This prospective observational study reports on diaphragm excursion, velocity of diaphragm contraction, and changes in pleural pressure that occur with thoracentesis.
Twenty-eight patients with pleural effusion underwent therapeutic thoracentesis. Diaphragm excursion and velocity of diaphragm contraction were measured with M-mode ultrasonography of the affected hemidiaphragm. Pleural pressure was measured at each aliquot of 250 mL of fluid removal. Fluid removal was continued until no more fluid could be withdrawn, unless there was evidence of nonexpandable lung defined as a pleural elastance greater > 14.5 cm H2O/L and/or ipsilateral anterior chest discomfort.
Twenty-three patients had expandable lung, and five patients had nonexpandable lung. Velocity of diaphragm contraction (mean ± SD) increased from 1.5 ± 0.4 cm/s to 2.8 ± 0.4 cm/s pre-thoracentesis and post-thoracentesis, respectively (CI, 0.93-1.61; P < .001) in subjects with expandable lung. Velocity of diaphragm contraction (mean ± SD) increased from 2.0 ± 0.4 cm/s to 2.3 ± 0.4 cm/s pre-thoracentesis and post-thoracentesis (P = .45) in subjects with nonexpandable lung. Diaphragm excursion was significantly increased in subjects with expandable lung at the end of thoracentesis; diaphragm excursion did not increase to a significant extent in patients with nonexpandable lung.
The velocity of diaphragm contraction and diaphragm excursion increased in association with fluid removal with thoracentesis in patients with expandable lung, whereas it did not significantly change in patients with nonexpandable lung. This may derive from improvement in loading conditions of the diaphragm in patients with expandable lung related to its preload and length-tension characteristics.
Pleural effusion has adverse effects on respiratory function through a variety of mechanisms that include alteration in the elastic equilibrium volumes of lung and chest wall, increased respiratory resistance, increased work of breathing, and hypoxemia due to intrapulmonary shunt.
Pleural fluid accumulation causes cranial displacement of the ribs, which results in impairment of intercostal muscle function. The removal of pleural fluid increases end-expiratory lung volume, decreases shunt fraction, improves respiratory mechanics, and reestablishes the force-length relationship of the diaphragm with improvement in its contractile function. In an animal model, stepwise introduction of fluid into the chest cavity caused caudal displacement of the diaphragm with significant impairment in the capacity of transdiaphragm pressure generation.
The objective of the current study was to measure diaphragm excursion and velocity of contraction using ultrasonography to correlate these parameters with pleural pressures measured during stepwise drainage of pleural fluid.
Patients and Methods
This prospective observational study was performed between February 2016 and April 2017 at Long Island Jewish Medical Center and North Shore University Hospital. The study population was a convenience sample, as inclusion into the study was determined by the availability of the primary investigator. The institutional review board of the Northwell Health System approved this study (IRB No. HS15-0682). Study subjects or their appropriate surrogate gave written informed consent to participate in the study.
The primary medical team made all decisions related to the indication and the timing of the thoracentesis. The investigator was informed by the primary medical team that a high-volume thoracentesis was planned as part of the ongoing medical care of the patient. If the investigator was available, the patient or the appropriate surrogate was asked to give informed consent to make the experimental measurements during the performance of the thoracentesis. If the patient or the appropriate surrogate gave informed consent to participate, the patient was enrolled in the study.
Exclusion criteria were lack of informed consent, age < 18 years, pregnancy, diaphragm paralysis, or recent thoracic surgery.
Thoracentesis and Pleural Manometry
All thoracenteses were performed by two investigators (Y. A. G. and A. P.) who were fully qualified in the procedure. Patients were placed in a sitting position with their arms resting on a flat bedside table with full exposure of the posterolateral hemithorax. The best site, angle, and depth for needle insertion were determined via ultrasonography by targeting the most dependent safe site for device insertion into the pleural fluid. Using a standard thoracentesis technique, an 8F pleural drainage catheter (Pleural-Seal thoracentesis kit, AK-01000; Arrow-Clark) was inserted with a sterile technique following local lidocaine infiltration. There was no departure from standard operating procedure for insertion of the pleural drainage catheter.
One side arm of the two-way stopcock attached to the pleural drainage catheter was used to measure serial pleural pressures while the other side arm was used to remove the pleural fluid. The pleural pressure was measured by using a fluid-filled low-compliance tubing set attached to a pressure transducer (TruWave disposable pressure transducer; Edwards Lifesciences). The tubing and pressure transducer were purged of air bubbles with normal saline. The transducer was zeroed to a reference point defined as the level at which the thoracentesis catheter was inserted into the thorax. A laser level was used to zero the pressure transducer at the level of the insertion site. The patient did not change position during the procedure.
The analog signal from the pressure transducer, following signal conditioning (DI-5B38; DATAQ Instruments), was digitized at a sample rate of 120 MHz (DI-710; DATAQ Instruments) and analyzed by using a personal computer (WinDaq Software; DATAQ Instruments). Pleural pressure was measured before any fluid removal and following removal of each subsequent 250-mL aliquot of pleural fluid. The pleural pressure curve was recorded over 10 to 15 respiratory cycles at each measurement point. A satisfactory tracing was defined by using a group of stable consecutive respiratory cycles in which the end-expiratory pleural pressure returned to the same baseline.
The final pleural pressure measurement was recorded only if residual pleural fluid was present on ultrasonography examination. Fluid removal was continued until no more fluid could be withdrawn or unless there was evidence of nonexpandable lung defined as a final pleural elastance greater > 14.5 cm H2O/L and/or ipsilateral anterior chest discomfort.
Pleural pressure measurements were derived from five stable breaths at each fluid collection point. Maximal inspiratory pleural liquid pressure (Pliq-ins) and maximal end-expiratory pleural liquid pressures (Pliq-exp) were determined over five consecutive stable breaths following each 250 mL of fluid removal with the values reported as a mean of the five breaths. The difference between Pliq-ins and Pliq-exp was defined as the pleural pressure swing (ΔPliq), and mean pleural liquid pressure (Pliq- ) was reported as (Pliq-ins + Pliq-exp)/2. Pleural elastance was calculated as the change in mean pleural pressure divided by the amount of fluid removed after the initial 500 mL of fluid and at the last aliquot of fluid that was removed as defined by Huggins et al.
Opening pleural elastance was reported after the initial 250 mL of fluid was removed in cases when the maximal fluid removed was 500 mL. Nonexpandable lung was defined as a final pleural elastance > 14.5 cm H2O/L and new or worsening oppressive chest pain during fluid removal.
We did not perform CT chest and/or chest radiographs with air contrast to assess visceral thickening and lung expansion. The etiology of the effusion, the amount of fluid withdrawn, and demographic information was recorded.
Diaphragm Ultrasonography and Measurements
Two-dimensional and M-mode ultrasonography were performed with patients in a sitting position with a 3.5 MHz phased-array probe (P31 M-Turbo; SonoSite) using a longitudinal scanning plane with the probe placed immediately below the costal margin in the midclavicular to anterior axillary line. The posteromedial diaphragm was identified and the M-mode interrogation line adjusted to be perpendicular to the diaphragm. Diaphragm excursion was measured from the M-mode image between end-expiration and end of inspiration and reported in centimeters.
The mean excursion and velocity were measured from a group of five stable breaths before fluid removal and following each subsequent 250-mL fluid aliquot. In each individual patient, the position and angle of the tomographic plane were duplicated with each measurement.
Quantitative data were expressed as means and SDs or medians and interquartile range (IQR) as appropriate. The Student t test and Mann-Whitney test were considered for quantitative parameters according to Student t test assumptions. Kolmogorov-Smirnov and Levene tests were used to check the normal distribution of the data and the equality of variance, respectively. The Pearson correlation coefficient was used to test the correlation between continuous normally distributed variables and the Spearman coefficient in the absent of normality. The paired Student t test was used for pairwise observations. Analyses were performed by using SPSS version 22 (IBM SPSS Statistics, IBM Corporation). A P value < .05 (two sided) was considered significant.
Twenty-eight subjects were enrolled in the study (18 male subjects, 10 female subjects; mean ± SD age, 70 ± 17 years; age range, 35-98 years). Thoracentesis was right-sided in 22 subjects and left-sided in six subjects. The diagnoses, amount of fluid removed, Pliq-, and pleural elastance are summarized in Table 1. Thoracentesis was terminated before complete fluid removal in five patients due to ipsilateral anterior chest discomfort consistent with nonexpandable lung. All five of these patients had a final pleural elastance > 14.5 cm/L, indicating nonexpandable lung.
Table 1Characteristics of Subjects (N = 28)
Etiology of Pleural Effusion
Pleural Fluid Removed (Lt)
Opening Mean Plpr (cm H2O)
Closing Mean Plpr (cm H2O)
Initial Pleural Elastance (cm H2O/L)
Final Pleural Elastance (cm H2O/L)
Expandable lung (n = 23)
Parapneumonic (n = 3)
1.5 (0.5 to 3)
9.8 (–5.5 to 13.9)
– 9.7 (–19.2 to 0.5)
9.7 (5.7 to 14.9)
6.2 (5.7 to 6.5)
Liver cirrhosis (n = 4)
1.3 (1 to 2.1)
–1.4 (–7.7 to 7.9)
–10.7 (–15 to –5)
5.5 (2.3 to 16)
4.7 (2.6 to 8.1)
Malignancy (n = 13)
1.5 (1 to 2)
–1.3 (–3 to 2.7)
–12.4 (–15.5 to –3.9)
5.2 (2.2 to 9)
4.7 (2.4 to 6.6)
Chronic heart failure (n = 2)
1.1 (1 to 1.3)
–6.7 (–8.9 to –4.5)
–10.6 (–16.5 to –4.7)
3.9 (2.7 to 5)
1.7 (0.3 to 3.1)
Uremic (n = 1)
Nonexpandable lung (n = 5)
Parapneumonic (n = 2)
0.75 (0.5 to 1)
–9.25 (–13 to –5.5)
–25.9 (–27.9 to –23.8)
17 (14.5 to 19.4)
21.4 (17.3 to 25.5)
Chronic heart failure (n = 1)
Malignancy (n = 2)
0.5 (0.5 to 0.5)
5.4 (–3.3 to 14)
–10.5 (–16.5 to –4.4)
9.8 (5.6 to 14)
27.4 (26.6 to 28.2)
Data are presented as median (interquartile range) unless otherwise indicated. Lt = liters; Plpr = pleural liquid pressure.
Effect of Thoracentesis on Diaphragm Excursion and Velocity of Diaphragm Contraction
Diaphragm excursion increased from a median of 0.84 cm (IQR, 0.57-1.2 cm) to 1.85 cm (IQR, 1.1-2.0 cm) pre-thoracentesis and post-thoracentesis, respectively (P ≤ .01) in the subjects with expandable lung. Diaphragm excursion increased from a median of 0.59 cm (IQR, 0.53-0.67 cm) to 0.73 cm (IQR, 0.59-0.79 cm) pre-thoracentesis and post-thoracentesis (P = .15) in the subjects with nonexpandable lung (Fig 1). Velocity of diaphragm contraction (mean ± SD) increased from 1.5 ± 0.4 cm/s to 2.8 ± 0.4 cm/pre-thoracentesis and post-thoracentesis (CI, 0.93-1.61; P < .001) in the subjects with expandable lung. Velocity of diaphragm contraction (mean ± SD) increased from 2.0 ± 0.4 cm/s to 2.3 ± 0.4 cm/s pre-thoracentesis and post-thoracentesis (P = .45) in the subjects with nonexpandable lung (Fig 2). ΔPliq increased from a median of 14.0 cm H2O (IQR, 9.1-16.3 cm H2O) to 19.5 cm H2O (IQR, 14-23.4 cm H2O) pre-thoracentesis and post- thoracentesis (P < .001) in the subjects with expandable lung. ΔPliq increased from a median of 15.4 cm H2O (IQR, 9.3-21.2 cm H2O) to 21 cm H2O (IQR, 14.8-25.6 cm H2O) pre-thoracentesis and post-thoracentesis (P = .30) in the subjects with nonexpandable lung (Fig 3). The ΔPliq changes were significantly related to the change in velocity of diaphragm contraction pre-thoracentesis and post-thoracentesis (rho = 0.40; r2 = 0.16; P = .03) (Fig 4).
Pleural Pressure and Amount of Fluid Removed
Patients with nonexpandable lung had a significant drop in the Pliq- , less amount of fluid removed, and increased pleural elastance pre-thoracentesis and post-thoracentesis compared with patients with expandable lung (Table 2).
The results of this descriptive study indicate that the removal of pleural effusion is associated with a significant increase in diaphragm excursion and velocity of contraction in subjects with expandable lung. The improvement in the velocity of the diaphragm contraction was correlated with an increase in respirophasic changes in pleural pressure. The improvement in the diaphragm excursion and velocity may reflect an improvement in contractile function of the diaphragm. This outcome could be explained by an improvement in the length-tension of the diaphragm muscle as well as greater improvement in the apposition of the diaphragm to the chest wall. The improvement in diaphragm function that occurred with pleural fluid removal is closely related to diaphragm configuration, as shown in animal models.
Regarding nonexpandable lung, the increase in diaphragm excursion and velocity of contraction following fluid removal did not reach statistical significance. There was a minimal increase in diaphragmatic excursion with pleural fluid drainage in subjects with nonexpandable lung. This may reflect that the amount of fluid removed in patients with nonexpandable lung was less than could be achieved with subjects with expandable lung. Alternatively, the presence of nonexpandable lung may have restricted the inspiratory movement of the diaphragm. Speculatively, this finding suggests that lack of an increase in diaphragm excursion might have clinical utility for identification of nonexpandable lung during performance of thoracentesis. The small number of subjects with nonexpandable lung does not allow for definitive conclusions.
The change in mean pleural pressure was significantly greater in patients with nonexpandable lung at the end of thoracentesis compared with subjects with expandable lung. This result was predictable, given that nonexpandable lung is defined by an increase in pleural elastance. ΔPliq was increased in subjects with expandable lung at the end of thoracentesis. A possible explanation for the increase in ΔPliq values is that, with removal of pleural fluid, the chest wall configuration changed in such a way to improve the intercostal muscle length-tension curve relationship, resulting in greater ΔPliq values
in combination with improvement in diaphragmatic function (as discussed earlier). Previous research has shown an improvement in the mean vital capacity and end-expiratory lung volume following thoracentesis.
This improvement in lung volume during thoracentesis may lead to better coupling of the lung and chest wall, reestablishing the physiological pleural pressure pattern in subjects with expandable lung. This improvement in diaphragm function may contribute to dyspnea relief in patients with expandable lung.
This study has limitations. First, the impact of diaphragm function in dyspnea relief was not examined in this study; future studies of different design are needed to assess this relevant outcome. Second, the measurement of diaphragm excursion was not checked for interobserver or intra-observer variability. Concerning the former, having two separate investigators make the measurements was not permitted by protocol. It would have prolonged the duration of the thoracentesis with violation of patient comfort and safety. Concerning intra-observer variability of diaphragm excursion, Boussuges et al
reported intra-observer reproducibility of 96% and 94%, and 95% and 91%, during quiet breathing for the right and left diaphragm, respectively. Third, the velocity of diaphragm contraction is not a well-validated measurement of diaphragm function; further research is needed to define the utility of this measurement. Fourth, there was no measurement of the diaphragm thickness and its thickening fraction, both of which are well-established methods for assessing diaphragm function. Measurement of these parameters could have provided additional insight in the etiology of diaphragm dysfunction caused by pleural effusion. We were concerned about the utility of these measurements given that the removal of pleural fluid might alter the position of the diaphragm and/or the zone apposition. Any observed change in diaphragm thickness or diaphragm thickening might reflect a change in the position or geometry of the diaphragm as opposed to a change resulting from contractile function of the diaphragm.
Fifth, there was no “gold standard” used for determination of diaphragm function compared with the experimental measurement. This would have required measurement of transdiaphragmatic pressures. Sixth, measurements of respiratory drive and measurement of the chest wall configuration were not obtained. Seventh, the number of patients with nonexpandable lung was low, reflecting the relative rarity of this condition in patients requiring thoracentesis. Eighth, the study method did not include measurement of changes in diaphragm configuration, as there is not yet a validated means to do so using ultrasonography. Finally, measurement of diaphragm excursion and contraction velocity is affected by factors such as sex, effort, and angle when using M mode, which may influence the results.
In current study, the operator placed the M-mode interrogation line as perpendicular to the axis of diaphragm movement as possible and performed the measurements only during quiet breathing. This method was aided by the presence of pleural fluid, which improved the resolution of the diaphragm.
Diaphragm excursion and velocity of contraction increased significantly with pleural fluid removal in patients with expandable lung. In patients with nonexpandable lung, diaphragm excursion and velocity of contraction did not increase to a significant extent prior to or following pleural fluid removal. Respirophasic changes in pleural pressure increased in subjects with expandable lung with fluid removal. This may result from changes in the length-force relationship of the intercostal muscles and the diaphragm. The measurement of contraction velocity with M mode is a novel method for assessing the diaphragm function, the clinical utility of which will require further examination.
Author contributions: Y. A. G. is the guarantor of the entire manuscript and was the primary data collector, data analysis and author of manuscript. A. P. was responsible for data collection and review of the manuscript. S. J. K., M. N., and P. H. M. were responsible for review of the manuscript and data analysis.