RICHARD WILLIAM STEPHEN RONALD
In this study the incidence and course of pleural effusions (parapneumonic effusions) in patients with acute bacterial pneumonia were prospectively evaluated. Bilateral decubitus chest x-ray films were obtained within 72 hours of admission in 203 patients with an acute febrile illness, purulent sputum and an infiltrate evident on the chest film. Ninety of the 203 patients (44 percent) had pleural effusions. Parapneumonic effusions, which required chest tubes for resolution and/or on which the pleural fluid cultures were positive, were classified as complicated parapneumonic effusions. The IO patients with complicated parapneumonic effusions had clinical characteristics similar to the remainder of the group and could be separated from the 80 with uncomplicated effusions only by pleural fluid analysis. A pleural fluid pH below 7.00 and/or a glucose level below 40 mg/lOO ml are indications for immediate tube thoracostomy. In patients with pleural fhdd pH between 7.00 and 7.20 or lactic dehydrogenase (LDH) above 1,000 IU/l,OOO ml, tube thoracostomy should be considered, but each case should be individualized; serial studies of the pleural fluid are useful in some of these cases. Patients with pleural fluid pH above 7.20 and pleural fluid LDH below 1,000 mg/lOO ml rarely have complicated parapneumonic effusions and do not require serial therapeutic thoracenteses.
W. LIGHT, M.D. M. GIRARD, M.D.* G. JENKINSON, M.D. B. GEORGE, M.D.
Patients with bacterial pneumonia sometimes have an associated pleural effusion (parapneumonic effusion). The mortality of patients with pneumonia and pleural involvement is higher than that of those without such involvement [l]. Most parapneumonic effusions clear spontaneously without the use of chest tubes [2,3]. However, some do not. It is important to be able to identify those patients who will need tube thoracostomy as early as possible since drainage of the pleural space becomes progressively more difficult the longer the fluid is present [4-61. The purpose of this prospective study was to determine the incidence of parapneumonic effusion and to analyze the factors related to its occurrence. In addition, the value of various laboratory tests in identifying those effusions which would not resolve without tube thoracostomv was studied. From the Departments of Medicine, Veterans Adminstration Medical Center. Lone Beach University of California Irvine, Irvine,vCalifornia; and the Louisiana State University Shreveport, Shreveport, Louisiana. Requests for reprints should be addressed to Dr. Richard W. Light, Pulmonary Disease Section, Veterans Administration Medical Center, Long Beach, California 90822.
*Present address: YIODO, P.O. Box 45, Seol150, Korea.
The study population consisted of patients admitted to the medical service at the Louisiana State University Medical Center between June 1, 1975, and November 1, 1977, with the diagnosis of bacterial pneumonitis. The patients were identified by screening the list of medical admissions each day. All patients who had an admission diagnosis compatible with bacterial pneumonia
were evaluated. To be included in the study, each patient was required to have an acute febrile illness, purulent sputum and an infiltrate on the chest x-ray film. All patients who met these three criteria were included in the study.
Journal of Medicine
Bilateral decubitus chest x-ray films were obtained within 72 hours of admission on all study patients. If pleural fluid was present on the decubitus x-ray film, its amount was semiquantitated by measuring the thickness of the fluid on the decubitis x-ray film. A diagnostic thoracentesis was performed when the thickness of the pleural fluid on the decubitus film was more than 10 mm. Patients with pleural effusions were classified as either having complicated or uncomplicated parapneumonic effusions. For this study, a complicated parapneumonic effusion was defined as a parapneumonic effusion for which tube thoracostomy was necessary for its resolution and/or on which the bacterial cultures of the pleural fluid were positive. The decision to place chest tubes was made by the patient’s attending physician. All patients were followed until they were discharged from the hospital or their pulmonary infiltrates resolved. Sputum gram stains and cultures, and blood cultures were obtained on all patients; routine aerobic and anaerobic pleural fluid cultures were obtained on all patients who had thoracentesis. Based on these data, the patients were placed in one of the following six categories: (1) Definite pneumococcal pneumonia (blood culture positive for Streptococcus pneumoniae); (2) probable pneumococcal pneumonia [sputum smear compatible with and sputum culture positive for Strep. pneumoniae); (3) possible pneumococcal pneumonia (sputum smear compatible with, sputum culture negative for Strep. pneumoniae); (4) other gram-positive pneumonia (sputum smear and culture positive for Staphylococcus aureus or strep. pyogenes); (5) anaerobic lung infection (blood or pleural fluid culture positive for anaerobic organisms or the presence of a putrid lung abscess]; and (6) gram-negative pneumonia (compatible sputum, smear with confirmatory sputum, pleural fluid or blood cultures).
During the study period, a total of 203 patients fulfilled our criteria for inclusion in the study. Of these, 90 (44.4 percent] had pleural effusions. The clinical characteristics of the patients with and of the patients without pleural effusions were very similiar (Table I). There was no significant difference between the two groups in any of the characteristics tabulated. Thoracentesis was performed on 37 of the 90 patients with parapneumonic effusions. According to our definition of a complicated parapneumonic effusion, in 10 of these 37 patients the parapneumonic effusions (27 percent] were complicated. These 10 patients constitute 5 percent of the original group of patients admitted to the study. Of these 10 patients with complicated parapneumonic effusions, pleural fluid cultures were positive in seven: in four, anaerobic bacteria were recovered and in one each, Pseudomonas aeruginosa, Strep. pneumoniae and Strep. pyogenes. Three patients were classified as having complicated parapneumonic effusions because chest tubes were required for the resolution of their parapneumonic effusions even though the pleural fluid cultures were negative. In each
Comparison of Patlents with Pneumonia With and Without Parapneumonic Effusion Wtth Effusion
Number Age WI* White blood cell count (per mm3)’ Peak temperature Pleuritic chest pain (“/o) Lobes involved (no.) Lower lobe involvement (%) Right-sided pneumonia (%) Left-sided pneumonia (%I Bilateral pneumonia (%I Bronchopneumonia (%I l
90 51.4 f 22.3 17,809 f 14,040 102.2 f 1.5 64
Elklslon 113 45.0 f 20.2 17,088 f 11,682 102.0 f 1.3 59
P Value >0.05 >0.05 >0.05
of these three patients, thoracentesis yielded thick pus with a very low pleural fluid glucose level (1 to 7 mg/lOO ml], very high pleural fluid LDH level (2,000 to 13,000 mg/lOO ml) and a very low pleural fluid pH (5.64 to 6.90). All three had received antibiotics before the initial thoracentesis. As already stated, the clinical characteristics of the patients with complicated and uncomplicated parapneumonic effusions who underwent thoracentesis were similar (Table II). It is noteworthy that only five of the 10 patients with complicated parapneumonic effusions had pleuritic chest pain. The patients with complicated parapneumonic effusion could be differentiated from those with uncomplicated parapneumonic effusion with pleural fluid analysis. Patients with complicated parapneumonic TABLE II
Characteristics of Patients with Complicated and Uncomplicated Parapneumontc Effusion Complkxrtsd
White blood cells (serum)* Peak temperature+ Pleuritic chest pain (%I Lobes involved (mean no.) Lower lobe involvement (%) l
53.3 f 17.8 22,240 f 8,590 102.1 f 1.7 50
27 51.1 f 24.4 19,600 l 23,910 101.9 f 1.5 67 1.44 89
>0.05 >0.05 >0.05 >0.05 >0.05 >0.05
. . . .
Figure 1. complicated
The pleural fluid pH, LDH and glucose in unand complicated parapneumonic effusions.
effusions tended to have higher pleural fluid LDH and lower pleural fluid pH and glucose than did patients with uncomplicated parapneumonic effusion (Figure 1). All the patients with a pleural fluid pH below 7.00 or a pleural fluid glucose level below 40 mg/lOO ml had complicated parapneumonic effusions. All the patients who had a pleural fluid pH above 7.20 or a pleural fluid LDH level below 1,000 IV/l,000 ml had uncomplicated parapneumonic effusions. Patients with pleural fluid TABLE III
Pleural Fluid Flndinas
6.64 6.76 7.20 7.22 7.20 7.33 7.34 7.44 7.25 7.33
2 3 4 5
IO/14 217 218 3120 3121 313 3110 10125 10127
with Serial Thoracentesis
1 2 0 95 93 127 75 74 99 a3 54
in Five Patients
Glucoss (mg/lOO ml) 3,780 13,360 1,368 1,074 774 534 1,180 318 1,330 484
pH between 7.00 and 7.20, an LDH level above I,OOO mg/lOO ml or a glucose level above 40 mg/lOO ml could have either complicated or uncomplicated parapneumanic effusions. In contrast, the pleural fluid white blood cell count and protein levels were not useful in differentiating complicated from uncomplicated parapneumonic effusions. Several patients had more than one thoracentesis during their hospitalization (Table III]. One patient (Case 1) was admitted with an acute febrile illness and a pleural effusion. Although gram stains and cultures of his pleural fluid were persistently negative, chest tubes were inserted when his pleural fluid pH and glucose level remained very low, and the pleural fluid LDH level continued to increase. Another patient (Case 2) had a relatively low pleural fluid pH and high LDH level at the time of his initial thoracentesis and, when the initial pleural fluid culture was positive for group A Strep. pyogenes, a repeat thoracentesis was performed. This second thoracentesis revealed a lower pleural fluid LDH level, a slightly higher pH and no organisms on gram stain. Chest tubes were not inserted, and the patient’s pneumonitis and pleural disease resolved without undue delay. This was the only person in this series who was classified as having a complicated parapneumonic effusion who did not undergo tube thoracostomy. Three patients [Cases 3,4 and 5) had a second thoracentesis when the original thoracentesis resulted in some pleural fluid findings which were worrisome. All were thought to have pneumonia secondary to Strep. pneumonia and antibiotic therapy had been started. The second thoracentesis in each showed a higher pleural fluid pH and a lower pleural fluid LDH level. The bacteriologic findings in our patients are summarized in Table IV. The incidence of parapneumonic effusions in patients with presumed pneumococcal pneumonia (40 percent) was significantly (X2 = 4.32, p < 0.05) less than that in the patients with nonpneumococcal pneumonia (58 percent). Indeed, in none of the 15 patients who had positive blood cultures for Strep.
. uncomplicated A Complicated
While Blood Cells (per mm31
2,300 6,500 72,000 21,000 28.000 32,100 16,600 2,400 126,500 46.800
Negative Negative Negative Strep. pyogenes Negative Negative Negative Negative Negative Negative Negative
Journal of Medicine
pneumonia was a pleural effusion large enough to perform a thoracentesis. Complicated parapneumonic effusions were significantly less common in patients thought to have pneumococcal pneumonia than in the remainder of the patients (X2 = 14.4, p < O.OOl].Of the 153 patients thought to have pneumococcal pneumonia, only two had a complicated parapneumonic effusion. In one of these the pneumonia was distal to a bronchus occluded by a bronchogenic carcinoma. The other probably had a complicated parapneumonic effusion secondary to anaerobic bacteria. His sputum smear revealed gram-positive diplococci, but his sputum culture was negative. His pleural fluid was characterized by a low pH (6.72) and glucose level (5 mg/lOO ml] but a negative culture. Seven of our 10 complicated parapneumonic effusions were thought to be due to anaerobic bacteria. The pleural fluid cultures were positive for anaerobic bacteria in four of the patients [one was also positive for aerobic bacteria). Anaerobic bacteria were probably also responsible for the three patients with complicated parapneumonic effusions who had negative cultures. All had received antibiotics before the initial thoracentesis was performed. The incidence of pleural effusion with gram-negative pneumonias was high. Of the 15 patients with probable gram-negative pneumonia, nine (60 percent) had pleural effusion and in seven (47 percent) there was sufficient fluid for thoracentesis. Of the six patients with gramnegative pneumonia and no pleural fluid, five were thought to have pneumonia secondary to H. influenzae. Mortality in the patients with pleural effusion was higher than the mortality in those without pleural effusions. Only one of the 113 patients with pneumonia without effusion died whereas seven of the 90 patients with pneumonia and effusion died (X2 = 4.61, p < 0.05). However, none of the seven patients with pneumonia and pleural effusion who died were thought to have died from pleural sepsis. The only patient with a complicated parapneumonic effusion who died had Pseudomonas pneumonia and Pseudomonas bacteremia. Five patients with uncomplicated parapneumonic effusions were thought to have died from their initial pneumonitis. One patient died with renal failure.
on 203 Patients Fluid Abselll
Strep. pneumoniae Definite Probable Possible Other gram-positive Anaerobes Gram-negative
7 28 15
Pleural effusion frequently occurs with bacterial pneumonia. The incidence of parapneumonic effusions in the present study (44 percent) is comparable to that found by other investigators. Taryle et al.  obtained decubitus chest x-ray films every 24 to 72 hours on 35 patients with presumed pneumococcal pneumonia during their hospitalization. They found that 20 of the 35 patients (57 percent) had pleural effusion. Bartlett and Finegold  reported that 51 of 142 patients (36 percent) with anaerobic infections of the lungs had pleural fluid. Small parapneumonic effusions are even common with nonbacterial pneumonia. Fine et al.  demonstrated pleural fluid with decubitus chest roentgenograms in 12 of 59 patients (20 percent) with Mycoplasma or viral pneumonia. Although more bacterial pneumonias are caused by Strep. pneumoniae than by any other bacteria, pleural infection by this organism is uncommon. Bartlett et al. [lo] reported that Strep. pneumoniae grew from only five of 83 pleural fluids with positive bacterial cultures
with Pneumonia Fluld Present
92 15 68 70
In general, the parapneumonic effusions tended to be near their maximum size at the time that the original decubitus chest x-ray films were obtained. In only four patients did the amount of pleural fluid increase enough to warrant a subsequent thoracentesis when there were less than 10 mm pleural fluid at the time of the initial decubitus chest x-ray film. Two of these patients, both of whom were thought to have pneumococcal pneumonia, had no fluid originally and had uncomplicated hospitalizations. However, after discharge and within three weeks of their initial admission, both had pleuritic chest pain and pleural effusion. Thoracentesis in each revealed exudates with more than 15 percent eosinophils. The effusions in both patients resolved spontaneously without further antibiotic therapy and with only symptomatic treatment. The other two patients had a small amount of fluid on the original decubitus chest x-ray film which increased substantially over the next several days. Thoracentesis in each patient revealed sterile exudative pleural effusions with no eosinophils. The pleural effusions in both subsequently resolved without tube drainage.
11 43 38 3 12 6
4 25 32 4 16 9
2 0 4 11
1 5 6
Performed Complicated 0 1 1 1 6 1
whereas Snider and Saleh [II] found Strep. pneumoniae in only six of 92 patients with positive pleural fluid cultures. In the present series, Strep. pneumoniae grew from only one of the 10 complicated effusions even though it was thought to be responsible for 75 percent of the pneumonias. Noteworthy is the observation that none of the 15 patients with positive blood cultures for Strep. pneumoniae had pleural effusions that were large enough to justify thoracentesis. The findings in the present series in which anaerobic bacteria were grown from four of the 10 complicated effusions emphasize the importance of anaerobic bacteria in pleural infection. The high incidence of anaerobic pleural infections has been documented by Bartlett et al. [lo]. They reported that anaerobic organisms alone were cultured in 23 of 83 cases (35 percent) of pleural infection, and anaerobic plus aerobic organisms were cultured in an additional 34 (41 percent). The three complicated parapneumonic effusions in the present series with negative pleural fluid cultures were probably also due to anaerobes as most sterile empyemas are thought to be due to anaerobic organisms [IO]. The pleural fluid in all cases of parapneumonic effusions should be carefully cultured for anaerobic bacteria. The pleural fluid in the majority of patients with parapneumonic effusions resolves without tube thoracostomy or repeated thoracentesis. We reported previously  that parapneumonic effusions, large enough on which to perform thoracentesis, resolved in 19 of 24 patients (79 percent) without tube thoracostomy or repeated thoracentesis. None of Fine’s 12 patients with nonbacterial pneumonia and pleural effusion  and only three of Taryle’s 20 patients with pneumococcal pneumonia and pleural effusion required tube thoracostomy . In the present series, 81 of the 90 parapneumonic effusions (90 percent) resolved without therapeutic thoracentesis or tube thoracostomy. The parapneumonic effusions in all the patients with small pleural effusions (pleural fluid less than 10 mm thick on lateral decubitus roentgenogram) and 28 of the 37 patients (76 percent] with larger effusions resolved without therapeutic thoracentesis or tube thoracostomy. Since there is a high incidence of parapneumonic effusions with bacterial pneumonia, but since most resolve without tube thoracostomy, what is the optimal approach to patients with parapneumonic effusions? Many different approaches have been recommended. Vianna  recommended the institution of closed tube drainage in all patients with parapneumonic effusion in which the pleural fluid white blood cell count was greater than 15,000/mm3 and the pleural fluid protein level was greater than 3.0 g/100 ml. Van De Water  recommended that patients with parapneumonic effusions undergo two therapeutic thoracentesis with complete drainage of the pleural fluid. If pleural fluid subsequently recurs, he recommended tube thoracostomy without regard for the characteristics of the pleural fluid. An American Thoracic Society Committee 
recommended intermittent closed drainage by repeated thoracentesis for pleural effusions in the exudative phase with the implementation of tube drainage when fluid reaccumulates rapidly. We have previously recommended  that tube thoracostomy be instituted when one or more of the following pleural fluid findings are present: (1) the presence of gross pus: (2) organisms visible on gram stain; (3) glucose level less than 50 mg/lOO ml; or (4) pleural fluid pH below 7.20. Since complicated parapneumonic effusions tend to become more and more loculated with time , it is clear that the earlier tube thoracostomy can be instituted, the easier it will be to drain the pleural space. Therefore, it is important to identify as early as possible those persons who will eventually require tube thoracostomy so that it can be instituted without delay. Although neither the pleural fluid protein level nor the white blood cell count helps to differentiate complicated from uncomplicated parapneumonic effusions, the pleural fluid pH, LDH level and glucose level all appear to be useful [Figure I). All patients with a pleural fluid pH below 7.00 or a glucose level below 40 mg/100 ml had complicated parapneumonic effusions. All patients with a pleural fluid pH above 7.2 or an LDH level below 1,000 mg/lOO ml had uncomplicated parapneumonic effusions. Patients with pleural fluid pH between 7.00 and 7.20 or LDH levels above 1,000 IU/l,OOO ml could have either complicated or uncomplicated parapneumanic effusions. Based on the present series, we recommend the following course of action for managing patients with parapneumonic effusions. The possibility of a parapneumonic effusion should be considered in all patients with bacterial pneumonia. If the posterior costophrenic angle is obliterated on the lateral chest roentgenogram, a decubitus chest x-ray film should be obtained. If the thickness of the pleural fluid on the decubitus film is greater than 10 mm, a diagnostic thoracentesis should be performed. Chest tubes should be placed immediately when the pleural fluid has any one of the following four characteristics: (I) is grossly pus; (2) shows organisms on gram stain; (3) has a glucose level below 40 mg/lOO ml; or (4) has a pH less than 7.00. If the pleural fluid LDH level is below 1,000 mg/lOO ml and if the pleural fluid pH is above 7.20, no further diagnostic or therapeutic maneuvers need be directed towards the pleura. The patients in the present series with pleural effusions who met these criteria did not have serial therapeutic thoracentesis. Their hospital courses were not prolonged, and they were not left with residual pleural disease. Therefore, we do not recommend serial therapeutic thoracentesis as have others [5,6]. We also do not increase the dose of antibiotics in these patients. Patients who have a pleural fluid pH between 7.00 and 7.20 or an LDH level above 1,000 mg/lOO ml present a special problem. Since some will need chest tubes whereas others will not, each case should be considered
Journal of Medicine
individually. If a patient has a very large effusion and the pH is close to 7.00, that patient should probably undergo tube thoracostomy. Alternatively, serial thoracenteses are indicated in some patients. If the pleural fluid pH and glucose level tend to rise and the LDH level tends to fall with successive thoracentesis, the patient probably will not require chest tubes as illustrated by the last four patients in Table III. Conversely, if the pH and glucose tend to decrease and the LDH to increase, tube thoracostomy should be instituted. A few words of caution should be made concerning the use of pleural fluid pH measurements: (1) The
pleural fluid must be collected anaerobically and placed on ice during its transfer to the blood gas laboratory; (2) the pH of the pleural fluid is influenced by the arterial pH (2). and a low pleural fluid pH indicates insertion of chest tubes only when the patient does not have systemic acidosis; (3) the use of the pleural fluid pH as a criteria for the placement of chest tubes is valid only with parapneumonic effusions. Patients with tuberculous, rheumatoid or malignant pleural effusions may also have low pleural fluid pHs; in these conditions, the placement of chest tubes is not influenced by the pleural fluid pH.
REFERENCES 1. Brewin A, Arango L, Hadley WK, Murray JF: High-dose penicillin therapy and pneumococcal pneumonia. JAMA 1974; 230: 409-413. 2. Light RW, MacGregor MI, Ball WC Jr. Luchsinger PC: Diagnostic significance of pleural fluid pH and Pcnz. Chest 1973; 64: 591-596. 3. Potts DE, Levin DC, Sahn SA: Pleural fluid pH in parapneumonic effusions. Chest 1976; 70: 328-331. 4. Andrews NC, Parker EF. Shaw RR, Wilson NJ, Webb WR: Management of nontuberculous empyema. Am Rev Respir Dis 1962; 85: 935-936. 5. Van De Water JM: The treatment of pleural effusion complicating pneumonia. Chest 1970; 57: 259-262. 6. Vianna NJ: Nontuberculous bacterial empyema in patients with and without underlying diseases. JAMA 1971; 215:
69-75. 7. Taryle DA, Potts DE, Sahn SA: The incidence and clinical correlates of parapneumonic effusions in pneumococcal pneumonia. Chest 1978; 74: 170-173. 8. Bartlett JG, Finegold SM: Anaerobic infections of the lung and nleural snace. Am Rev Resnir Dis 1974: 110: 56-77. 9. Fine NL, StnithLR, Sheedy PF: Fiequency of pleural effusions in mvconlasma and viral pneumonias. N Engl- 1. Med 1970;283:-796-793. 10. Bartlett JG, Gorbach SL, Thadepalli H, Finegold SM: Bacterioloev of emovema. Lancet 1974: 1: 338-340. 11. Snider Gc, Saleh I%: Empyema of the thorax in adults: review of 105 cases. Dis Chest 1968; 54: 410-415. 12. Light, RW: Management of parapneumonic effusions. Chest 1978;70:325-326.