufflation or other surgical interventions that might be more effective than chemical pleurodesis [8, 9, 20]. Talc insufflation incurs costs for sterilization of the talc, thoracoscopy, general anesthesia, and the operating room, which substantially increase the overall cost.

Our study has some limitations. The reading of chest radiographs was not blinded, the diagnosis of trapped lung done using conventional chest radiography was not assessed for reliability, and the predictive power of measurements of elastance was only marginally greater than that of effusion pH and glucose level. Further studies with improved design are needed.

In summary, our study shows that measurement of the elastance of the pleural space in patients with symptomatic malignant pleural effusion is simple and effective in the diagnosis of trapped lung. A more important finding is that measurement of the elastance enhances prediction of the outcome of chemical pleurodesis. Patients who have an elastance less than 19 cm H₂O can be treated by tube thoracostomy followed by intrapleural injection of bleomycin in low doses (30 mg) for pleurodesis. Patients with an elastance of 19 cm H₂O or more have a high incidence of trapped lung and are less likely to have successful pleurodesis if bleomycin is used as the sclerosant. These patients should be managed with other, more effective (although more invasive) methods.