ORIGINAL ARTICLE

ACTA RADIOLOGICA

Pneumothorax and the Value of Chest Radiography after Ultrasound-Guided Thoracocentesis

KI, S. MARJELUND & T. TIKKAKOSKI K. PIHLAJAMAA, M. K. BODE, T. PUUMALAINEN, A. LEHTIMA

Department of Radiology, Oulu University, Oulu, Finland; Department of Radiology, Central Hospital, Kokkola, Finland Pihlajamaa K, Bode MK, Puumalainen T, Lehtima ki A, Marjelund S, Tikkakoski T. Pneumothorax and the value of chest radiography after ultrasound-guided thoracocentesis. Acta Radiol 2004;45:828832. Purpose: To determine the incidence, the operators experience, and other variables that may inuence the development of pneumothorax or re-expansion edema after ultrasound (US)-guided thoracocentesis. Material and Methods: The medical records of 264 procedures in 212 patients who had undergone US-guided thoracocentesis in our radiology department or intensive care unit during the period 19962001 were retrospectively reviewed. Results: Post-thoracocentesis pneumothorax occurred in 11 cases, the incidence being 4.2% (11/264). None of the pneumothoraces occurred in the 10 mechanically ventilated patients. All but one patient with pneumothorax were asymptomatic or had only minor symptoms. Chest tube drainage was needed in one patient with a large pneumothorax. No re-expansion edema was recorded, although 1500 ml or more pleural uid was aspirated in 29 patients. The operators experience had no effect on the complication rate. Needle size was the only signicant variable that contributed to the pneumothorax rate. Conclusion: US-guided thoracocentesis can be done equally as safely by residents as by senior radiologists. The safety and feasibility of the method are evident among mechanically ventilated intensive care patients. Our results do not support the routine use of post-thoracocentesis chest radiography. Key words: Fluid; pleura; pneumothorax; thoracocentesis; ultrasound K. Pihlajamaa, Oulu University, Department of Radiology, Kajaanintie 50, FIN-90220 Oulu, Finland (e-mail. kai.pihlajamaa@mnet.) Accepted for publication 1 August 2004

Pleural effusion in chest radiography is a common indication for diagnostic thoracocentesis. Patients with abundant pleural effusion may require a therapeutic thoracocentesis by a single puncture or a pleural catheter to relieve dyspnea. Thoracocentesis can be done with conventional techniques or under ultrasound (US) guidance. Pneumothorax is the most common complication of this procedure, occurring in 3% to 20% of patients who undergo thoracocentesis (1). Less common complications of thoracocentesis include bleeding in the pleural cavity, re-expansion pulmonary edema, and hepatic or splenic puncture. Chest radiography is routinely performed after thoracocentesis to evaluate the presence of pneumothorax (4, 8, 9). However, the need for routine chest radiography after thoracocentesis is questionable according to recent studies because of the low

incidence of pneumothorax; the number of cases that require subsequent intervention is also small (1, 3, 5, 6, 10). In previous reports, the premedication (8), the technique of thoracocentesis, and the experience of the operator have varied (4, 8), and the results have been inuenced by selection bias (9, 12). In addition, the literature on US-guided thoracocentesis in ventilated patients is scant (6, 9). In our department, both residents and senior radiologists have carried out US-guided thoracocentesis using a standardized technique without premedication. However, the contribution of operator experience has been considered in only one previous study (4). The aim of our current study was to assess the occurrence of pneumothorax and re-expansion edema and to evaluate the possible risk factors for pneumothorax after US-guided thoracocentesis among unselected general hospital patients.
DOI 10.1080/02841850410008270 # 2004 Taylor & Francis

US-Guided Thoracocentesis

829

Material and Methods The medical records of patients who underwent USguided thoracocentesis in Keski-Pohjanmaa Central Hospital during the period 19962001 were reviewed. This retrospective study covered 264 procedures on 212 patients. The study patients included 84 women and 128 men (mean age 68.7 years, range 1897 years). Post-procedure chest radiography was a criterion for inclusion. Despite our routine having included post-procedure chest X-ray, some patients may have been missed. Patients who underwent pleural biopsy were excluded. The procedures were performed by several on-duty radiology residents with less than 2 years experience in radiology or senior radiologists with 515 years experience of interventional US. In our hospital, residents perform their rst two or three US-guided (Powervision 6000; Toshiba, Tokyo, Japan) thoracocentesis procedures under the supervision of a senior radiologist and thereafter work independently. Of 245 procedures performed in the Department of Radiology and 19 in the intensive care unit, 10 were receiving mechanical ventilation. Thoracocentesis was done using either a 0.8 mm ne needle attached to a syringe or a 1.8 mm (OD) stiff needle attached to a tube, a 3-way stopcock, a syringe, and a collector bag (Pleurox1; Braun, Melsungen, Germany). A ne needle was used if the physician asked for diagnostic aspiration or if the amount of pleural uid was estimated to be small in US examination. All therapeutic drainages were done using a 1.8 mm (OD) stiff needle. Local anesthesia was used. The puncture was generally performed with the patient seated, entering the pleural cavity through a dorsal intercostal space using a freehand technique. In the case of critically ill patients, the puncture was done laterally, with the patient in a supine position, or, posteriorly, with the patient in a lateral decubitus position. No premedication was used. We drained up to 1500 ml or even more if the patient was cooperative and tolerated the procedure. Our routine included postprocedure chest radiography taken 4 h after the intervention. A chest radiograph was taken in inspiration and the majority in the standing position. All the patients in the intensive care unit and some very ill patients were imaged in a lying position. We used the Fisher exact test and Pearsons chisquare test to dene the signicance of individual variables. For analyzing the interaction of uid volume and needle size we used the logistic regression model. The t test was used to compare the mean volumes of uid in pneumothorax

patients and the others. P values of less than 0.05 were considered signicant. SPSS 11.0 was used to conduct statistical analyses. Results Altogether 104 procedures were performed on women and 160 on men. Of the 43 (20%) patients who underwent multiple procedures, 37 needed 2 punctures, 5 needed 3 punctures and 1 needed 6 punctures. The volume of uid removed ranged from 2 to 1850 ml, the mean being 442 ml. In 29 patients, 1500 ml or more pleural uid was aspirated. A ne needle was used in 107 procedures and a large needle in 157, mean volumes being 26 ml (range 2220) and 874 ml (range 101850), respectively. The predominant causes of pleural uid are presented in Table 1. Post-thoracocentesis pneumothorax developed in 11 procedures, the incidence being 4.2% (11/264). All these procedures were carried out in the Department of Radiology. There were 10 patients (10/212) who had pneumothorax. One patient had pneumothorax twice, in his rst and third procedures (Table 2). No bleeding complications were recorded. The mean age of the patients with pneumothorax was 70 years. In the 10 mechanically ventilated patients, no pneumothoraces occurred. The mean volume of aspirated uid was 711 ml among the pneumothorax patients and 437 in the rest (P~0.281). Needle size was the only signicant variable contributing to the incidence of pneumothorax (Table 3). Interaction between amount of uid and needle size was not signicant. Odds ratio of thick needle was 10.2 when controlled for uid volume (Table 4). Large volumes of aspirated uid tended to be associated with higher frequency of pneumothoraces, but the difference was not statistically signicant (Table 3). Thus, pure diagnostic aspirations of small volumes with a ne needle proved safer than therapeutic drainages with a stiff needle. All but one patient with pneumothorax were
Table 1. Etiologies of pleural effusion Procedures (n / %) Benign serous uid Malignancy Pulmonary (n~25) Pleural (n~12) Other (n~17) Empyema Post-traumatic or postsurgery Unknown etiology Total 155/58,7 54/20.5

14/5, 3 12/4, 5 29/11.0 264/100

Acta Radiol 2004 (8)

830

K. Pihlajamaa et al.

Table 2. Characteristics of patients with pneumothorax Sex F M M1 F F F F2 M M M3

1 2 3

Age 64 72 97 80 72 89 72 21 70 66

Aspirated uid volume (ml) 200 1000 800/600 1500 1500 500 600 15 1100 10

Etiology of pleural effusion Empyema Benign serous uid Metastasis of pancreatic cancer Benign serous uid Benign serous uid Benign serous uid Benign serous uid Empyema Metastasis of urinary bladder carcinoma Unknown

The same patient had pneumothorax in two procedures. A chest tube was required. The patient had asthma.

Table 3. Effect of different variables on the occurrence of pneumothorax in 264 procedures carried out on 212 patients Occurrence of pneumothorax Total Needle Fine needle (0.8 mm) Thick needle (1.8 mm) Puncture site Right side Left side Mechanical ventilation No Yes Obstructive lung disease No Yes Performer Resident Senior Place of procedure Radiology department Intensive care unit Amount of uid removed (ml) 250 51500 5011850
1 2

n 1 10 9 2 11 10 1 4 7 11 2 2 7

% 0.9 6.4

Signicance 0.0261

107 157 140 124 254 10 222 42 137 127 245 19 104 49 111

0.0652 6.4 1.6 0.6481 4.3 0.4541 4.5 2.4 0.2291 2.9 5.5 0.4321 4.5 0.3003 1.9 4.1 6.3

One-sided signicance of the Fisher exact test. Two-sided signicance of the Fisher exact test. 3 Two-sided Pearson chi-square test.

Table 4. Logistic regression model with occurrence of pneumothorax as dependent variable and needle size and amount of uid as independent variables. Interaction between amount of uid and needle size was not signicant (OR~1.05, P~0.598) 95% CI for OR OR Amount of uid removed (ml) 1.0 Needle Fine needle (0.8 mm) 1.0 Thick needle (1.8 mm) 10.2 Lower Upper Signicance 1.0 1.0 0.514 0.050

asymptomatic or had only minor symptoms. Of the 11 pneumothoraces noted in chest radiography, 10 were small, and only 1 pneumothorax required chest tube placement. In the 10 mechanically ventilated patients, no pneumothoraces occurred. Discussion Our results show, for the rst time, that the procedure is easy to learn. The incidence of

1.0

103.0

Acta Radiol 2004 (8)

US-Guided Thoracocentesis

831

pneumothorax after US-guided thoracocentesis was 4.2% in our study (11/264). In our hospital, radiology residents perform their rst two or three procedures under the supervision of a senior radiologist. The operators experience had no inuence on the pneumothorax rate in our study, which is in line with the results of a recent study (4). In two previous studies, US-guided thoracocentesis yielded pneumothorax rates of 2.5% when the operator was an interventional radiologist and 3% when a resident radiologist, respectively (8, 11). BOLAND et al. reported an 8% pneumothorax rate in a selected series. However, the operators experience was not evaluated (2). Two previous studies have demonstrated that the incidence of pneumothorax can be markedly reduced by using US guidance (7, 11). In one study, however, US guidance did not reduce the risk for pneumothorax (10). The strength of the present ndings is that our results were not inuenced by selection bias, as all patients requiring thoracocentesis are referred to our department and no preselection of patients between residents and senior radiologist was done. Only one pneumothorax required a chest tube placement because of its large size, while no therapy was necessary in the other cases. Two previous studies revealed no correlation between the pneumothorax rate and the side of the procedure (1, 5) as in our study. GERVAIS et al. reported a 2.4% overall pneumothorax rate in their study, but a rate of 7% among mechanically ventilated patients (6). The increased risk of pneumothorax in association with mechanical ventilation was due to the different pressure conditions in the airways and the pleural cavity. Lung mobility is also greater in mechanical ventilation than in spontaneous breathing, which increases the risk of direct needle contact with the lung. Positive pressure in the airways during expiration allows continuous air leakage from the lung into the pleural space (6). Contrary to CERVAIS et al., and in line with LICHTENSTEIN et al., USguided thoracocentesis here proved a safe procedure for patients on mechanical ventilation (6, 9). The drawback of our study was that the number of ventilated patients was small. Needle size was the only signicant variable contributing to the pneumothorax rate. A ne needle was used in 41% of the procedures, but only one pneumothorax occurred after a ne needle puncture. We used either a 1.8 mm (OD) stiff needle or a ne needle in every procedure; we have no experience on plastic catheters in this context. RAPTOPOULOS et al. (11) also reported a signicantly

lower pneumothorax rate when a ne needle (1 mm) was used and also when abundant effusion was present. In our study, the volume of removed uid was greater among the pneumothorax patients (711 ml versus 437 ml, Pw0.05), which is in agreement with RAPTOPOULOS et al. (472 ml versus 246 ml). While on the one hand JONES et al. (8) recently reported that the incidence of pneumothorax requiring tube drainage increased when over 1100 ml of uid was removed, on the other hand the overall rate of pneumothoraces and other complications did not increase even when over 2000 ml of uid was removed. They also reported that premedication with atropine is unnecessary given the low incidence of vasovagal reactions. In our study, more than 1500 ml of uid was aspirated in 29 patients without an increased rate of complications. We did not use premedication. In a recent study, neither a clinical indication for thoracocentesis nor chronic obstructive pulmonary disease (COPD) was an independent predictor of pneumothorax (1). Our study concurs with this. Patients with malignant lung parenchymal disease who undergo therapeutic thoracocentesis will develop asymptomatic hydropneumothoraces due to poor lung compliance. BOLAND et al. recommended that routine drainage of asymptomatic pneumothoraces should be avoided in this subgroup of patients, as the pneumothorax will resolve spontaneously or will be replaced by re-accumulating effusion (2). Previous studies have not explained why pneumothorax occurs most frequently in patients undergoing multiple thoracocentesis (4). In our study, too, 43 patients underwent repeated procedures, 4 (9%) with pneumothoraces. In conclusion, US-guided thoracocentesis is occasionally, but rarely, complicated by pneumothorax, especially pneumothorax requiring chest tube placement. US-guided thoracocentesis is easy to learn and can be done equally as safely by residents as by senior radiologists. The safety and feasibility of the method were especially evident among mechanically ventilated intensive care patients. Our results do not advocate the routine use of post-thoracocentesis chest radiography.

Acknowledgement We acknowledge Marianne Haapea, M.Sc., for statistical analyses.

Acta Radiol 2004 (8)

832

K. Pihlajamaa et al. 7. Grogan D, Irwin R, Channick R, Raptopoulos V, Curley F, Bartter T, et al. Complications associated with thoracentesis. A prospective, randomized study comparing three different methods. Arch Intern Med ??;150:87377. 8. Jones P, Moyers J, Rogers J, Rodriguez R, Lee Y, Light R. Ultrasound-guided thoracentesis: is it a safer method? Chest 2003;123:41823. 9. Lichtenstein D, Hulot J-S, Rabiller A, Tostivint I, Mezie ` re G. Feasibility and safety of ultrasound-aided thoracentesis in mechanically ventilated patients. Int Care Med 1999;25:9558. 10. Petersen W, Zimmerman R. Limited utility of chest radiograph after thoracentesis. Chest 2000;117: 103842. 11. Raptopoulos V, Davis L, Lee G, Umali C, Lew R, Irwin R. Factors affecting the development of pneumothorax associated with thoracentesis. Am J Roentgenol 1991;156:91720. 12. Yu C-J, Yang P-C, Chang D-B, Luh K-T. Diagnostic and therapeutic use of chest sonography: value in critically ill patients. Am J Roentgenol 1992;159:695701.

References
o V, 1. Aleman C, Alegre J, Armadans L, Andreu J, Falc Recio J, et al. The value of chest roentgenography in the diagnosis of pneumothorax after thoracentesis. Am J Med 1999;107:3403. 2. Boland G, Gazelle G, Girard M, Mueller P. Asymptomatic hydropneumothorax after therapeutic thoracentesis for malignant pleural effusions. Am J Roentgenol 1998;170:9436. 3. Capizzi S, Prakash U. Chest roentgenography after outpatient thoracentesis. Mayo Clin Proc 1998;73:94850. 4. Colt H, Brewer N, Barbur E. Evaluation of patientrelated and procedure-related factors contributing to pneumothorax following toracentesis. Chest 1999;116:1348. 5. Doyle J, Hnatiuk O, Torrington K, Slade A, Howard R. Necessity of routine chest roentgenography after thoracentesis. Ann Intern Med 1996;124:81620. 6. Gervais D, Petersein A, Lee M, Hahn P, Saini S, Mueller P. US-guided thoracentesis: requirement for postprocedure chest radiography in patients who receive mechanical ventilation versus patients who breath spontaneously. Radiology 1997;204:5036.

Acta Radiol 2004 (8)