CCL17 Is a Useful Biomarker for Discriminating Acute Eosinophilic Pneumonia From Other Causes of Acute Lung Injury

CCL17 Is a Useful Biomarker for Discriminating Acute Eosinophilic Pneumonia From Other Causes of Acute Lung Injury

CHEST Original Research CRITICAL CARE MEDICINE Circulating Thymus- and ActivationRegulated Chemokine/CCL17 Is a Useful Biomarker for Discriminating ...

886KB Sizes 0 Downloads 0 Views

Recommend Documents

A Case of Acute Eosinophilic Pneumonia
We examined an 86-year-old man with acute respiratory failure. A chest roentgenogram showed diffuse reticular shadows. T

Flavored Cigar Smoking Induces Acute Eosinophilic Pneumonia
Two cases of acute eosinophilic pneumonia (AEP) following smoking of flavored cigars were analyzed for characteristic fe

Sulfasalazine: A rare cause of acute eosinophilic pneumonia
Sulfasalazine is a compound of 5-aminosalicylic acid (5-ASA) and sulfapyridine joined by an azo bond. It is a widely use

Acute Lung Injury
Initiated by numerous factors, acute lung injury is marked by epithelial and endothelial cell perturbation and inflammat

DCoH: A novel biomarker for diagnosing acute kidney injury
Initial diagnosis of acute kidney injury (AKI) is usually based on measuring serum-creatinine and blood urea nitrogen le



Circulating Thymus- and ActivationRegulated Chemokine/CCL17 Is a Useful Biomarker for Discriminating Acute Eosinophilic Pneumonia From Other Causes of Acute Lung Injury* Eishi Miyazaki, MD; Shin-ichi Nureki, MD; Emiko Ono, MD; Masaru Ando, MD; Osamu Matsuno, MD; Tetsujiro Fukami, MD; Takuya Ueno, MD; and Toshihide Kumamoto, MD

Background: The presentation of acute eosinophilic pneumonia (AEP) closely resembles that of acute lung injury (ALI)/ARDS, including its idiopathic form, acute interstitial pneumonia (AIP). AEP usually lacks peripheral eosinophilia at the acute phase; therefore, the establishment of serum biomarkers for AEP would be clinically useful. Methods: We measured the levels of thymus- and activation-regulated chemokine (TARC)/CCL17, eotaxin/CCL11, KL-6, and surfactant protein-D (SP-D) in serum for patients with acute parenchymal lung diseases including AEP (n ⴝ 17), AIP (n ⴝ 13), pneumonia-associated ALI/ARDS (n ⴝ 12), and alveolar hemorrhage (n ⴝ 7). To evaluate diagnostic ability, each marker was estimated by measuring the area under the receiver operating characteristic curve (AUC). Results: Serum TARC/CCL17 levels of AEP patients were much higher than those of patients in other disease groups. More importantly, high circulating TARC/CCL17 levels were observed in AEP even at acute phase when peripheral eosinophilia was absent. TARC/CCL17 showed the largest AUC, and the TARC/CCL17 levels with cutoff points from 6,259 to 7,039 pg/mL discriminated AEP from other syndromes with sensitivity and specificity of 100%. The KL-6 level was low in most patients with AEP, and the sensitivity was 81.6% in cutoff with 100% specificity. The AUC for eotaxin/CCL11 and SP-D was small, with values of 0.73 (95% confidence interval [CI], 0.60 to 0.86) and 0.53 (95% CI, 0.31 to 0.64), respectively. Conclusions: This study indicates that the measurement of circulating TARC/CCL17 and KL-6 is useful for discriminating AEP from other causes of ALI. (CHEST 2007; 131:1726–1734) Key words: acute eosinophilic pneumonia; acute lung injury; KL-6; serum biomarker; thymus- and activation-regulated chemokine Abbreviations: AEP ⫽ acute eosinophilic pneumonia; AHS ⫽ alveolar hemorrhage syndrome; AIP ⫽ acute interstitial pneumonia; ALI ⫽ acute lung injury; AUC ⫽ the area under the receiver operating characteristic curve; CI ⫽ confidence interval; p-ALI ⫽ pneumonia-associated ALI/ARDS; SP-D ⫽ surfactant protein-D; TARC ⫽ thymus- and activation-regulated chemokine

eosinophilic pneumonia (AEP) is characterA cute ized by an acute febrile illness with severe hypoxemia, diffuse pulmonary infiltrates, and an *From the Divisions of Pulmonary Diseases (Drs. Miyazaki, Nureki, Ono, Ando, Matsuno, Fukami, and Ueno) and Neurology and Neuromuscular Disorders (Dr. Kumamoto), Department of Brain and Nerve Science, Oita University Faculty of Medicine, Oita, Japan. All work was performed at Oita University Faculty of Medicine. The authors have no personal or financial conflicts of interest to disclose. Manuscript received October 25, 2006; revision accepted January 19, 2007. 1726

increase in BAL eosinophils.1,2 AEP may be a common pathway of lung inflammation in response to a variety of possible antigens such as cigarette smoke,3,4 dusts,2,5 fungi,6 and drugs,7 and it usually Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. org/misc/reprints.shtml). Correspondence to: Eishi Miyazaki, MD, PhD, Division of Pulmonary Disease, Department of Brain and Nerve Science, Oita University Faculty of Medicine, 1–1 Idaigaoka, Hasama-machi, Oita 879-5593, Japan; e-mail: [email protected] DOI: 10.1378/chest.06-2596 Original Research

demonstrates an excellent prognosis. The pulmonary presentation of AEP closely resembles that of acute lung injury (ALI) and its more severe form, ARDS and its idiopathic form, acute interstitial pneumonia (AIP).8,9,10 AEP should be distinguished from ALI/ ARDS and AIP because of its uniform prompt response to steroid therapy and uniformly good prognosis in AEP, as opposed to the low rate of steroid responsiveness and the high mortality rate in ARDS and AIP.11 A previous investigation12 defined the histologic pattern of AEP as eosinophilic infiltration with diffuse alveolar damage, which is shared by ARDS and AIP, and only massive eosinophil infiltration can distinguish both conditions. An analysis of the BAL fluid can help in distinguishing AEP from AIP and ALI/ARDS in that the latter is usually associated with neutrophils without a significant number of eosinophils.11 In contrast to the presence of conspicuous BAL eosinophils in AEP, blood eosinophilia is usually not prominent at presentation3; therefore, clinicians may sometimes misdiagnose AEP. Moreover, the BAL procedure is not always available at every hospital, while this procedure also sometimes may even worsen the patient’s condition. Hence, the establishment of diagnostically useful serum biomarkers for AEP would be of great value. Previous investigations have shown several molecules to be involved in AEP pathogenesis. In particular, some chemokines are considered to contribute to eosinophilic inflammation, which may be detectable in serum as well as BAL fluids. Eotaxin/CCL11, the most potent chemokine for eosinophils, was exclusively elevated in the BAL fluids of patients with eosinophilic pneumonia.13 Thymus- and activation-regulated chemokine (TARC)/CCL17, a functional ligand for CCR4, was also elevated in the BAL fluids only from patients with eosinophilic pneumonia among diffuse lung diseases.14,15 Serum levels of eotaxin/CCL11 and TARC/CCL17 have been shown to be elevated and associated with disease activity in other types of allergic diseases, such as atopic dermatitis or bronchial asthma, thus suggesting that these chemokines in the circulation may reflect in situ eosinophilic inflammation in these syndromes,16 –19 whereas KL-6 and surfactant protein-D (SP-D) are expressed on regenerated type II pneumocytes and moved into the bloodstream in patients with interstitial pneumonia20,21 as well as ALI/ ARDS.22,23 The elevation of KL-6 and SP-D may reflect alveolar epithelial cell damage or re-epithelialization in the pathogenesis. Although the histopathology of AEP has previously been demonstrated to be diffuse alveolar damage with eosinophilic infiltration,12 circulating KL-6 levels were not elevated in five cases in the publication by Daimon et al.24 The aim of this study was to establish diagnostically useful

biomarkers of AEP in order to discriminate this syndrome from other causes of ALI such as AIP, pneumonia-associated ALI/ARDS, and alveolar hemorrhage. Materials and Methods Patient Recruitment The diagnosis of AEP was established based on a modification of the criteria proposed by Philit et al,25 as follows: (1) acute febrile illness; (2) bilateral diffuse infiltrates on chest radiography; (3) hypoxemia with Pao2 on room air ⬍ 60 mm Hg, and/or oxygen saturation on room air ⬍ 90%; (4) lung eosinophilia, with ⬎ 25% eosinophils on BAL differential cell count; and (5) no evidence of infection. We excluded any case with an exacerbation of allergic bronchopulmonary aspergillosis by measuring the specific Ig levels against Aspergillus fumigatus. ALI and ARDS were diagnosed according to the North American-European Consensus Conference definition of ALI/ARDS including the acute onset of bilateral pulmonary infiltrates, a Pao2/fraction of inspired oxygen ratio ⬍ 300 mm Hg for ALI and ⬍ 200 for ARDS, and no evidence of left atrial hypertension.26 The exclusion criteria in this study were ALI/ARDS cases secondary to extrapulmonary origin such as sepsis, trauma, burns, or a postsurgical operation. Any patients having severe immunocompromised conditions were also excluded. AIP was diagnosed based on the clinical, radiologic and histopathologic findings as characterized by acute respiratory failure of unknown etiology with severe hypoxemia, diffuse lung infiltrates, and evidence of diffuse alveolar damage either at lung biopsy or at autopsy. The diagnosis of alveolar hemorrhage was made by harvesting bloody samples with BAL and then determining the presence of hemosiderinladen macrophages.27 BAL and Blood Sample Collection After informed consent was obtained from the subjects, BAL was performed principally to make a diagnosis. Briefly, a fiberoptic bronchoscope was wedged into the right middle lobe bronchus or into the left lingula. Saline solution was instilled in two to three aliquots of 50 mL, and then BAL fluid specimens from the subjects were collected. The cells were stained with May-Gru¨nwald-Giemsa solution, and a differential count was performed on 300 cells. Blood samples were obtained from clotted blood following centrifugation at 1,500g at 4°C for 10 min, and then were stored at ⫺ 80°C until the measurements were performed. Measurement of TARC/CCL17, Eotaxin/CCL11, KL-6, and SP-D The level of each marker was measured using commercially available specific kits according to the protocol of each manufacturer. The concentrations of TARC/CCL17 and eotaxin/CCL11 were measured using an enzyme-linked immunosorbent assay (R&D Systems; Minneapolis, MN). The coefficients of variation for the chemokines assays were within 10%. The upper and lower limits of detection for TARC/CCL17 were 2,000 pg/mL and 7 pg/mL, respectively. The detectable doses of eotaxin/CCL11 by the enzyme-linked immunosorbent assay kit were from 5 to 1,000 pg/mL. The levels of KL-6 were measured by a sandwich-type electrochemiluminescence immunoassay kit (Picolumi KL-6; Sanko CHEST / 131 / 6 / JUNE, 2007


Junyaku; Tokyo, Japan). The concentration of SP-D was measured by a sandwich-type enzyme immunoassay kit (Yamasa Shoyu; Tokyo, Japan). We used nondiluted serum for the detection of each molecule. When the values were beyond the upper limits of detection, we analyzed either 20-fold or 40-fold diluted samples. The people performing the assay were blinded to the diagnosis.

pletely resolved within 1 week, and all 17 patients were discharged without any symptoms from the hospital within a month. We also enrolled 32 patients with ALI due to diffuse parenchymal lung diseases other than AEP, which included 13 patients with AIP, 12 patients with pneumonia- associated ALI/ARDS (p-ALI), and 7 patients with alveolar hemorrhage syndrome (AHS). Four AIP cases were diagnosed without histologic confirmation based on clinical, radiologic, and the BAL findings, which demonstrated prominent neutrophils and only a few eosinophils. The AHS group consisted of two patients with allergic granulomatous angiitis as an underlining cause, who had current symptoms of bronchial asthma.

Statistical Analysis The results are presented as median values, with minimum and maximum values as the range. The Kruskal-Wallis test was used to compare the values of the different groups. In cases of a significant difference between groups, intergroup comparisons were assessed by nonparametric methods using the MannWhitney U test. The diagnostic ability of each marker, the power to discriminate between patients with AEP and patients with other types of ALI was estimated by measuring the area under the receiver operating characteristic curve (AUC). Statistical analysis was performed using the statistical software (Statistical Package for Social Science for Windows; SPSS; Chicago, IL); p ⬍ 0.05 was considered statistically significant.

TARC/CCL17, Eotaxin/CCL11, KL-6, and SP-D Concentrations in Sera None of the patients with AEP received corticosteroid therapy before blood sample retrieval; however, three patients with AIP and two patients with p-ALI had already been treated with corticosteroids. The results of TARC/CCL17, eotaxin/CCL11, KL-6, and SP-D assays using the sera obtained when the patients visited us due to symptoms of respiratory distress are shown in Figure 1. The median value of serum TARC/CCL17 in healthy volunteers was 248 pg/mL (range, 106 to 752 pg/mL). Patients with AEP had extremely high concentrations of TARC/CCL17 in sera (median, 11,895 pg/mL; range, 7,040 to 42,094 pg/mL), which were significantly higher than those from healthy volunteers as well as patients with AIP (median, 92 pg/mL; range, 31 to 661 pg/mL), patients with p-ALI (median, 210 pg/mL; range, 8 to 6,258 pg/mL), and patients with AHS (median, 244 pg/mL; range, 122 to 1,961 pg/mL). The differences demonstrated statistical significance (p ⬍ 0.001). The median value of circulating eotaxin/CCL11 was 144 pg/mL (range, 58 to 325 pg/mL) in healthy volunteers, 118 pg/mL (range, 55 to 207 pg/mL) in

Results Patient Characteristics This study was based on a retrospective analysis of frozen samples stored from 1995 until 2005. The patient characteristics are shown in Table 1. Seventeen patients met the criteria for AEP in this study, 4 of whom were described in a previous study.14 Suspected causes of AEP were cigarette smoke in 12 patients, drugs in 2 patients, and unknown in 3 patients. Three patients with AEP had a history of bronchial asthma; however, no asthmatic symptoms had occurred in the individuals within the previous 10 years and at the time of presentation. All patients were hospitalized, but no patients received mechanical ventilation. Six patients required intervention with corticosteroids for a few days, while the remaining 11 patients all spontaneously recovered. The pulmonary infiltrates on chest images were com-

Table 1—Characteristics of the Study Population* Characteristics





Healthy Volunteers

Male/female gender, No. Age, yr Smoker Pao2/fraction of inspired oxygen ratio Peripheral eosinophil Peripheral eosinophil, /␮L BAL fluid eosinophil Spontaneous improvement Mortality

11/6 19 (14–53)† 94 272 (182–299) 3 (0–12) 234 (0–745) 65 (42–56)† 65 0

5/8 67 (51–79) 46 222 (126–288)‡ 2 (1–9) 182 (86–974) 2 (1–16) 0 85

9/3 68 (51–84) 75 258 (84–293) 2 (0–8) 234 (0–1103) 1 (0–9) 0 45

3/4 59 (41–73) 43 260 (153–290) 3 (1–36) 201 (92–3290) 2 (0–27) 0 28

9/9 35 (22–58) 50 3 (1–8) 138 (52–320) 1 (0–3)

*Data are presented as the median (range) or % unless otherwise indicated. †p ⬍ 0.01, compared with the other groups (Mann-Whitney U test). ‡p ⬍ 0.001, compared with the AEP group (Mann-Whitney U test). 1728

Original Research

Figure 1. Serum concentrations of TARC/CCL17 (top left, a), eotaxin/CCL11 (top right, b), KL-6 (bottom left, c), and SP-D (bottom right, d) obtained from healthy volunteers (HV) and patients with AEP, AIP, p-ALI/ARDS, and AHS. A short horizontal line represents the median value. *p ⬍ 0.001, compared with the other groups (Mann-Whitney U test); **p ⬍ 0.001, compared with the groups of AIP and p-ALI (Mann-Whitney U test).

patients with AEP, 79 pg/mL (range, 27 to 203 pg/mL) in patients with AIP, 79 pg/mL (range, 13 to 707 pg/mL) in patients with p-ALI, and 161 pg/mL (range, 60 to 214 pg/mL) in patients with AHS. There was no difference in eotaxin/CCL11 levels among the AEP group and the other groups.

The levels of circulating KL-6 in patients with AEP (median, 126 U/mL; range, 101 to 265 U/mL) were comparable to those of healthy volunteers (median, 140 U/mL; range, 107 to 343 U/mL). Patients with AIP, p-ALI, and AHS had elevated serum KL-6 levels (AIP: median, 1,599 U/mL; range, CHEST / 131 / 6 / JUNE, 2007


387 to 6,580 U/mL; p-ALI: median, 512 U/mL; range, 169 to 1,607 U/mL; and AHS: median, 288 U/mL; range, 129 to 719 U/mL) when compared with patients with AEP and healthy volunteers. The median circulating SP-D level in healthy volunteers was 62 U/mL (range, 32 to 91 U/mL). In comparison to the healthy volunteers, raised SP-D concentrations were obtained from patients with AEP (median, 342 U/mL; range, 96 to 855 U/mL), patients with AIP (median, 379 U/mL; range, 144 to 1,440 U/mL), patients with p-ALI (median, 320 U/mL; range, 28 to 842 U/mL), and patients with AHS (median, 131 U/mL; range, 86 to 314 U/mL). No differences were observed between the AEP group and the other disease groups. Circulating TARC/CCL17 Levels in Association With Time Course in Patients With AEP Although patients with AEP generally lack peripheral blood eosinophilia, when patients are observed over the course of hospitalization, most patients tend to show a mild-to-moderate increase in the number of peripheral blood eosinophils.3,25 The serum levels of TARC/CCL17 as well as the number of peripheral eosinophils may change according to the clinical course. We thus plotted all measurements of TARC/ CCL17 as well as KL-6 and peripheral eosinophil numbers after the onset of dyspnea in AEP patients (Fig 2). When the peripheral eosinophilia was measured within 3 days after the onset of dyspnea, only two patients showed eosinophilia ⬎ 350/␮L. The eosinophil numbers were increased with time. In contrast, the value of TARC/CCL17 was not associated with the presence of peripheral eosinophilia, and the circulating TARC/CCL17 levels were invariably high even when the measurements were done within 3 days after the onset. According to chronologic observations, TARC/CCL17 levels in the AEP patients remained elevated for at least 10 days after the onset of dyspnea. The KL-6 level time course disclosed that the KL-6 concentrations were low whenever measured. The elevated TARC/CCL17 levels at presentation returned to normal when the condition completely resolved (Fig 3). Circulating TARC/CCL17 levels were measured during a rechallenge test in two AEP patients after informed consent was obtained from each patient. The quick increase in the circulating TARC/CCL17 levels in response to certain stimuli was verified in the challenge tests in the two patients. Results were shown in Figure 4. Patient 1 was exposed to cigarette smoke, while patient 2 was challenged with causal agent, minocycline. The circulating TARC/CCL17 value reached beyond the cutoff level of 8,000 pg/mL at least within 16 h after the challenge. 1730

Receiver Operating Characteristic Curve Analysis Receiver operating characteristic curves were used to evaluate the discriminative value of serum TARC/ CCL17, eotaxin/CCL11, KL-6, and SP-D among AEP and other acute types of diffuse parenchymal lung disease. Serum TARC/CCL17 levels resulted in the largest AUC: TARC/CCL17, 1.00 (95% confidence interval [CI], 1.00 to 1.00); eotaxin/CCL11, 0.73 (95% CI, 0.60 to 0.86); KL-6, 0.97 (95% CI, 0.94 to 1.00); and SP-D, 0.53 (95% CI, 0.31 to 0.64) [Fig 5]. The cutoff levels were set as the closest point to 100% sensitivity and 100% specificity, and the TARC/CCL17 levels with cutoff points from 6,258 to 7,040 pg/mL discriminated AEP from other syndromes with sensitivity and specificity of 100%. The KL-6 level was low in most patients with AEP, and the sensitivity was 81.6% in cutoff with 100% specificity.

Discussion Our AEP patients met the clinical definitions of ALI/ARDS except for the presence of high percentages of eosinophils in BAL fluid. Unlike the patients with ALI/ARDS and AIP, which showed high mortality rates, all AEP patients rapidly and completely recovered either spontaneously or with corticosteroid administration. No study to date has attempted to seek biochemical markers in patients with AEP. Based on a receiver operating characteristic curve analysis, TARC/CCL17 showed the largest AUC and the TARC/CCL17 levels with cutoff points from 6,258 to 7,040 pg/mL discriminated between AEP and other acute types of diffuse lung diseases including AIP, pneumonia-associated ALI/ARDS, and AHS with 100% sensitivity and 100% specificity. This indicates that TARC/CCL17 is a useful diagnostic biomarker of AEP among the acute types of diffuse parenchymal lung disease. Patients with AEP generally lack peripheral blood eosinophilia, and only a minority of these patients have a blood eosinophil count ⬎ 350/␮L on presentation.25,28 In this series, only 2 of 17 patients had peripheral blood eosinophilia at presentation. Importantly, increased TARC/CCL17 was detected in the circulation of every patient even when peripheral blood eosinophilia was absent. An increase in the circulating TARC/CCL17 in patients with AEP is thus considered to indicate an increase in the chemokine expression at the inflammatory sites of the lungs. In fact, the circulating TARC/CCL17 levels were correlated with those of BAL fluids (data not shown). Regarding the cell sources of the circulating TARC/CCL17, we recently showed that TARC/ Original Research

Figure 2. Eosinophil numbers (top, a), and circulating TARC/CCL17 (center, b) and KL-6 levels (bottom, c) after the onset of dyspnea in patients with AEP. All measurements are plotted.

CCL17 is generated by alveolar dendritic cells and macrophages in AEP.29 Our AEP patients did not have apparent organ failures other than lung, such as liver dysfunction or skin eruption. Whether or not the peripheral blood mononuclear cells or platelets contributed to the elevation of serum TARC/CCL17

has yet to be determined in AEP. It also should be noted that the serum TARC/CCL17 concentration was found to increase and also be associated with the disease activity in patients with bronchial asthma19 in which TARC/CCL17 is highly expressed in the bronchial epithelium30; however, the TARC/CCL17 CHEST / 131 / 6 / JUNE, 2007


Figure 3. Serum TARC/CCL17 concentrations at the time when the condition was completely resolved (inactive) in comparison to those at admission (active) [p ⬍ 0.001].

values reported in bronchial asthma are much lower than those in our AEP patients.18,19 Our challenge tests documented that TARC/ CCL17 increased in the bloodstream within 16 h after provocation. In addition, chronologic observations indicate that TARC/CCL17 levels in the AEP patients remained elevated for at least 10 days after disease onset and thereafter decreased according to the reduction in symptoms. Therefore, TARC/ CCL17 may facilitate the assessment of the degree of disease activity in AEP. Unlike a previous article3 that demonstrated that two thirds of the patients require mechanical ventilation, our 17 AEP patients did not need any ventilatory support. We do not know what affects the apparent difference in severity between individuals from Western countries and those in Japan. Most AEP cases are associated with cigarette smoking in Japan, which might generate a clinically mild phenotype. In this regard, the performance of international comparative studies would thus be helpful and informative. Unexpectedly, the serum levels of eotaxin/CCL11 of patients with AEP were not higher than those of either normal volunteers or the other disease groups in this study. A previous study31 has shown that the BAL eotaxin/CCL11 levels increased in asthmatic subjects after an antigen challenge. In addition, an 1732

Figure 4. Results of the sequential measurements of serum TARC/CCL17 in the courses of a rechallenge test with either cigarette smoke (Case 1, top) or minocycline (Case 2, bottom) in patients with AEP.

increased concentration of eotaxin/CCL11 was detected in BAL fluids only from patients with eosinophilic pneumonia among various interstitial lung diseases.13 Therefore, eotaxin/CCL17 may play a substantial role in AEP as well as bronchial asthma; however, the circulating eotaxin/CCL11 does not appear to be diagnostically useful for discriminating AEP from other groups of diseases. KL-6 and SP-D have been reported to be useful biomarkers of interstitial lung diseases.20,21 It is thought that the increase of SP-D is due to the destruction of alveolar epithelium, whereas the elevation of serum KL-6 reflects the regeneration and proliferation of pneumocytes. High levels of these Original Research

Figure 5. The receiver operating characteristic curves of TARC/ CCL17, eotaxin/CCL11, KL-6, and SP-D concentrations in the serum for discriminating AEP from other acute types of parenchymal lung disease.

molecules, which reflect the severity of lung injury in patients with ALI/ARDS, are associated with poor clinical outcomes.22,23 In the present study, circulating KL-6 and SP-D levels increased in most patients with AIP and ALI/ARDS. A significant correlation has been reported between KL-6 and SP-D when investigating interstitial lung diseases such as idiopathic pulmonary fibrosis and collagen vascular disease-associated interstitial pneumonia.32,33 This was not true, however, in AEP patients in our series. In contrast to the increased serum SP-D levels in all AEP patients, none of AEP patients showed KL-6 elevation. This result supports the findings of a study24 that demonstrated no elevation of the KL-6 level in serum and BAL fluid of patients with AEP. In addition, this study disclosed that a low level of KL-6 would be diagnostically helpful marker for AEP among acute types of diffuse parenchymal lung disease. AIP is a syndrome that can be easily confused with AEP because both present as rapidly progressive respiratory failure without any obvious underlying causes. According to Allen et al,1 the diagnosis of AEP can never be made without BAL, which provides the presence of conspicuous BAL eosinophilia, a finding not seen in AIP. Similarly, our results showed a high level of circulating TARC/CCL17 to be seen only in AEP, not in AIP. This finding suggests the possibility that the presence of pulmonary eosinophilia can be recognized by measuring the circulating TARC/CCL17 concentration when BAL procedure cannot be performed. High-resolution CT scans have the ability to discriminate AEP from other diffuse lung diseases, such as AIP and pneumonia-associated ALI/ARDS.3,34,35 In contrast, a lower accuracy in the differential diagnosis of AEP and AHS using high-resolution CT was reported.35 The present study suggested that the

circulating TARC/CCL17 made it possible to distinguish AEP from AHS including allergic granulomatous angiitis, even though the number of AHS cases investigated in this study was small. In conclusion, we herein demonstrated the levels of TARC/CCL17 to be preferentially high in the serum of the patients with AEP. A low level of KL-6 was also evident in AEP. As a result, the TARC/ CCL17 and KL-6 values made it possible to discriminate AEP from other types of acute parenchymal lung disease such as AIP, pneumonia-associated ALI/ARDS, and AHS. However, each group in this study consisted of a relatively small number of cases. Future studies with a larger number of individuals are thus called for to confirm our findings.

References 1 Allen JN, Pacht ER, Gadek JE, et al. Acute eosinophilic pneumonia as a reversible cause of noninfectious respiratory failure. N Engl J Med 1989; 321:569 –574 2 Badesch DB, King TE, Schwarz MI. Acute eosinophilic pneumonia: a hypersensitivity phenomenon? Am Rev Respir Dis 1989; 139:249 –252 3 Allen J. Acute eosinophilic pneumonia. Semin Respir Crit Care Med 2006; 27:142–147 4 Nakajima M, Manabe T, Niki Y, et al. Cigarette smokeinduced acute eosinophilic pneumonia. Radiology 1998; 207: 715–719 5 Rom WN, Weiden M, Garcia R, et al. Acute eosinophilic pneumonia in a New York City firefighter exposed to World Trade Center dust. Am J Respir Crit Care Med 2002; 166:797– 800 6 Miyazaki E, Sugisaki K, Shigenaga T, et al. A case of acute eosinophilic pneumonia caused by inhalation of Trichosporon terrestre. Am J Respir Crit Care Med 1995; 151:541–543 7 Solomon J, Schwarz M. Drug-, toxin-, and radiation therapyinduced eosinophilic pneumonia. Semin Respir Crit Care Med 2006; 27:192–197 8 Shorr AF, Scoville SL, Cersovsky SB, et al. Acute eosinophilic pneumonia among US military personnel deployed in or near Iraq. JAMA 2004; 292:2997–3005 9 Giacoppe GN, Degler DA. Rapidly evolving adult respiratory distress syndrome with eosinophilia of unknown cause in previously healthy active duty soldiers at an Army training center: report of two cases. Mil Med 1999; 164:911–916 10 Weng TI, Yuan A, Tsai KC, et al. A patient of adult respiratory distress syndrome presenting initially with peripheral infiltration on chest radiographs and pulmonary eosinophilia. Am J Emerg Med 2001; 19:457– 460 11 Katzenstein AL, Myers JL, Mazur MT. A clinicopathologic, ultrastructual, and cell kinetic study. Am J Surg Pathol 1986; 10:256 –267 12 Tazelaar HD, Linz LJ, Colby TV, et al. Acute eosinophilic pneumonia: histopathologic findings in nine patients. Am J Respir Crit Care Med 1997; 155:296 –302 13 Katoh S, Matsumoto N, Fukushima K, et al. Elevated chemokine levels in bronchoalveolar lavage fluid of patients with eosinophilic pneumonia. J Allergy Clin Immunol 2000; 106: 730 –736 14 Miyazaki E, Nureki S, Fukami T, et al. Elevated levels of thymus- and activation-regulated chemokine in bronchoalveolar lavage fluid from patients with eosinophilic pneumoCHEST / 131 / 6 / JUNE, 2007


nia. Am J Respir Crit Care Med 2002; 165:1125–1131 15 Katoh S, Fukushima K, Matsumoto N, et al. Accumulation of CCR4-expressing CD4⫹ T cells and high concentration of its ligands (TARC and MDC) in bronchoalveolar lavage fluid of patients with eosinophilic pneumonia. Allergy 2003; 58:518 – 523 16 Jahnz-Rozyk K, Targowski T, Paluchowska E, et al. Serum thymus and activation-regulated chemokine, macrophagederived chemokine and eotaxin as markers of severity of atopic dermatitis. Allergy 2005; 60:685– 688 17 Kakinuma T, Nakamura K, Wakugawa M, et al. Thymus and activation-regulated chemokine in atopic dermatitis: serum thymus and activation-regulated chemokine level is closely related with disease activity. J Allergy Clin Immunol 2001; 107:535–541 18 Sekiya T, Yamada H, Yamaguchi M, et al. Increased levels of a TH2-type CC chemokine thymus- and activation-regulated chemokine (TARC) in serum and induced sputum of asthmatics. Allergy 2002; 57:173–177 19 Leung TF, Wong CK, Lam CWK, et al. Plasma TARC concentration may be a useful marker for asthmatic exacerbation in children. Eur Respir J 2003; 21:616 – 629 20 Kohno N, Awaya Y, Oyama T, et al. KL-6, a mucin-like glycoprotein, in bronchoalveolar lavage fluid from patients with interstitial lung disease. Am Rev Respir Dis 1993; 148:637– 642 21 Honda Y, Kuroki Y, Matsuura E, et al. Pulmonary surfactant protein D in sera and bronchoalveolar lavage fluids. Am J Respir Crit Care Med 1995; 152:1860 –1866 22 Greene KE, Wright JR, Steinberg KP, et al. Serial changes in surfactant-associated proteins in lung and serum before and after onset of ARDS. Am J Respir Crit Care Med 1999; 160:1843–1850 23 Ishizaka A, Matsuda T, Albertine KH, et al. Elevation of KL-6, a lung epithelial cell marker, in plasma and epithelial lining fluid in acute respiratory distress syndrome. Am J Physiol Lung Cell Mol Physiol 2004; 286:L1088 –1094 24 Daimon T, Tajima S, Oshikawa K, et al. KL-6 and surfactant proteins A and D in serum and bronchoalveolar lavage fluid in patients with acute eosinophilic pneumonia. Intern Med 2005; 44:811– 817 25 Philit F, Etienne-Mastroianni B, Parrot A, et al. Idiopathic










34 35

acute eosinophilic pneumonia: a study of 22 patients. Am J Respir Crit Care Med 2002; 166:1235–1239 Bernard GR, Artigas A, Brigham KL, et al. The AmericanEuropean Consensus Conference on ARDS: definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994; 149:818 – 824 De Lassence A, Fleury-Feith J, Escudier E, et al. Alveolar hemorrhage: diagnostic criteria and results in 194 immunocompromised hosts. Am J Respir Crit Care Med 1995; 151:157–163 Pope-Harman AL, Davis WB, Christoforidis AJ, et al. Acute eosinophilic pneumonia: a summary of fifteen cases and review of the literature. Medicine 1996; 75:334 –342 Nureki S, Miyazaki E, Ando M, et al. CC chemokine receptor 4 ligand production by bronchoalveolar lavage fluid cells in cigarette-smoke-associated acute eosinophilic pneumonia. Clin Immunol 2005; 116:83–93 Sekiya T, Miyamasu M, Imanishi M, et al. Inducible expression of a Th2-type CC chemokine thymus- and activationregulated chemokine by human bronchial epithelial cells. J Immunol 2000; 165:2205–2213 Lilly CM, Nakamura H, Belostotsky OI, et al. Eotaxin expression after segmental allergen challenge in subjects with atopic asthma. Am J Respir Crit Care Med 2001; 163:1669 – 1675 Ohnishi H, Yokoyama A, Kondo K, et al. Comparative study of KL-6, surfactant protein-A, surfactant protein-D, and monocyte chemoattractant protein-1 as serum markers for interstitial lung diseases. Am J Respir Crit Care Med 2002; 165:378 –381 Yokoyama A, Kohno N, Hamada H, et al. Circulating KL-6 predicts the outcome of rapidly progressive idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 1998; 158:1680 – 1684 King MA, Pope-Harman A, Allen JN, et al. Acute eosinophilic pneumonia: radiologic and clinical features. Radiology 1997; 203:715–719 Tomiyama N, Mu¨ller NL, Johkoh T, et al. Acute parenchymal lung disease in immunocompetent patients: diagnostic accuracy of high-resolution CT. Am J Roentgenol 2000; 174:1745– 1750

Original Research