Therapeutic thoracentesis is one of the most commonly performed procedures, with an estimated 173,000 thoracentesis performed yearly in the United States.1 In contrast to the relatively recent introduction of ultrasonography-guided thoracentesis now considered standard of care by most and supported by a large body of evidence,2–4 manometry was described >120 years ago2 but is not commonly performed by chest physicians. The measurement of pleural pressures during thoracentesis can assist in the diagnosis of unexpandable lung,2,5–8 and has been proposed as a potential safeguard against complications related to excessive transpulmonary pressures such as pneumothorax ex vacuo and reexpansion pulmonary edema (REPE). Pleural manometry may also help predict the success of pleurodesis.9
However, existing data have not convincingly that manometry prevents pneumothorax ex vacuo or REPE.10,11 The rarity of these events makes comparative studies using these as endpoints difficult to power adequately. In addition, demonstrating that the lung has completely reexpanded after maximal fluid removal (guided by patient symptoms and without manometry) has been suggested as an adequate way to diagnose unexpandable lung.12 Large-volume thoracenteses guided by patient symptoms without manometry have been reported as safe in the literature.13 Conversely, the correlation between pleural pressures and the development of chest discomfort during thoracentesis has been firmly established.14,15 As such, we hypothesized that using pleural manometry during thoracentesis may help anticipate the development of excessively negative pleural pressures and potentially prevent the development of chest discomfort, resulting in a more comfortable procedure for patients. The potential of manometry for preventing chest discomfort during therapeutic thoracentesis has, to the best of our knowledge, not been specifically addressed yet. We propose in this study to evaluate whether manometry allows for improved patient comfort and relief of symptoms after the procedure.
We performed a retrospective cohort study based on chart review of 214 consecutive adult patients referred for therapeutic thoracentesis in the Division of Pulmonary and Critical Care Medicine of Mayo Clinic, Rochester, MN, between January 1, 2011 and June 30, 2013. This study was approved by the Mayo Clinic Institutional Review Board (IRB 13-006216). All patients included in the study provided a general informed consent to the use of their clinical data for research purposes during the first clinical encounter at Mayo Clinic.
Procedures performed by 3 proceduralists (F.M., J.J.M., and C.E.D.) experienced in performance and interpretation of manometry were included in this study. Therapeutic thoracenteses were defined as thoracenteses during which there was an attempt at maximum fluid drainage. Thoracenteses performed by other proceduralists who do not use manometry were not included. After a standard procedural pause (during which patient identity, type, and laterality of the procedure are confirmed in the presence of the patients and medical staff involved), the ultrasound machine (Micromaxx Ultrasound System with P17/5-1 MHz 17 mm phased array probe; Sonosite Inc., Brothell, WA) was used to identify the most appropriate site for needle insertion with the patient sitting in the upright position. The site was marked and the skin was then cleaned using 2% chlorhexidine gluconate and 70% isopropyl alcohol formulation and draped in standard manner. Local anesthesia was then performed using 1% lidocaine with a maximum of 20 mL injected. The volume of lidocaine used was left at the discretion of the operator. Pleural pressures were measured using a standard pressure transducer (TruWave Disposable Pressure Transducer; Edwards LifeSciences Corp., Irvine, CA) and a portable patient arterial blood pressure monitoring system (Philips IntelliVue MP50 Patient Monitor; Philips Corporation, Amsterdam, Netherlands). Pressure parameters are optimized to measure most accurately between 20 and −20 mm Hg and all data are converted to cm H2O for analysis. This measurement technique is validated internally against a water column manometer and a clinical transducer calibration system (XCaliber Transducer Calibration System; Gould Inc., Oxnard, CA). The electronic manometry system was then set up and zeroed using the marked needle entry site. Typically, setting up the manometry montage takes approximately 5 minutes before the procedure, and an additional 5 to 10 minutes of procedure time depending on the amount of pleural fluid removed. As no evidence supports the systematic use of manometry in therapeutic thoracentesis, decision to proceed was left at the discretion of the proceduralist and this was explained to all patients. Although there are no standard criteria guiding the decision to proceed with manometry, we typically perform manometry in the following situations:
- Evidence of large pleural effusions by ultrasound with the intent to drain as much fluid as possible, in order to avoid large swings in intrapleural pressure.
- When the etiology of a recurrent pleural effusion remains unclear after several thoracenteses.
- When nonexpandable lung is suspected, because of the occurrence of pain, pneumothorax ex vacuo, or evidence of incomplete drainage after prior thoracentesis.
All thoracenteses were performed using a customized thoracentesis tray assembled for our institution by Cardinal Health, Waukegan, IL and a 5.0 Fr Yueh Centesis Needle (Cook Medical Inc., Bloomington, IN). The pleural fluid was drained by connection to an evacuation bottle. When manometry was used, an initial (opening) pressure was measured followed by drainage that was interrupted to measure a mean pleural pressure every 250 mL in the setting of a positive initial pressure, and every 100 mL with a negative initial pressure. The drainage was stopped when the pleural pressure became more negative than −20 cm H2O, as recommended by expert opinion, or until the patient developed symptoms (chest discomfort and intractable cough) or until complete fluid drainage.14 Postprocedural ultrasound was performed and a chest x-ray was only obtained when the ultrasound examination suggested a pneumothorax or in case of increased symptoms (chest discomfort and dyspnea). We do not use an absolute volume threshold for discontinuing therapeutic thoracentesis at our institution, as it does not appear supported by current evidence.13,16 Without manometry, only patient symptoms were used to guide the drainage, which was otherwise continued until a maximum pleural volume was removed.
We attempted to study the effect of manometry on patient important outcomes by comparing the change in chest discomfort and dyspnea experienced by patients who underwent manometry versus those who did not. As part of our usual clinical process, preprocedure and postprocedure chest discomfort and dyspnea scores using a linear analog scale from 0 (no symptoms) to 10 (maximum symptoms) are assessed by an assistant immediately before and 2 minutes after the procedure and recorded in the electronic medical records at Mayo Clinic. These data were subsequently abstracted by 2 independent investigators (J.P. and Z.S.D). We calculated that in order to demonstrate a minimally clinical important difference of 2 on the linear analog scale with a 90% power, 40 patients would be required in each group.
A total of 267 patients underwent outpatient thoracentesis during the time period specified above, under 3 separate proceduralists (J.J.M, C.E.D, and F.M). Twenty-eight patients with incomplete data and 9 patients without documented informed consent for research were excluded from our study (during standard registration to our health facility, patients are offered a chance to consent to the use of their clinical information for research purposes). Patients were eligible for inclusion only once, and repeated procedures were excluded. Sixteen patients who had a therapeutic pneumothorax performed at the end of the procedure for chest discomfort with evidence of negative pressure by manometry were excluded. We justified the decision to exclude these patients by the fact that manometry clearly did not anticipate the development of chest discomfort in these patients, and post-thoracentesis pain scores were confounded by the therapeutic pneumothorax performed (Fig. 1).
Continuous data are described using mean and standard deviation and categorical data are described using counts and percentages. Association of covariates with dyspnea and chest discomfort was assessed with linear regression. Covariates were selected for inclusion into multivariable regression models following the Hosmer Lemeshow approach.17 For all statistical tests, a P-value of ≤0.05 was considered statistically significant. SAS 9.3 was used for all analyses (SAS Institute, Cary, NC).
Manometry was performed in 82 of 214 thoracenteses (38%) included in our study. The mean patient age was 67 years (±14.8 y). No significant difference was found between the manometry and nonmanometry groups in terms of age, sex, pleural fluid appearance, laterality of thoracentesis, and baseline chest discomfort and dyspnea scores. Significant differences existed in the amount of local anesthetic used, volume of fluid removed, and operator performing the procedure (Table 1).
The spectrum of diagnosis of our patient population has been described in Table 2. Although statistically significant differences were noted in distribution of diagnosis between the 2 groups, still in a majority of patients in both groups the etiology of the effusions remained unknown. On univariate analysis, manometry was not associated with change in chest discomfort or dyspnea. Additional covariates were tested for their association with change in discomfort and dyspnea. Baseline discomfort (P<0.01) and volume removed (P=0.03) were associated with change in discomfort. Baseline dyspnea was significantly associated with change in dyspnea before and after the procedure (P<0.01) (Table 3).
A multivariate regression model was used to adjust for the possible confounding effects of: amount of local anesthetic used, volume of fluid removed, and operators performing the procedure as well as baseline risk factors. Covariates with a univariate association (P-value <0.1) with the outcome were considered baseline risk factors for fitting the multivariable model. No interactions were detected between covariates. The final multivariable model as adjusted for baseline chest discomfort and lidocaine volume and showed that manometry was not significantly associated with change in discomfort (P=0.12). No other covariates contributed to the model fit. Similarly, when adjusted for baseline dyspnea, manometry was not significantly associated with change in dyspnea (P=0.24).
Subgroup analyses were also performed on the group of patients who had large-volume thoracenteses as defined by volume drained >1000 mL. In total, 76 patients underwent large-volume thoracentesis and 35 (46.1%) of these underwent manometry. No significant differences were found in the change in symptoms experienced by the group of patients undergoing manometry as compared with those who did not (Pdiscomfort=0.32, Pdyspnea=1.0) on univariate as well as multivariate analysis (Table 4). Another subgroup analysis including the patients who received therapeutic pneumothorax was also performed. Again no significant difference could be demonstrated among the above 2 groups(Pdiscomfort=0.1, Pdyspnea=0.2).
Our study suggests that manometry, as currently performed, may not prevent the development of chest discomfort during therapeutic thoracentesis. As manometry has not been definitively shown in prior studies to prevent other complications related to increased transpulmonary pressures, such as REPE and pneumothorax ex vacuo, our data suggest that further prospective studies may be warranted before recommending that manometry be performed systematically in all therapeutic thoracenteses.
The clinical role of manometry during thoracentesis has been a topic of debate in recent years.12,18,19 Occurrence of complications such as REPE and pneumothorax ex vacuo are difficult to predict and have led experts to recommend limiting pleural fluid drainage to 1000 to 1500 mL, unless manometry is used in which case pleural fluid drainage could be continued in the absence of patient symptoms or until the pressure drops below −20 cm H2O.14,15,20–22 It is important to emphasize that neither the absolute volume nor the pressure thresholds are supported by strong evidence.2,14,23 However, there is clear evidence that excessively negative pleural pressures, that is, increased transpulmonary pressures, correlate well with the diagnosis of unexpandable lung and the development of symptoms such as chest discomfort.2,5–8,14,15
At resting state (ie, functional residual capacity), the pleural pressure is slightly subatmospheric, approximately −3 to –5 cm H2O.9,24 Depending on the pathophysiology of the pleural effusion, accumulation of pleural fluid generally results in an increase in pleural pressure. Drainage of the effusion will therefore typically lead to a gradual return of pleural pressures to a more physiological level, as long as the volume removed can be replaced by expansion of the lung and/or return of the diaphragm muscle and chest wall to their normal resting states. A sharp drop in pleural pressures during pleural drainage, whether due to trapped lung, endobronchial lesion limiting lung reexpansion, or decreased compliance of the lung (eg, fibrotic lung diseases), reflects a marked increase in transpulmonary pressures, which are thought to be associated with complications. Although intuitively appealing, monitoring pleural pressures during thoracentesis was not found to absolutely prevent REPE or pneumothorax ex vacuo.10,11
The largest prospective cohort study performed to date on manometry-guided therapeutic thoracentesis by Feller-Kopman et al10 included 185 patients and confirmed that that the incidence of REPE is very rare, with only 1 clinically relevant and 4 radiologic cases reported in the study. Interestingly, REPE was independent of pleural pressures, pleural elastance, and volume of fluid removed. Likewise, Heidecker et al11 concluded from their study that occurrence of unintentional pneumothorax in thoracentesis cannot be prevented by monitoring pleural pressures. This may be explained by the variable distribution of stress for a given transpulmonary pressure leading to unpredictable regional complications in less compliant regions of the lung. Another perhaps more likely explanation is based on the fact that all current manometry systems require that pleural flow be interrupted during measurements.5 As such, pleural pressure data points are obtained intermittently during the procedure, leaving substantial periods of “blind time” during which acute change in pleural elastance cannot be observed. These proposed mechanisms might explain the negative findings in our study as well. Continuous pleural pressure monitoring with real-time pleural elastance curves may circumvent these limitations. In addition, it is fair to point out that as drainage was stopped in these studies after reaching a predefined negative pressure threshold, additional complications could have arisen in the absence of manometry, and that comparative studies are clearly needed.
Feller-Kopman et al14 analyzed the relationship between development of chest tightness and pleural pressure changes. Their study demonstrated a good correlation between chest tightness and absolute change in pleural pressures and lower closing pressures. However, only 22% of patients with chest discomfort and 9% of asymptomatic patients in this study had pleural pressure below −20 cm H2O. These results suggest that pleural manometry technique, as currently performed, with intermittent pressure measurements and an arbitrarily chosen threshold of −20 cm H2O, may not be optimal.
The aim of our study was to clarify the potential role of manometry, as generally performed, in preventing chest discomfort during thoracentesis. As opposed to REPE and pneumothorax ex vacuo, which are relatively rare complications and do not lend themselves well to adequately powered studies, our study was sufficiently powered to identify any difference in this patient-important endpoint between both groups. We found that there was no significant difference in the change in chest discomfort and dyspnea experienced by patients who underwent manometry as compared with those who did not. The results were consistent on multivariate analysis after adjusting for potential confounding factors. Large-volume thoracenteses have been linked to a perceived increased risk of complications in the literature.2,10,11,14,15 Hence we performed a subgroup analysis including those patients who had >1000 mL of pleural fluid removed. Again, no significant difference was observed between the 2 groups on both univariate and multivariate analyses.
Our study has limitations. The decision to proceed with manometry was left at the discretion of the proceduralists. This likely introduced selection bias, as a high pretest probability of unexpandable lung could have motivated the decision to proceed with manometry. It is therefore possible that patients in the manometry group would have experienced more complications in the absence of manometry, that is, the absence of difference in change in dyspnea scores could, in fact, support the use of manometry. Although plausible, this explanation remains hypothetical and we acknowledge that appropriately designed prospective comparative studies are needed. In addition, our study was retrospective and therefore subject to the usual limitations inherent in this study design. Likewise, even if our study is one of the largest studies on the use of manometry, its impact on the development of REPE and pneumothorax could not be analyzed, given the rarity of these complications. However, it may be difficult to adequately power studies on REPE or pneumothorax ex vacuo. It is very plausible that the manometry technique generally used may not be adequate to anticipate the development of excessively negative pressures associated with chest discomfort and other complications during thoracentesis. Continuous manometry using a double-lumen catheter may for instance allow identification of an inflection point on the pleural elastance curve, which may address the current limitations of pleural manometry. In this study, we followed a protocol similar to those described in other manometry studies, acknowledging the limitations of such protocol and the need for additional research. In addition, descriptions and grading of symptoms such as chest discomfort or dyspnea by patients are by nature subjective with risks of both underestimation and overestimation. Finally, we recognize that 16 patients with chest discomfort that resolved with therapeutic pneumothorax were excluded in our study. However, manometry evidently did not anticipate the development of chest discomfort in these patients, which was the object of our study. We acknowledge that manometry may be needed to understand the pathophysiology of the chest discomfort in these cases. We also recognize that excluding these patients may bias the results in favor of manometry. Hence an analysis including these patients was also performed, which did not show any significant difference of change in symptoms with or without performing manometry. We did not address the clinical practice of therapeutic pneumothorax in our study, but do believe that future prospective comparative studies are needed as well.
In summary, our study suggests that using manometry, as currently performed by pulmonologists during thoracentesis, may not anticipate the development of chest discomfort related to excessively negative pressures. These results may be due to technical limitations, or the retrospective nature of our study design introducing selection bias, and future prospective studies using real-time pleural elastance curves and patient-centered outcomes are needed to clarify the utility of pleural manometry. Although we believe that the adoption of a clinical intervention as standard of care should generally be based on the evidence for improved outcomes, not on the absence of evidence to the contrary, we also recognize that manometry is a simple technique that offers benefits. Additional data are needed to clarify its utility and determine whether it should be systematically performed or reserved for selected patients.
1. Owings MF, Kozak LJ. Ambulatory and inpatient procedures in the United States, 1996. Vital and Health Statistics. Series 13, Data From the National Health Survey. 1998;139:1–119.
2. Feller-Kopman D. Therapeutic thoracentesis: the role of ultrasound and pleural manometry. Curr Opin Pulm Med. 2007;13:312–318.
3. Sachdeva A, Shepherd RW, Lee HJ. Thoracentesis and thoracic ultrasound: state of the art in 2013. Clin Chest Med. 2013;34:1–9.
4. Mercaldi CJ, Lanes SF. Ultrasound guidance decreases complications and improves the cost of care among patients undergoing thoracentesis and paracentesisultrasound guidance in thoracentesis/paracentesis. Chest. 2013;143:532–538.
5. Doelken P, Huggins JT, Pastis NJ, et al.. Pleural manometrytechnique and clinical implications. Chest. 2004;126:1764–1769.
6. Villena V, Lopez-Encuentra A, Pozo F, et al.. Measurement of pleural pressure during therapeutic thoracentesis. Am J Respir Crit Care Med. 2000;162:1534–1538.
7. Huggins JT, Doelken P, Sahn SA. The unexpandable lung. F1000 Med Rep. 2010;2:77.
8. Huggins JT, Sahn SA, Heidecker J, et al.. Characteristics of trapped lung pleural fluid analysis, manometry, and air-contrast chest CT. Chest. 2007;131:206–213.
9. Feller-Kopman D, Parker MJ, Schwartzstein RM. Assessment of pleural pressure in the evaluation of pleural effusions. Chest. 2009;135:201–209.
10. Feller-Kopman D, Berkowitz D, Boiselle P, et al.. Large-volume thoracentesis and the risk of reexpansion pulmonary edema. Ann Thorac Surg. 2007;84:1656–1661.
11. Heidecker J, Huggins JT, Sahn SA, et al.. Pathophysiology of pneumothorax following ultrasound-guided thoracentesis. Chest. 2006;130:1173–1184.
12. Maldonado F, Mullon JJ. Counterpoint: should pleural manometry be performed routinely during thoracentesis? No. Chest. 2012;141:846–848discussion 848-849.
13. Jones PW, Moyers JP, Rogers JT, et al.. Ultrasound-guided thoracentesis: is it a safer method? Chest. 2003;123:418–423.
14. Feller-Kopman D, Walkey A, Berkowitz D, et al.. The relationship of pleural pressure to symptom development during therapeutic thoracentesis. Chest. 2006;129:1556–1560.
15. Light R, Jenkinson S, Minh V, et al.. Observations on pleural fluid pressures as fluid is withdrawn during thoracentesis. Am Rev Respir Dis. 1980;121:799–804.
16. Abunasser J, Brown R. Safety of large-volume thoracentesis. Conn Med. 2010;74:23–26.
17. Hosmer DW Jr, Lemeshow S, Sturdivant RX. Applied logistic regression: Wiley.com, 2013.
18. Feller-Kopman D. Point: should pleural manometry be performed routinely during thoracentesis? Yes. Chest. 2012;141:844–845.
19. Khosla R, Kistler CR. Pleural manometry: our point of view. Chest. 2012;142:1357–1358.
20. Heller BJ, Grathwohl K. Contralateral reexpansion pulmonary edema. South Med J. 2000;93:828–831.
21. Trapnell D, Thurston JB. Unilateral pulmonary oedema after pleural aspiration. Lancet. 1970;295:1367–1369.
22. Matsuura Y, Nomimura T, Murakami H, et al.. Clinical analysis of reexpansion pulmonary edema. Chest. 1991;100:1562–1566.
23. Genofre EH, Vargas FS, Teixeira LR, et al.. Reexpansion pulmonary edema. Jornal de Pneumologia. 2003;29:101–106.
24. Broaddus V, Light R. Disorders of the Pleura: General Principles and Diagnostic Approach. Textbook of Respiratory Medicine. 1994:2nd ed.Philadelphia, PA:WB Saunders;2156–2160.