Patterns of pleural pressure amplitude and respiratory rate changes during therapeutic thoracentesis

Large volume pleural effusion leads to an increase in pleural pressure, negatively affects lung volumes and induces clinical symptoms (e.g. dyspnea and cough) [1‐3]. Conversely, therapeutic thoracentesis usually results in a decrease in pleural pressure and has beneficial effects on pulmonary function [2, 4, 5]. Although the number of thoracenteses performed in the USA is reported between 127,000 and 173,000 procedures per year [6, 7], some physiological aspects of pleural fluid removal have not been adequately studied. This is partly because there is no universal and commonly accepted animal model for studying pleural pathophysiology. The relationship between pleural fluid volume and pulmonary function has been evaluated in several human studies. These studies showed that the increase in forced expiratory volume at first second (FEV₁) and forced vital capacity (FVC) after therapeutic thoracentesis approximates 20–30% of the withdrawn pleural fluid volume [4, 8‐10]. Changes in spirometric parameters in patients undergoing therapeutic thoracentesis were related to pleural pressure (Ppl). In a study by Light et al. [4], higher post-thoracentesis Ppl and smaller Ppl decrease after pleural fluid removal were associated with a more significant improvement in FVC.

Some other studies focused on the effect of therapeutic thoracentesis on blood gases. A study by Brandstetter and Cohen [11] showed that PaO₂ decreases significantly early after therapeutic thoracentesis and returns to pre-thoracentesis value after 24 h. Different results were reported in later studies. Perpina et al. found a significant increase in PaO₂ after pleural fluid removal reaching a maximum at 24 h, while Agusti et al. reported no significant effect of therapeutic thoracentesis on blood gases [12, 13]. Changes in blood gases after therapeutic thoracentesis were also studied in relation to pleural pressure and pleural elastance [14, 15]. Chen et al. reported a significantly larger increase in PaO₂ and PaO₂/FiO₂ ratio in patients with normal pleural elastance (PE < 14 cmH₂O/L) compared to patients with high (> 14 cmH₂O/L) pleural elastance.

The use of digital pleural manometers not only allows to precisely assess pleural elastance during pleural fluid removal but also enables monitoring Ppl changes in different phases of the respiratory cycle. In our previous studies we observed some changes in Ppl curve characteristics during pleural fluid withdrawal [16, 17]. These included an increase in Ppl amplitude during the respiratory cycle and changes in respiratory rate. To our knowledge, this effect has not been evaluated and described in details so far. Therefore, we undertook a study aimed at the evaluation of changes in the respiratory pattern and pleural pressure amplitude associated with therapeutic thoracentesis.

Twenty-six patients were initially enrolled. However, three patients had to be excluded from final analysis due to low quality of measurements, i.e. high variability of pleural pressure curve that did not allow reliable calculation of Ppl and Ppl_ampl. Thus, the investigated group included 23 patients (14 M, 9F, median age 68, range 46–85). There were 11 patients with right- and 12 patients with left-sided pleural fluid. Malignant, parapneumonic and rheumatoid pleural effusion was diagnosed in 20, 2 and 1 patient, respectively.

Our study showed different patterns of Ppl_ampl and RR changes during pleural fluid withdrawal. Despite an extensive search of the literature, we could not find any earlier publications reporting the above aspects of therapeutic thoracentesis. The only study that reported Ppl_ampl measurements (described as pleural pressure swings) was that of Boshuizen et al. However, the authors presented only the difference between Ppl_ampl after removal of 200 ml of pleural fluid in terms of expandable vs unexpandable lung [18]. Hence, we believe, our paper may be the first report on Ppl_ampl during therapeutic thoracentesis.

The principle finding of the study is a relatively consequent increase in Ppl_ampl that follows pleural fluid withdrawal. The Ppl_ampl at the end of thoracentesis was higher in all but one patient (95.6%) compared to Ppl_ampl measured directly after pleural catheter insertion. In 65% of patients the Ppl_ampl increased steadily during the procedure. In 13%, there was a transient decrease in the third phase (between points 2 and 3) of the procedure with further significant increase in the fourth phase (between points 3 and 4), while in the remaining 22% patients steady increase in Ppl_ampl was noted in the first two or three phases with a decrease in the last phase of the procedure. Thus, we believe, three different patterns of Ppl_ampl changes during pleural fluid withdrawal can be distinguished. When patients with pattern 1 (steady increase) were compared to patients with pattern 3 (initial increase followed by a decrease) in terms of withdrawn pleural fluid volume, pleural pressure changes during the procedure and other parameters, no differences between these two groups were noted, except median Ppl_ampl at the completion of pleural fluid withdrawal (11.5 IQR 8.3–15.6 vs 5.7 IQR 5.7–9.9, respectively; p = 0.04). When patients were divided into two groups based on median baseline Ppl_ampl (below 3.4 and above 3.4 cm H₂O) we found that the volume of withdrawn pleural fluid and pleural elastance were lower in low median baseline Ppl_ampl group (1760 vs 2150 ml and 8.05 vs 12.68 cmH₂O/L, respectively). Hence, we may speculate about the potential relationship between these parameters and the difference between Ppl_ampl at the end of pleural fluid withdrawal.

The variability in RR changes during the procedure was noticeably higher than that found for Ppl_ampl. In general, an increase in RR was found between the first (baseline) and the last measurement point (p = 0.0097). This was the case in 20/23 (87%) patients. The most common pattern of RR changes during pleural fluid withdrawal was its initial decrease followed by a steady increase. Nonetheless, this pattern was demonstrated in only approximately 1/3 of patients. The other characteristics of RR changes found in 17% and 22% of patients were steady and slow increase from point 0 to point 1 and initial increase with decrease in the last point (or two last points) of measurements, respectively. In the remaining 26% of patients different, difficult to categorize RR changes were observed.

As there were no earlier studies on Ppl_ampl changes during therapeutic thoracentesis, we could not compare our results with other reports. The data on RR changes are also scarce with one study showing a decrease in RR from 19.4± 6.5 to 15.5± 6.3 after drainage of significant pleural effusion [19]. It should be emphasized, however, that the above study was performed in critically ill surgical patients who were on volume-regulated mechanical ventilation.

The paucity of data, together with the initial character of our report and relatively small study group, makes the explanation of our findings challenging. It could have been hypothesized that the overall increase in the respiratory system compliance associated with pleural fluid removal would result in decreased pleural pressure amplitude rather than in its increase. With higher respiratory system compliance, lower transpulmonary pressure should be sufficient to maintain or even increase tidal volume and minute ventilation. On the other hand, it could have been argued that the changes in Ppl_ampl would supposedly be mainly associated with the elastic properties of the lung and visceral pleura. If so, the expansion of lung elastic elements associated with increasing lung volume due to decreasing Ppl that follows pleural fluid withdrawal would result in higher transpulmonary pressure required to maintain tidal volume and minute ventilation. The decline in lung compliance resulting from stretching of its elastic elements associated with lung re-expansion may be particularly significant in the atelectatic lung characterized by relative surfactant deficiency or dysfunction. We believe, the above reasoning is fully consistent with our findings of increasing Ppl_ampl and RR during therapeutic thoracentesis. Significant negative correlation between Ppl_ampl and Ppl and positive correlation between Ppl_ampl and pleural elastance may further support the correctness of our interpretations.

We think that even if the differences in RR were relatively small and their clinical relevance might be questioned, our findings may raise new questions on dyspnea perception, changes in voluntary ventilation and work of breathing after therapeutic thoracentesis. Although pleural fluid withdrawal reduces dyspnea in nearly all patients, the mechanism of this phenomenon is unclear. As improvement in lung function parameters is relatively small and data on blood gases are equivocal, it has been postulated that this is mainly related to the reduced pressure on the diaphragm and subsequent restoration of its normal shape which enables inspiratory muscles to work on a more advantageous position [20, 21]. One study showed that during pleural fluid drainage, the neural impulses from mechanoreceptors located in muscles, chest wall and lung parenchyma significantly contributed to the perception of breathlessness and indirectly influenced the respiratory rate [22]. Paradoxically, the increasing Ppl_ampl during therapeutic thoracentesis found in our study may suggest an increased work of breathing rather than its reduction. It is not known whether the increased Ppl_ampl is associated with increased tidal volume and minute ventilation. These, and other questions should be answered in further studies which would involve continued measurement of air flow through the airways enabling estimation of tidal volume and the minute ventilation.

A prerequisite for our study on the physiological effects of pleural fluid and therapeutic thoracentesis is pleural manometry. Although pleural pressure measurement was commonly applied to create and maintain artificial pneumothorax used to treat tuberculosis a century ago, its modern use in studying pleural pathophysiology began in early 1980s and was associated with impressive results presented by Light and colleagues [23]. The authors demonstrated three different patterns of pleural elastance which were further confirmed in other studies [24‐26]. However, water manometers or modified overdamped water manometers used in earlier studies had some significant limitations. They did not allow either for the observation of pleural pressure oscillations associated with respiratory cycle or for the precise measurement and registration of instantaneous pleural pressure [25]. Introduction of new systems based on electronic pressure transducers opened new possibilities to study pleural pathophysiology during pleural fluid removal [14, 26‐29]. This refers not only to the measurement of instantaneous pleural pressure but also to data display, registration and analysis [2, 16, 18, 30]. Electronic manometers had very low inertia and high frequency of measurements (50 Hz) allowing reliable measurements of pleural pressure even during cough [17]. The use of electronic systems of pleural pressure registration allows very precise and detailed analysis of pleural pressure curve, as showed in our study and the paper by Boshuizen et al. [18]. We believe that further developments in monitoring of different parameters of respiratory system will help to precisely explain the pathophysiological effects of pleural fluid accumulation and its removal.

We are aware about some limitations of our study. First, the investigated group was relatively small. The limited number of patients does not allow a full assessment of Ppl_ampl and RR characteristics in subgroups presenting different patterns of changes. Very small groups of patients with inconsecutive increase in Ppl_ampl (only 3 patients) and Ppl_ampl decrease in the last phase of the procedure (5 patients) do not allow a reliable comparative analysis between of different variables in those groups. Second, as the vast majority of our patients had malignant pleural effusion we cannot be sure that the characteristics of Ppl_ampl and RR changes would be similar in patients with non-malignant pleural effusion. Third, the analysis of Ppl_ampl and RR changes in relation to the relative volume of withdrawn pleural effusion (% of the total volume removed in individual patients) may be questioned. However, this method of analysis was applied to make our results relatively independent of the absolute volume of removed pleural fluid. This approach was based on the assumption that in all patients therapeutic thoracentesis is a complete procedure independently of the volume of the withdrawn pleural fluid and justified the comparison of the procedures of removal of different volumes of pleural fluid. In our opinion, the analysis of Ppl_ampl and RR as the function of absolute volume of removed pleural fluid in patients with removed pleural fluid volume ranging from 500 to 4250 ml would be even more controversial.

In conclusion, therapeutic thoracentesis is associated with significant changes not only in pleural pressure but also in pleural pressure amplitude during the respiratory cycle. Contrary to pleural pressure, the pleural pressure amplitude steadily increases in the vast majority of patients. The significance of Ppl_ampl and RR changes associated with pleural fluid withdrawal should be elucidated in further studies.