Usefulness of Carcinoembryonic Antigen in Pleural Effusion Diagnosis

REFERENCES

REFERENCES

1 Miller AC, Harvey JE. Guidelines for the management of spontaneous pneumothorax. BMJ 1993; 307:114-16 2 Baumann MH, Strange C. Treatment of spontaneous pneumothorax: a more aggressive approach? Chest 1997; 112:789804 3 Schramel FMNH, Sutedja TG, Braber JCE, et al. Costeffectiveness of \~cleo-assisted thoracoscopic surgery versus conservative treatment for first time or recurrent spontaneous pneumothorax. Eur Respir J 1996; 9:1821-25 4 Miller AC. Management of spontaneous pneumothorax: back to the future . Eur Respir J 1996; 9:1773-74 5 Sharma TN, Ahnihotri SP, Jain NK, et al. Intercostal tube thoracostomy in pneumothorax. Indian J Chest Dis Allied Sci 1988; 30:32-35

1 Wang PH. When should ACE inhibitors be given to normotensive patients with IDDM? Lancet 1997; 349:1782-83 2 Tomaki M, Ichinose M, Miura M, e t al. Angiotensin converting enzyme (ACE ) inhibitor-induced cough and substance P. Thorax 1996; 51:199-201 3 Nakagawa T, Ohrui T, Sekizawa K, et al. Sputum substance P in aspiration pneumonia. Lancet 1995; 345:1447 4 Ebihara T, Sekizawa K, Nakazawa H , et al. Capsaicin and swallowing reflex. Lancet 1993; 341:432 5 Nishino T, Takizawa K, Yokokawa N, et al. Depression of the swallowing reflex during sedation and/or relative analgesia produced by inhalation of 50% nitrous oxide in oxygen. Anesthesiology 1987; 67:995-98

ACE Inhibitor and Swallowing Reflex To the Editor:

Usefulness of Carcinoembryonic Antigen in Pleural Effusion Diagnosis To the Editor:

Angiotensin-converting enzyme (ACE) inhibitors have been shown to have beneficial effects such as antihypertension, vasodilator in congestive heart failure, and renoprotection. 1 ACE inhibitors are also reported to increase substance P (SP) concentration in sputum in patien ts wi th hype rtension,2 and a reduction of SP concentrations in sputum is observed in patients with aspiration pne umonia. 3 Since SP might stimulate the swallowing reflex,< we have investigated whether an ACE inhibitor improves the swallowing re flex in elderly patients with aspiration pneumonia. Twenty-two normotensive patients, mean age 75 (SE, ±2) years had at least one episode of aspiration with chest radiographic evidence of inflammation in the lower pulmonary segments. The 10 control subjects, mean age 75 (±3) years, were healthy volunteers. The swallowing reflex was induced by a bolus injection of 1 mL of distilled water into the pharynx through a nasal catheter. Swallowing was identified by submental electromyographic acti~ty. 4 · 5 The swallo,~ng reflex was evaluated by the latency of response, which was timed from the injection to the onset of swallowing 4 ·s In a randomized, double-blind, crossover design, the subjects were allocated a 5-mg tablet of imidapril or placebo daily for 2 weeks. The swallowing refl ex was measured three times at an interval of5 min after the treatment period, and an average value was used for analysis. The latency of response did not differ between placebo and imidapril in the controls (1.2 [± 0.1] vs 1.4 [±0.2] s; p > 0.50). However, imidapril significantly improved the latency of response compared '~th placebo in patients (2.7 [±0.3] vs 6.3 [±l.l] s, p<0.002). In conclusion, our results suggest that ACE inhibitors have beneficial effects on the impaired swallowing reflex in patients \~th aspiration pneumonia and help to prevent aspiration pneumonia in these patients. Katsutoshi Nakayama, MD Kiyohisa Sekizatca, MD Hidetada Sasaki, MD, FCCP Department of Geriatric Medicine Tohoku University School of Medicine Sendai, Japan Reprint requests: Hidetada Sasaki, MD, FCCP, Dept of Geriatric Medicine, Tokohu Univenity School of Medicine, Aoba-Ku Seiryo-machi 1-1 , Sendai 980 Japan

We read with interest the report by Garcia-Pachon and colleagues (March 1997)! on the level of carcinoembryonic antigen (CEA) in nonmalignant effusions. The authors found that CEA level was elevated (> lO nglmL) in 17 of 182 pleural effusions (9%) O\~ng to benign diseases, especially in empyemas and in complicated parapneumonic effusions (12/17, 70%). The authors concluded that knm~ng the causes and characteristics of nonmalignant pleural effusions is useful in interpreting the results of elevated levels of CEA. We \~Sh to report the results of our prospective study that were presented in part at the 59th Annual International Scientific Assembly of the American College of Chest Physicians, Orlando, in October of 1993. We studied 341 consecutive patients. Pleural fluid and serum CEA concentrations were determined by immunofluorescent assay. The patients were di~ded into various groups on the basis of the final diagnosis, which rested on clinical, radiological, and laboratory findings ,2 ·3 and follow-up. In 15 of 341 patients (4.4% ), suffi cient fluid for the planned diagnosti c tests was not obtained. Of the remain ing 326 patients, 131 (40%) had malignancy, 73 (22%) had transudates, 86 (26%) had benign exudates, 11 (3%) had pa ramalignant effusions , and 2.5 (8%) had ple ural effusions of undetermined cause . The patients with ple ural effusions of undete rmin ed cause and those with paramalignant effusions were not included in our study. Of th e 131 malignant pleural effusions , 14 were due to lymphomas and were not included in the study. Of th e 25 pleural effusions of undetermined cause, 1 had a p leural fl uid CEA level of 64.8 ng/m L (in the follow-up no malignancy was found ). None of th e patients with paramalignant effusions and lymphomas had a p leural .fluid CEA level > 10 nglmL. The causes of the remaining 276 pleural effusions are reported in Table l. The group \\~th 159 nonmalignant pleural effusions comprised 93 men and 66 women, with an average age of 68± 19 years (range, 17 to 98 years). We selected a CEA cutoff level of 10 nglmL for the diagnosis of malignant pleural effusion. Pleural fluid CEA concentrations were significantly higher in cases of malignant pleural effusions than in those of benign pleural effusions (mean±SD, 202.2 ±655.2 [ranging from 0.10 to 4,970] vs 3.5± 18.4 [ranging from 0.10 to 229], p < 0.0001 ). Pleural fluid CEA concentrations > 10 ng/mL were obse1ved in 72 of 117 patients '~th malignant pleural effusions (62%). Among these, CHEST I 113 I 5 I MAY, 1998

1425

Table !-Etiology of the 276 Pleural Effusions* Effusions

No.

%

Transudates CHF Liver cirrhosis Nephrotic syndrome Exudates Malignant Parapneumonic Empyema Tuberculosis Traumatic Nonspecific pleuritis Pulmonary em bolism Rheumatoid arthriti s Other exudates

73 .511

(26) 21 4 l (74) 42 14 4 4 3 3

1.3 2

203 117 40 ll 1l 8 11 2 2 4

L 2

*Total percentages are reported in parentheses.

CEA levels were elevated in 49 of 70 patients \\~th adenocarcinomas (70%), in compatison with 23 of 47 patients with other histologic types (49%). Pleural fluid CEA concentrations > 10 ng/m L were observed in 5 of 159 patients with benign pleural effusion s (3%). Four of them (80%) had empyemas and one had pulmonary embolism. The serum CEA levels in these patients were, respectively, ].5, 1.6, 1, and 2.5 ng/mL. Sensitivity of CEA for the diagnosis of malignant pleural effusions was 62%, specificity was 97%, positive pred ictive value was 94%, negative predictive value was 77%, and diagnostic accuracy was 82%. Our results agree with those reported by Garcia-Pachon and colleagues, though there are some differences. First, in our seties, sensitivity and specificity for the diagnosis of malignant pleural effu sion are higher than those observed by the authors (62% vs 47% and 97% vs 91%, respectively). Second, empyemas had elevated pleural fluid concentrations of CEA, but parapneumonic effusion s did not. The mechanism that causes the hi gh levels of CEA in empyemas is unclear, and it seems due to some local factor, because CEA serum concentrations were normal in 75% of our patients. Third, since in our expe rience, high levels of CEA in pleural fluid are found in only 3% of nonmalignant pleural effusions , we think that the presence of high levels of plemal fluid CEA in a patient with pleural effusion not clue to empyema is very suggestive of malignan cy. To conclude, we think that in order to inte'11ret causes of pleural effusions , we must consider that high levels of pleural fluid CEA rarely are found in benign effusions , particularly so in empyemas.

Valeriano Foresti, MD, FCCP Nadia Scolari, MD Ill Medical Department Antonio Villa, MD Emergency Department Fatebenefmtelli Hospital Milan, Italy REFERENCES Garcia-Pachon E, Padilla-Navas l , Dosda D , et a!. Elevated level of carcinoemhryonic antigen in nonmalignant pleural effusions. C hest 1997; 111:643-47 2 Foresti V, Villa A, Terenzio C, et a!. Thoracen tesis: clinical

1426

value and complications. Rassegna di Patologia deii'Apparato Respiratorio 1991 ; 6:123-27 3 Collins TR, Sahn SA. Thoracocentesis: clinical value, complications, technical proble ms, ancl patient experience . Chest 19R7; 91 :Rl7-22

To tlue Editor: \Ve wish to thank Foresti and colleagues for their interest in our article (March 1997). 1 They present th ei r experience \\~th the use of carcinoemhryonic antigen (CEA) in pleural effusion diagnosis and conclude the following: (l) that, contrary to our findings , false positive results are almost excl usively found in empyemas, ancl (2) that the sensitivity and specificity of CEA are slightly highe r than those we obtained. We think that these differences may he due to methodological reasons. In our study, both e mpyemas and complicated parapneumonic e ffusions were the le ading causes of false-positive CEA tests. According to the scheme of Light, 2 we defined empyema as th e presence of frank pus in the plenral space; nonpurul ent effusions with positive bacterial cultnres were considered to be "complicated parapneumonic e ffusions. " We do not know if Foresti and colleagues apply the same criteria, which could obviously influence th e results. With respect to the diffe rence in sensitivity and specificity of CEA, it is clearly due to the diffe rent composition of cliagnostic groups. Foresti and colleagues did not include malignant effusions due to lymphoma and excluded one patient with a benign pleurisy of undetermined origin and elevated CEA level. IF these cases are included, and if we exclude from our series paramalignant ple ural effusions in order to have the same diagnostic groups, th en we find that differences between both stuclies (Table 1) are not statistically significant (x 2 test).

Table !-Sensitivity and Specificity of CEA in Pleural Fluid* Source Gareia-Pachon et Foresti et al

al 1

Sensith~ty

Specificity

0.48 (0.34-0.60) 0.55 (046-0.63)

0.91 (0 11.5-0 95) 0.96 (O 92-0.911)

*Ninety five percent confirlence interval is shown in parentheses.

Eduardo Garcia-Pachon, MD Hospital Vega Baja Orilwela-Alicante, Spain Isabel Pndilla-Navas, MD Hospital General Universitari d'Elx Alicante, Spain REFERENCES Garcia-Pachon E, Padilla- Navas I , Dosda MD, eta!. Elevated level of carcinoembryonic antigen in nonmalignant pleural effusions. Chest 1997; 111:643-47 2 Light RW. A new classification of parapneumoni c effu sions and empyema. Chest 1995; 108:299-31

Pleural Fluid pH Determination To the Editor: vVe read with intPrest the article hv Lesho and Roth (Novembe r 1997) 1 that addressed the unreli;bility of litmus paper as an Communications to the Editor