INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In normal humans, a small amount of pleural fluid is present (1, 2). The exact volume of this fluid is unknown. Careful measurements in rabbits and dogs have yielded volumes of pleural fluid of 0.1 to 0.3 ml/kg (2), and a similar volume is thought to be present in normal humans (1).

Normal pleural fluid, at least in laboratory animals such as rabbits and dogs (3, 4), contains a significant number of cells. Total cell counts vary from 1,500 to 2,450/µl, with a high variance observed in differential cell counts of 9% to 70% mesothelial cells, 28% to 70% macrophages, 2% to 11% lymphocytes, and 0% to 2% polymorphonuclear leukocytes (3, 4).

No reliable data are available on the cellular content of pleural fluid in normal humans. The only study addressing this issue was that of Yamada (5), published in 1933, who punctured the 9th or 10th intercostal space on the dorsal axillary line in a group of healthy Japanese soldiers. In about 30% of cases, some liquid was retrieved after a period of rest, and in about 70% of cases some liquid was retrieved after exercise. Usually, only a few drops of foam were collected, but in a few (probably abnormal) cases, up to 20 ml of fluid was collected. In this study, Yamada found a total white cell count of 4,500 cells/µl (range: 1,700 to 6,200 cells/µl). A differential cell count showed 3% mesothelial cells, 53.7% cells "similar to monocytes," 10.2% lymphocytes, and 3.6% granulocytes, but also 29.5% "deteriorated cells of difficult classification." Hence, because of the obvious technical difficulty in retrieving nontraumatically the few milliliters of fluid present, the composition of normal pleural fluid remains one of the few last secrets of human body fluid composition.

In our clinical practice, we weekly perform thoracoscopic interventions for the treatment of severe essential hyperhidrosis in otherwise healthy subjects (6), thus having access to normal pleural spaces through a minimally invasive technique. In analogy with bronchoalveolar lavage (BAL) (7), which allows retrieval and examination of the few milliliters of epithelial lining fluid in the examined lung, we designed a new technique, pleural lavage (PL), enabling retrieval and analysis of the few milliliters of pleural fluid present in normal humans.

The purpose of the present study was to use PL to: (1) measure the volume of pleural fluid present in the pleural space of normal humans; and (2) determine the total and differential cell contents of this normal pleural fluid.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

After having successfully performed a feasibility study in four patients, we performed PL in 34 consecutive patients referred for thoracoscopic treatment of severe essential hyperhidrosis. These patients included 21 nonsmokers (nine male and 12 female), aged 25.4 ± 8.0 yr (mean ± SD) (range: 17 to 45 yr), and 13 smokers (five male and eight female), aged 26.5 ± 8.7 yr (range: 18 to 50 yr). Total and differential counts were compared in the nonsmoking (n = 21) and smoking (n = 15) subjects. Among the nonsmoking subjects, cell counts were compared in male (n = 9) and female (n = 12) subjects. In five subjects, PL was performed bilaterally for within-subject right-left pleural fluid comparison.

In a subgroup of 16 nonsmoking subjects (five male and 11 female), values of right pleural fluid volume (Vpv) were determined with a modified urea dilution method. In five of these subjects, sequential right- and left-sided Vpv determinations were also performed for within-subject comparisons.

The subjects had no history of prior pulmonary or pleural disease, and a detailed history and physical examination, chest radiography, and complete pulmonary function testing confirmed that all subjects were in perfect health except for the presence of invalidating essential hyperhidrosis at the palmar (100%) and/or axillary (40%) level. Informed consent was obtained from every patient, and the study was approved by our local ethics committee.

PL Technique

PL was performed during the thoracoscopic sympathicolysis procedure. All procedures were performed under strict aseptic conditions in a fully equipped operating theater. After intramuscular premedication with midazolam (0.1 mg/kg) and glycopyrolate (0.2 mg), patients were anesthetized with propofol, alfentanil, and atracurium. Patients were in the supine, 30-degree trunk-up position. Intubation was performed with a single-lumen Hi-Lo jet endotracheal tube (Mallinckrodt, Northampton, UK), and ventilation and oxygenation were assured by use of an AMS 1000 high-frequency jet ventilator (Acutronic, Jona, Switzerland).

A right-sided pneumothorax was created by blunt puncture of the parietal pleura at the 2nd intercostal space with a Küss or Boutin needle and insufflation of ~ 250 ml of ambient air. The puncture site was bluntly enlarged to ~ 10 mm, and a 7-mm trocar was gently introduced into the pleural cavity. After quick inspection with a thoracoscope (Richard Wolf, Knittlingen, Germany), a second 5-mm trocar was gently and bluntly introduced 3 to 4 cm laterally from the first trocar. A nontraumatic 14-gauge polyvinyl chloride catheter (Medicoplast, Jillingen, Germany) was introduced via the second trocar and positioned under direct thoracoscopic control in the lowest portion of the lateroposterior costophrenic angle.

A volume of 150 ml of prewarmed, sterile saline was injected, and was immediately and gently aspirated (minimal dwell time). The aspirated fluid was collected in a sterile jar and immediately transported on ice to the laboratory for processing.

After the PL procedure, sympathicolysis was performed as described elsewhere (6). In the subjects in whom PL was performed bilaterally for within-subject comparison, the same procedure as described earlier was performed on the left hemithorax, after termination of the right-sided intervention.

PL Fluid Processing

Cellular analysis. The PL fluid was processed in a manner similar to bronchoalveolar lavage fluid (BALF) (except for mucus filtration) (8). After measurement of the total volume of PL fluid, a drop of the well-mixed fluid was introduced into a Bürker chamber for total cell counting. Red and white blood cells were counted separately.

The PL fluid was then centrifuged (3 min at 1,000 × g). Ten milliliters of supernatant fluid were removed and frozen at 80° C for later solute analysis. The cell pellet was washed twice with 1 ml of phosphate-buffered saline (PBS)-1% albumin. Thereafter, the cells were resuspended in PBS-5% albumin until a suspension of approximately 10⁶ cells/ml was obtained. Cytospins were then made with a cytocentrifuge (Cytospin-3; Shandon Scientific Limited, Astmoor, UK). The cytospin preparations were air-dried and one slide was stained with May-Grünwald-Giemsa and the other slides were frozen at 20° C.

Differential counts were made on the May-Grünwald-Giemsa-stained slide by counting 500 cells by light microscopy (×1,150). Macrophages, mesothelial cells, neutrophils, eosinophils, basophils, and lymphocytes were differentiated.

Immunophenotyping of the cells was performed with monoclonal antibodies and immunogold-silver staining (9). Air-dried cytocentrifuge preparations were incubated with monoclonal antibodies (mAbs) obtained from Becton-Dickinson Immunocytometry (Mountain View, CA) and directed against leukocyte cell-surface antigens. Surface antigens studied were CD3 (Leu4), CD4 (Leu3a), CD8 (Leu2a), and CD19 (Leu12); mouse IgG₁ and CD45 (Hle-1) were used as negative and positive controls. Only the CD4⁺ to CD8⁺ cell ratio is reported here. Colloidal gold-labeled goat antimouse IgG + IgM antibodies were purchased from Amersham-Pharmacia-Biotech (Roosendaal, The Netherlands). Light microscopy-grade reagens with 5-mm gold particles were used. Silver enhancement was performed with the silver enhancing kit (British Biocell International, Cardiff, UK), following the recommendations of the manufacturer.

Details of the immunogold-silver staining procedure performed on the cytocentrifuge preparations are summarized elsewhere (10). In short, air-dried preparations were fixed for 30 s at room temperature (RT) with phosphate-buffered 9.25% formol and 45% acetone, pH 6.6. The preparations were rinsed in PBS and then incubated in the horizontal position for 10 min in a humidified chamber with 25 µl of 0.01 M PBS containing 5% nonfat dry milk. Following this, 25 µl of the appropriate dilution of mAb was added and the mixture was left for 30 min at RT. The preparation was rinsed with PBS to remove the excess reagent, and was incubated with PBS-milk-mAb. Twenty-five microliters of a 1:75 dilution of the goat antimouse IgG + IgM reagent was added and left for 30 min at RT. After being rinsed with PBS, the preparations were postfixed with buffered formol-acetone for 2 min at RT. After rinsing with distilled water (thrice for 5 min each), silver enhancement was performed for 30 min at 26° C. The preparations were then rinsed again, were air dried, and were counterstained with May-Grünwald-Giemsa stain and mounted in DPX Mountant (BDH, Montreal, Canada).

The labeled preparations were then examined with a Dialux 22 EB brightfield microscope (Leitz, Wetzlaar, Germany). The cells were identified by their morphology. Four hundred cells were examined, and the percentage of positive cells was determined.

Volume Analysis

In a subgroup of 16 nonsmoking patients (five male and 11 female; age 24.1 ± 6.5 yr; range 17 to 37 yr), volume analysis of the pleural fluid present in the right hemithorax was performed. In five of these patients, volume measurements were made in the right and left hemithorax.

In analogy with the estimation method of volume of epithelial lining fluid recovered by BAL with urea used as an endogenous marker of dilution (11), the subsequently described methodology was designed.

Measurement of the effective volume of pleural fluid present in a hemithorax was based on the assumption that the urea concentrations in plasma and in normal pleural fluid are equal (since urea diffuses freely throughout the body, including passage through alveolar membranes [12] and probably also through pleural membranes [13]), on the principle of conservation of mass, and on the assumption that after induction of a pneumothorax, most of the pleural fluid present would drain by gravity to the lowest portion of the pleural space (3).

Starting from the original situation of V_PV in milliliters of pleural fluid present, the total amount of urea (M_U) present in this fluid can be expressed as M_U = C_PV × V_PV (1), where C_PV equals the concentration of urea in the original pleural fluid (in mg/ml). After pleural lavage with 150 ml of saline, the conservation of mass principle states that M_U = (150 + V_PV) × C_PL, where C_PL is the measured concentration of urea (in mg/ml) in the retrieved PL fluid.

Therefore, C_PV × V_PV = (150 + V_PV) × C_PL. Hence, V_PV = (150 × C_PL)/(C_PV  C_PL). Since it is assumed that C_PV equals C_plasma (where C_plasma is the concentration of urea in plasma) (12): V_PV = (150 × C_PL)/(C_plasma  C_PL). To minimize the possible effects of urea influx into the pleural space during the lavage procedure, the dwell time of the lavage fluid was kept to a minimum (a few seconds to 1 min) (14).

C_plasma was measured on the day of PL with an endpoint urease method and a Vitros 950 analyzer (Johnson & Johnson Clinical Diagnostics Inc., Rochester MA), using dedicated dry chemistry reagents. To determine the (low) concentrations of urea in the retrieved PL fluid (C_PL), a modified kinetic enzymatic method, including urease and glutamate dehydrogenase (Unimate 5 Urea; Hoffmann-La Roche, Basel, Switzerland) and adapted to a Cobas Mora S analyzer (Hoffmann-La Roche), was used. The sensitivity of the method, originally designed for the measurement of urea concentrations in serum, plasma, or urine, was improved by increasing the sample volume fraction from 0.01 to 0.16. All samples were measured in a single analytical run. Within-run coefficients of variation (n = 8) ranged from 4.7% to 8.8% for urea concentrations between 1.1 mg/dl and 0.4 mg.dl. The limit of detection was 0.024 mg/dl. The V_PV in one hemithorax is expressed per kilogram of body mass.

Statistical Analysis

Cellular data are expressed as medians and interquartile ranges (IRs) because of the apparent nonnormal distribution of the results. Comparisons between data for smokers and nonsmokers were made with chi-square analysis for sex and the Mann-Whitney U test for cell counts (15). Comparisons of data for males and females were made with the Mann-Whitney U test. Within-subject right-left paired comparisons were made with Wilcoxon's signed ranks test. Volume data are expressed as mean ± SD because of the apparent normal distribution of data. Within-subject paired comparisons of right- and left hemithorax pleural fluid volumes were made with a paired t test. Significance was accepted at p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All PL procedures were uneventful and prolonged the thoracoscopy procedure by only 3 to 5 min. The recovered volume of fluid was 141.2 ± 10 ml (mean ± SD) (94% of the injected volume). Postoperative chest radiographs never showed visible residual fluid. In all subjects, the macroscopic appearance was that of clear fluid, and there was no macroscopically visible contamination with blood.

Results of Cellular Analysis

Results of total (red and white blood cells) and differential cell counts in nonsmokers and smokers are summarized in Table 1.

There were no significant differences in cell counts for nonsmokers and smokers, except for neutrophils, which were present in slightly but significantly higher numbers in the smoking group (median: 1% versus 0%; IR: 1% to 3% versus 0% to 1%; p = 0.015, Mann-Whitney U test). Within the smoking group, cigarette consumption was 11.4 ± 14.1 pack-years. There was no correlation between pack-years smoked and PL neutrophil content. (Spearman's rank correlation, p > 0.05).

Within-subject right-left comparison showed no significant differences in total or differential cell counts between the right and left pleural cavities (data not shown). Within the nonsmoking group, total and differential cell counts of the male and female subjects were also compared. There were no significant sex differences (data not shown).

Results of Analysis of Vpv

In all cases studied, plasma and PL fluid urea concentrations could be measured. The right C_PL was 1.16 ± 0.41 mg/dl (range: 0.51 to 2.12 mg/dl).

C_plasma was 23.07 ± 4.47 mg/dl (range: 16 to 31 mg/dl). Hence, the original Vpv in the right pleural cavity was 8.4 ± 4.3 ml (range: 3.7 to 17.8 ml). PL with an instilled volume of 150 ml of saline therefore represents a dilution factor of (150 + 8.4): 8.4 = 18.86.

Hence, the median total WBC count of 91 × 10³ cells/ml in the diluted PL fluid corresponds to a total WBC count of 91 × 18.86 = 1,716 × 10³ WBC/ml in the original pleural fluid. Expressed per kilogram of body weight, the right-sided Vpv was 0.13 ± 0.06 ml/kg (range: 0.06 to 0.25 ml/kg).

In five subjects, measured values of Vpv in the right and left hemithorax were not significantly different (p = 0.6), at 12.6 ± 5.3 ml (range: 3.9 to 17.9 ml) and 11.8 ± 6.8 ml (range: 2.8 to 18.3 ml), respectively.

Therefore, the total (right + left) Vpv calculated for the whole nonsmoking study population and expressed per kilogram of body weight, averaged 0.26 ± 0.1 ml/kg.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This is the first report of measurement of the actual volume and total and differential cell contents of the pleural fluid present in normal subjects, done through a minimally invasive, atraumatic thoracoscopic PL technique performed on otherwise healthy hyperhydrosis patients.

Using urea as an endogenous dilution marker, we determined the Vpv of normal humans to be 0.26 ± 0.1 (mean ± SD) ml/kg body mass. This corresponds quite well with the animal-derived proposed volume of 0.3 ml/kg body weight (2). Urea has been used in the past to estimate the volume of epithelial lining fluid recovered by BAL, on the assumptions that the concentrations of urea in plasma and epithelial lining fluid are similar and constant across time (i.e., no movement of urea occurs between the vascular and air space compartments during lavage [11]). In the specific context of BAL, this technique, in which urea is used as an endogenous marker, has been largely abandoned because enough urea may enter the air spaces during the BAL procedure to significantly alter the calculated values of dilution, hence leading to more than a threefold overestimation of actual epithelial lining fluid volumes (14). This is mainly due to the relatively long dwell time of the BAL solution (2 min or more) and the often increased permeability of the barrier separating the plasma and air spaces in pathologic states (16). However, in the context of the PL technique, urea may be an accurate endogenous marker of dilution because: (1) dwell time is kept to a minimum (i.e., a few seconds to 1 min maximum, hence minimizing urea flux across the membranes during the procedure; and (2) measurements are made in healthy pleural cavities with intact visceral and parietal pleura. Furthermore, admixture with blood caused by the (small) trauma of creation of the pneumothorax could in our study be excluded, since only very small numbers of red blood cells were present in the PL fluid.

In the nonsmoking subjects in our study, a median of 91 × 10³ (IR: 74 to 198 × 10³) WBC/ml of PL fluid was measured. Since the originally measured Vpv was 8.4 ± 4.3 ml (corresponding to a lavage-induced dilution factor of 18.86), the total WBC count in the original pleural fluid can be calculated at 1,716 × 10³ cells/ml. This corresponds well with total cell counts made on original pleural fluid in rabbits and dogs, which varied from 1,500 to 2,450 × 10³ cells/ml (3, 4), but is in the lower range of the 1,700 to 6,200 × 10³ cells/ml measured in the only available (but poorly reliable) human study done by Yamada (5).

Differential cell counts done on the PL fluid in our study showed that the bulk of its cellular content consisted of macrophages (median: 75%; IR: 64% to 80%) and lymphocytes (median: 23%; IR: 18% to 36%), with a median CD4⁺-to-CD8⁺ T-cell ratio of 0.75 (IR: 0.6 to 1). Mesothelial cells accounted for only 1% (IR: 0% to 2%), and neutrophils (median: 0%; IR: 0% to 1%) and eosinophils (median: 0%; IR: 0%) were only marginally present.

The difference between these results and the differential count proposed by Yamada (5) may be due to the almost certainly present (3) contamination by blood in Yamada's series, owing to the traumatic retrieval technique that was used, and/or to the differences in staining techniques and interpretation (especially for mesothelial cells and macrophages) in Yamada's paper and the present study.

Interestingly, as compared with nonsmokers, smoking subjects showed a slight but statistically significant increase in pleural fluid neutrophils (median: 1%; IR: 1% to 3%; p = 0.02). An analogous increase in BAL neutrophils has been observed in smokers (17). Smoking, however, induces numerous other cellular and solute changes in BALF (e.g., a decrease in the CD4⁺-to-CD8⁺ lymphocyte ratio without a change in total T-cell number). In PL fluid, however, only changes in neutrophil content were observed. The mechanism by which this phenomenon occurs is unclear, but it probably reflects "transvisceral" extension of the neutrophil influx into the lower airways caused by smoking (18).

In conclusion, application of a new and simple PL technique performed during thoracoscopy has allowed the first measurement of volume and of total and differential cell counts in the pleural fluid of normal humans.