Scandinavian Journal of Infectious Diseases, 2011; 43: 430435

ORIGINAL ARTICLE

Outcome and clinical characteristics in pleural empyema: A retrospective study

JANNIE NIELSEN1, CHRISTIAN N. MEYER1 & SIGNE ROSENLUND2

From the 1Department of Internal Medicine, Roskilde Hospital, Copenhagen University Hospital, Roskilde, and 2Department of Orthopaedic Surgery, Kge Hospital, Copenhagen University Hospital, Kge, Denmark

Abstract Background: Pleural empyema is a serious condition with a considerable mortality rate and morbidity. This study evaluated the correlations between several potential prognostic factors (age, predisposing diseases, early drainage, insufcient initial antimicrobial therapy, thoracic surgical treatment, intrapleural brinolysis, and nosocomial status) and outcome. Methods: Danish patients with positive pleural cultures attending 3 hospitals over a 9-y period, were identied in the laboratory databases. Clinical details and outcome were evaluated retrospectively by audit of the medical records. Results: We included 158 patients in this study. The overall mortality was 27% and the median length of stay was 29 days. Mortality correlated independently with several factors: nosocomial infection (odds ratio (OR) 2.62, 95% condence interval (CI) 1.714.16), predisposing conditions (OR 2.17, 95% CI 1.503.14), and also with the possibly interventional factors of sufcient initial antimicrobial therapy (OR 0.45, 95% CI 0.310.65), thoracic surgery treatment (OR 0.27, 95% CI 0.140.52) and local brinolytic therapy (OR 0.13, 95% CI 0.060.28). Delay in chest tube drainage more than 2 days was not independently correlated with mortality. The initial biochemical diagnostics of non-purulent pleural effusions (63%) did not follow the current international guidelines. Conclusion: Factors correlating independently with survival included the possible interventional parameters of brinolytic therapy, insufcient initial antimicrobial therapy, and having thoracic surgery treatment. Keywords: Pleural infections, empyema, outcome, treatment

Introduction Pleural empyema is a serious clinical condition reported to have a 36-times increased case-fatality rate compared to patients with non-infected parapneumonic effusions [1]. Several factors including predisposing diseases, delay in targeted antimicrobial or surgical treatment, the choice of treatment, and the bacterial aetiology may affect outcome. A short delay in the diagnosis and treatment of pleural empyema is associated with a superior outcome and a low mortality rate of 4.3 15% [2,3]. A difference in mortality of from 3.4% to 16% has been described with delay in chest tube drainage of more than 3 days, and the detrimental effect on outcome of a further 1 day delay in chest tube drainage has been shown in animal models [4,5]. There appears not to be consensus on whether the primary treatment ought to be handled directly

by thoracic surgery or in a collaboration between physicians of respiratory medicine and radiology [6]. Thus recent guidelines have concluded that a multidisciplinary approach is recommended [79]. Recent clinical studies in the developed countries have focused on correlating outcome to treatment parameters such as drainage technique and the use of local brinolytics [10,11], but few recent studies on other clinical parameters from the Scandinavian countries have been published [12]. In this retrospective study, we evaluated several potential prognostic factors (predisposing diseases, patient age, early drainage, thoracic surgery treatment, nosocomial status, insufcient initial antimicrobial therapy, intrapleural brinolytic treatment) for correlation with outcome in Danish patients with pleural empyema.

Correspondence: C. N. Meyer, Department of Internal Medicine, Roskilde Hospital, Kgevej 713, 4000 Roskilde, Denmark. Tel: 45 4732 2041. Fax: 45 4635 2093. E-mail: cnm@regionsjaelland.dk (Received 7 November 2010 ; accepted 9 February 2011) ISSN 0036-5548 print/ISSN 1651-1980 online 2011 Informa Healthcare DOI: 10.3109/00365548.2011.562527

Pleural empyema and outcome Materials and methods Study population Patients with a positive culture in pleural uids from the 3 Danish hospitals without in-house facilities to perform thoracic surgery (Copenhagen University Hospital Hvidovre, Copenhagen University Hospital Bispebjerg and Roskilde Hospital) were identied in the laboratory database of the Department of Clinical Microbiology at Hvidovre Hospital and in the Statens Serum Institute for the periods 19962004 and 19992007, respectively. Pre-admission symptoms, pre-admission therapy, predisposing conditions, post-admission clinical and laboratory ndings, medical and surgical treatment, outcome and the clinical relevance of the microbiological ndings obtained from the medical records were evaluated retrospectively. Inclusion criteria included the clinical diagnosis of pleural empyema with microbiological verication. Exclusion criteria included the evaluation of a positive culture being a contamination and not being treated as a pleural infection, or the nding of Mycobacterium tuberculosis in the pleural uid. Empyema was dened as purulent pleural uid or the presence of relevant positive culture in pleural uid. Delay of pleural drainage was dened as time from admission exceeding 2 days. A nosocomial infection was dened according to the US Centers for Disease Control and Prevention (CDC) [13]: if the patient had been hospitalized within 4 weeks of admission, if the empyema was a complication to invasive procedures, or if the empyema or the underlying infection clinically had begun later than 2 days after admission. Sufcient recovery was dened as patients returning to their recent health status. Insufcient recovery was dened as not returning completely to their recent health status. An unfavourable outcome was dened as insufcient recovery or death. Death was dened as in-hospital death. More detailed data on excluded cases in the cohort, the microbiological ndings, antimicrobial susceptibility and on the initial antimicrobial therapy is presented in a separate publication [14]. Statistics For univariate analyses the Chi-square test and Fishers exact test were used, when appropriate. For the outcome analyses, the above-mentioned potential prognostic factors were included in a multiple logistic regression analysis using a backward selection technique (SPSS for Windows, v. 6.1.2, 1995; EpiInfo, World Health Organization, Zurich, Switzerland). The level of signicance was set at p 0.05. Results Background characteristics

431

In the 9-y period, 158 patients with veried pleural empyema were included in this study. Details on the clinical background, predisposing diseases, bacterial aetiology and outcome are presented in Tables I and II. In 64% of all cases (n 101) the infection was community-acquired; 27% of cases were nosocomial (n 43) and 9% of cases (n 14) could not be categorized. All patients were subjected to thoracentesis by denition. According to currently guidelines, a few of the non-purulent pleural uid samples were analysed biochemically for white blood cell count (17%), protein (31%), lactate dehydrogenase (LDH) (13%), and glucose (7%); no samples were analysed for pH. Our patients had a median length of hospital stay of 29 days (range 1122), and the overall mortality was 27%. Median time to death among the study patients was 21 days (range 3146). Dyspnoea and cough were less frequently documented symptoms among the patients who later died (dyspnoea 47% vs 65%, cough 28% vs 56%, p 0.04, see Table I). Therapeutic factors Pleural drainage was done in 91% of the cases (n 144) and was guided by ultrasound in 63% (n 91). Pig-tail drains were used in 64% (n 92), large bore drains ( 20 French) in 6% (n 9), and no drain specication (most likely pig-tail or similar drain) was recorded for 30% (n 43). Therapeutic thoracentesis with no tubing was done in 9% of the cases (n 14). The median time from admission to drainage was 5 days (IQ-range 113 days), and delay in drainage 2 days did not correlate with mortality or with an unfavourable outcome (p 0.50 and p 0.95, respectively). Median duration of drainage was 8 days. No antimicrobial therapy was administrated in 5 cases. In 11 patients, the initial antimicrobial therapy could not be categorized as either sufcient or insufcient. Sufcient initial antimicrobial therapy (n 65) did not correlate signicantly with mortality in the univariate analysis (OR 0.56, 95% CI 0.251.22, see Table III). Local brinolytic therapy was performed in 24% (37/154) of the cases, and this group had a lower mortality than patients not treated with brinolytic agents (5% vs 34%, p 0.002). The median age was similar in the 2 groups (66 y vs 64 y, p 0.51), and the group treated with brinolytics had a similar rate of co-morbidity (69% vs 79%, p 0.21). By univariate analysis, brinolytic therapy correlated inversely with mortality (OR 0.11, 95% CI 0.010.47),

432

J. Nielsen et al.

Table I. Background data concerning bacterial aetiology, clinical features, admission data, and outcome in 158 patients with pleural infection (115 patients who survived and 43 who died). All patients Age (interquartile range) Male sex Bacterial aetiology Streptococci Mixed infection Staphylococcus aureus Enterobacteriaceae Anaerobic bacteria Other bacteria Reason for hospital admission Pneumonia Dyspnoea Cancer Abdominal illness Chest pain Lung disease (not pneumonia) Other diagnoses Background for the infectiona Pneumonia Thoracic trauma Recent thoracentesis Abdominal focus Recent surgery Oesophageal rupture Unregistered Antimicrobial therapy before admission Antimicrobial therapy before thoracentesis Symptoms at admission Dyspnoea Cough Chest pain Fever Pleural uid, purulent Pleural uid, cloudy Intensive care unit stay Length of stay, days (median, range)
aBy

Patients who died 66 61% 26% 33% 23% 16% 2% 0% 23% 7% 7% 14% 5% 5% 40% 33% 16% 7% 7% 4.6% 2% 30% 26% 72% 47% 28% 35% 30% 33% 19% 28% 27 (3146)

Patients who survived 62 67% 36% 23% 16% 6% 9% 10%

p-Value 0.03 0.45 0.03

63 (5377) 63% (n 99) 33% 26% 18% 9% 7% 7% 32% 10% 8% 8% 6% 6% 30% 59% 11% 9% 7% 3% 0.6% 15% 26% 66% 60% 48% 44% 37% 37% 27% 17% (n 27) 29 (3220)

0.37 36% 11% 8% 5% 5% 6% 29% 0.001 68% 9% 10% 7% 2.6% 0 3.5% 26% 64% 65% 56% 49% 40% 39% 23% 13% 29 (5220)

0.95 0.47 0.03 0.002 0.32 0.25 0.45 0.59 0.03 0.08

post-admission evaluation.

but not with unfavourable outcome (OR 1.21, 95% CI 0.522.86) (see Tables III and IV). Patients were admitted to a department of thoracic surgery in 34% of the cases (n 54), and 63% of these (n 34) had a surgical intervention performed: 85% decortication, 3% VATS (video-assisted thoracoscopic surgery) alone (without decortication) and 9% rib resection or thoracoplasty. By univariate analysis, mortality was lower in the group having surgery (12% vs 31%, OR 0.29, 95% CI 0.070.91, Tables III and IV), with a lower median age (61 y vs 69 y, p 0.01) and with a trend towards a lower rate of predisposing diseases (68% vs 80%, p 0.12). Multivariate analyses of outcome Apart from early drainage and the age parameter, the hypothesized prognostic factors correlated independently with mortality (see Table III). Nosocomial

infection (OR 2.62, 95% CI 1.714.16) and predisposing conditions (OR 2.17, 95% CI 1.503.14) correlated to mortality and are non-preventable and non-interventional parameters for the individual patient. In contrast, sufcient initial antimicrobial therapy (OR 0.45, 95% CI 0.310.65), intrapleural thrombolysis (OR 0.13, 95% CI 0.060.28) and thoracic surgery treatment (OR 0.27, 95% CI 0.140.52) were all factors associated with survival. In the analysis with unfavourable outcome as the dependent variable, a similar picture occurred (Table IV), though the correlation to the parameters intrapleural thrombolysis, nosocomial infection and sufcient initial antimicrobial therapy did not independently reach the level of statistical signicance (p 0.83, p 0.09 and p 0.06, respectively), with increasing age (OR 1.77, 95% CI 1.472.14) and presence of predisposing conditions (OR 3.64, 95% CI 2.305.76) as independent correlating parameters.

Pleural empyema and outcome

Table II. Occurrence of predisposing diseases and conditions in 158 patients with pleural empyema. No. of patients (%) Cancer Severe weakening in general Alcoholism ( 50 g ethanol/day) COPD Diabetes Rheumatologic disease Aspiration due to neurological disease HIV infection Other immune defects Other 37 (23) 29 (18) 24 (15) 20 (13) 18 (11) 10 (6) 5 (3) 4 (3) 3 (2) 35 (22)

433

COPD, chronic obstructive pulmonary disease; HIV, human immunodeciency virus.

Discussion In our population, pleural empyema was a serious and debilitating disease with 36% recovering fully, 37% with insufcient recovery, and 27% dying during hospital admission. Several of the hypothesized prognostic factors correlated independently with mortality, including predisposing conditions, nosocomial infection, insufcient initial antimicrobial therapy, intrapleural thrombolysis, and thoracic surgery. Another possible important interventional factor, delay in pleural drainage, did not correlate signicantly with mortality. Previous studies have examined the use of brinolytic therapy versus tube drainage alone in pleural empyema with opposing results concerning mortality or the need for later surgical intervention [1517]. Early small-scale (n 2450) randomized trials [1821] and observational or non-randomized studies [10] in adults showed the effects of intrapleural urokinase or streptokinase on drained volume and radiological improvement and showed a reduced need for surgical intervention. A recent larger scale randomized trial (n 454) could not show a treatment effect of intrapleural streptokinase [15], and it is still unclear whether and when it may be of benet to use brinolytics [8,22,23]. In the present

study, we found a signicant correlation between local brinolytic therapy and lower mortality. Several studies have focused on the potential benet of early surgery versus a more conservative treatment algorithm [6,24,25]. A recent study [26] conrmed the results of earlier randomized trials [25,27] that VATS or thoracotomy as initial treatment for advanced empyema is associated with better outcomes, and this may be preferred for those selected patients without serious medical problems if such surgical assistance is easily available [8,22,23,28]. To our surprise, 63% of all infected samples from pleural empyema patients were described macroscopically as non-purulent. In this population, only a minority of samples was further analysed biochemically, and remarkably no documentation was found for the measurement of pH in any sample. The diagnostic approach in the emergency departments and in the departments of internal medicine seldom followed the current British Thoracic Society guidelines [7] or the recent 2010 update [9], stressing the importance of initial diagnostic pH measurement in non-purulent effusions (in addition to cytology, culture, protein, LDH and glucose) to identify patients needing early chest tube drainage. Thus there was the potential for improvement in the diagnostic work-up at the hospitals concerned. The nding of a median delay of 2 days for thoracentesis and of 5 days until pleural drainage was set into effect may seem excessive. Few publications have described these therapeutic delays, but Lindstrom and Kolbe reported similar delays of 3.4 days and 5.2 days, respectively, in an Australian set-up [2]. Our nding that co-morbidity and increasing age (though the latter non-signicantly) correlated with an unfavourable outcome was expected and has been described by others [26], but appears clinically less interesting as these factors are not preventable or targets for intervention. The design of our study was retrospective and non-randomized, and the results will have limitations

Table III. Correlation of in-hospital death (n 43) to hypothesized factors by univariate and multivariate analysis, in patients with pleural infection (n 133; no antimicrobials for 5 patients, missing data for 20 patients). Univariate OR Intrapleural brinolysis Nosocomial infection Age (in quartiles) Sufcient initial antimicrobials Thoracic surgery performed Predisposing conditions
aComparing bBefore

Logistic regression 95% CI 0.010.47 1.387.50 0.914.36 0.251.22 0.070.91 1.4727.91 OR 0.13 2.62 1.36 0.45 0.27 2.17 95% CI 0.060.28 1.714.16 0.892.08b 0.310.65 0.140.52 1.503.14

0.11 3.21 1.99a 0.56 0.29 5.20

above the median to below the median. backward selection.

434

J. Nielsen et al. References

[1] Smith JA, Mullerworth MH, Westlake GW, Tatoulis J. Empyema thoracis: 14-year experience in a teaching center. Ann Thorac Surg 1991;51:3942. [2] Lindstrom ST, Kolbe J. Community acquired parapneumonic thoracic empyema: predictors of outcome. Respirology 1999;4:1739. [3] Davies CW, Kearney SE, Gleeson FV, Davies RJ. Predictors of outcome and long-term survival in patients with pleural infection. J Respir Crit Care Med 1999;160:16827. [4] Sasse S, Nguyen TK, Mulligan M, Wang NS, Mahutte CK. The effects of early chest tube placement on empyema resolution. Chest 1997;111:167983. [5] Ashbaugh DG. Empyema thoracis. Factors inuencing morbidity and mortality. Chest 1991;99:11625. [6] Huang HC, Chang HY, Chen CW, Lee CH, Hsiue TR. Predicting factors for outcome of tube thoracostomy in complicated parapneumonic effusion for empyema. Chest 1999; 115:7516. [7] Davies CW, Gleeson FV, Davies RJ. BTS guidelines for the management of pleural infection. Thorax 2003;58(Suppl 2): ii1828. [8] Colice GL, Curtis A, Deslauriers J, Heffner J, Light R, Littenberg B, et al. Medical and surgical treatment of parapneumonic effusions: an evidence-based guideline. Chest 2000;118:115871. [9] Davies HE, Davies RJ, Davies CW; BTS Pleural Disease Guideline Group. Management of pleural infection in adults: British Thoracic Society Pleural Disease Guideline 2010. Thorax 2010;65(Suppl 2):ii4153. [10] Frey DJ, Klapa J, Kaiser D. Irrigation drainage and brinolysis for treatment of parapneumonial pleural empyema. Pneumologie 1999;53:596604. [11] Sen S, Senturk E, Pabuscu E, Cokpinar S, Yaman E. [Minimally invasive approach in the pleural uids]. Tuberk Toraks 2010;58:717. [12] Heindorff H, Drenck NE. Pleural empyema. 95 cases treated at a thoracic surgery department. Ugeskr Laeger 1987;149: 2936. [13] Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM. CDC denitions for nosocomial infections, 1988. Am J Infect Control 1988;16:12840. [14] Meyer CN, Rosenlund S, Nielsen J, Friis-Mller A. Bacteriological aetiology and antimicrobial treatment of pleural empyema. Scand J Infect Dis 2011;43:1659. [15] Maskell NA, Davies CW, Nunn AJ, Hedley EL, Gleeson FV, Miller R, et al. U.K. Controlled trial of intrapleural streptokinase for pleural infection. N Engl J Med 2005;352:86574. [16] Cameron R, Davies HR. Intra-pleural brinolytic therapy versus conservative management in the treatment of parapneumonic effusions and empyema. Cochrane Database Syst Rev 2004;(2):CD002312. [17] Diacon AH, Koegelenberg CF, Bolliger CT. A trial of intrapleural streptokinase. N Engl J Med 2005;352:22435. [18] Bouros D, Schiza S, Tzanakis N, Chalkiadakis G, Drositis J, Siafakas N. Intrapleural urokinase versus normal saline in the treatment of complicated parapneumonic effusions and empyema. A randomized, double-blind study. Am J Respir Crit Care Med 1999;159:3742. [19] Davies RJ, Traill ZC, Gleeson FV. Randomised controlled trial of intrapleural streptokinase in community acquired pleural infection. Thorax 1997;52:41621. [20] Tuncozgur B, Ustunsoy H, Sivrikoz MC, Dikensoy O, Topal M, Sanli M, et al. Intrapleural urokinase in the management of parapneumonic empyema: a randomised controlled trial. Int J Clin Pract 2001;55:65860.

Table IV. Correlation of unfavourable outcome (n 101) to hypothesized factors by univariate and multivariate analysis, in patients with pleural infection (n 132; no antimicrobials for 5 patients, missing data for 21 patients). Univariate OR Age (in quartiles) Predisposing conditions Nosocomial infection Sufcient initial antimicrobials Intrapleural brinolysis Thoracic surgery performed
aComparing bBefore

Logistic regression OR 1.77 3.64 2.16 0.985 0.95 0.55 95% CI 1.472.14 2.305.76 1.363.44 0.980.99b 0.761.19b 0.340.90b

95% CI 1.134.91 1.7910.02 1.297.94 0.693.13 0.522.86 0.180.94

2.35a 4.21 3.16 1.47 1.21 0.41

above the median to below the median. backward selection.

concerning possible treatment bias. Thus selection bias cannot be ruled out concerning treatment with thoracic surgery or VATS. Only cases with a positive culture of relevant microorganisms were included in the present study [14]. Thus, excluding culture-negative cases, which have been said to account for up to 40% of all clinical cases, may also be a possible selection bias [7,9]. The diagnostic and therapeutic management was not uniform or standardized in the 3 centres in practice, which may not have been fully controlled for in the study analyses. In a number of the medical charts (50%), 1 or more data sets were missing out of the 22 parameters collected. The missing data in the medical records will reduce the statistical strength to a minor extent, as a small proportion of cases (13%, n 20) was not included in the logistic regression analyses. Thus, possibly skewed results (caused by the missing data) in the logistic regression analyses cannot be ruled out. Still, important results were gained concerning the possible background for the signicant ndings, and possible suboptimal routines in the clinical handling of patients were identied. In conclusion, this study found an independent correlation between survival and the possible interventional factors of sufcient initial antimicrobial therapy, intrapleural brinolytic therapy and thoracic surgical therapy. The non-preventable and noninterventional factors of nosocomial infection and predisposing conditions were correlated with mortality. We found that suboptimal diagnostic action may have affected the delay in treatment, which theoretically can inuence outcome. Future prospective and thus more complete studies may further clarify the role of these factors. Declaration of interest: The authors state that there are no conicts of interest in connection with this article.

Pleural empyema and outcome

[21] Diacon AH, Theron J, Schuurmans MM, Van de Wal BW, Bolliger CT. Intrapleural streptokinase for empyema and complicated parapneumonic effusions. Am J Respir Crit Care Med 2004;170:4953. [22] Light RW. Parapneumonic effusions and empyema. Proc Am Thorac Soc 2006;3:7580. [23] Molnar TF. Current surgical treatment of thoracic empyema in adults. Eur J Cardiothorac Surg 2007;32:42230. [24] Cassina PC, Hauser M, Hillejan L, Greschuchna D, Stamatis G. Video-assisted thoracoscopy in the treatment of pleural empyema: stage-based management and outcome. J Thorac Cardiovasc Surg 1999;117:2348.

435

[25] Lim TK, Chin NK. Empirical treatment with brinolysis and early surgery reduces the duration of hospitalization in pleural sepsis. Eur Respir J 1999;13:5148. [26] Wozniak CJ, Paull DE, Moezzi JE, Scott RP, Anstadt MP, York VV, et al. Choice of rst intervention is related to outcomes in the management of empyema. Ann Thorac Surg 2009;87:152530. [27] Wait MA, Sharma S, Hohn J, Dal NA. A randomized trial of empyema therapy. Chest 1997;111:154851. [28] Koegelenberg CF, Diaconi AH, Bolligeri CT. Parapneumonic pleural effusion and empyema. Respiration 2008;75: 24150.

Copyright of Scandinavian Journal of Infectious Diseases is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.