대한내과학회지 : 제 83 권제 4 호 2012 http://dx.doi.org/10.3904/kjm.2012.83.4.458 결핵성심낭염진단에있어심낭액림프구 - 포도당비 1 인산의료재단선린병원내과, 2 서울아산병원심장내과, 3 인제대학교해운대백병원심장내과 양경호 1 송윤석 1 정선호 1 김태진 3 윤신의 2 정용석 1 조길현 1 김용복 1 The Ratio Between the Percentage of Lymphocytes and Glucose Levels in Pericardial Fluid as a Method to Diagnose Tuberculous Pericarditis Kyung Ho Yang 1, Yoon Suk Song 1, Sun Ho Jung 1, Tae Jin Kim 3, Shin Eui Yoon 2, Yong Suk Jeong 1, Kil Hyun Cho 1, and Young Bok Kim 1 1 Department of Internal Medicine, Good Samaritan Hospital, Pohang; 2 Division of Cardiology, Department of Internal Medicine, Asan Medical Center, Seoul; 3 Division of Cardiology, Department of Internal Medicine, Haeundae Paik Hospital of Inje University, Busan, Korea Background/Aims: Adenosine deaminase (ADA) is a valuable biochemical marker for pericardial effusion (PE) and may be useful for diagnosing tuberculous pericarditis (TPE) in patients with PE. However, no definite cut-off or borderline values for ADA currently exist to distinguish TPE from other PE etiologies. In this study, we identified other useful parameters and characterized their relationship with ADA as a method for diagnosing TPE. Methods: From June 2004 to November 2011, 42 patients underwent pericardiocentesis due to moderate or severe PE, as confirmed by echocardiography or chest computed tomography (CT). Patients were subdivided into TPE and non-tpe (NTPE) groups. We analyzed ADA (p) (the pericardial ADA) and %Lymph (p)/glucose (p) (the ratio between the percentage of lymphocytes and glucose levels in PE). Results: We defined the cut-off value of ADA (p) as 48.5 IU/L, and that of %Lymph (p)/glucose (p) as 0.678% ㆍ dl/mg. In a multivariate logistic regression analysis, an odds ratio (OR) of 44.24 and a 95% confidence interval (CI) of 2.85-686.97 were observed in patients with an ADA (p) 48.5 IU/L (p = 0.023). An OR of 20.39 and a 95% CI of 1.06-392.93 were observed in patients with a %Lymph (p)/glucose (p) 0.678% ㆍ dl/mg (p = 0.046). The combination of ADA (p) and %Lymph (p)/glucose (p) had a higher positive predictive value (PPV, 80.0%) and specificity (Sp, 93.8%) than either ADA (p) (PPV, 47.4%; Sp, 68.8%) or %Lymph (p)/glucose (p) (PPV, 69.2%; Sp, 87.5%) alone. Conclusions: %Lymph (p)/glucose (p) is a useful parameter for distinguishing TPE from other pericardial diseases if combined with an ADA (p) 48.5 IU/L. (Korean J Med 2012;83:458-467) Keywords: Glucose; Lymphocyte; Tuberculous pericarditis Received: 2012. 3. 12 Revised: 2012. 4. 2 Accepted: 2012. 7. 17 Correspondence to Young Bok Kim, M.D. Department of Internal Medicine, Good Samaritan Hospital, 43 Daesin-ro, Buk-gu, Pohang 791-704, Korea Tel: +82-54-254-5124, Fax: +82-54-245-5345, E-mail: dulos2003@yahoo.co.kr Copyright c 2012 The Korean Association of Internal Medicine This is an Open Access article distributed under the terms of the Creative Commons Attribution - 458 - Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0/) which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
- Kyung Ho Yang, et al. The ratio between lymphocyte and glucose - 서론결핵성심낭염 (tuberculous pericarditis, TPE) 은우리나라에서흔히접할수있는질환으로국내보고에따르면심낭염중결핵성인경우의빈도가 20-51.2% 로국외보고보다상당히높은편이다 [1-4]. 진단은심낭천자에서얻은심낭액에서결핵세포진검사와심낭생검으로확진할수있지만, 이상의방법을통해확진하지못한경우심낭액의생화학적검사가결핵진단의보조적방법이될수있다. 현재유용한심낭액생화학적검사로 ADA (adenosine deaminase) 가널리이용되고있다. 이는검사방법이비교적용이하고높은진단적감수성과특이성을가지고있으며 [2,5,6], 또한여러연구에서결핵성심낭염의진단에유의한 ADA 제한치 (cut-off value) 도 40-60 IU/L으로보고하고있다 [1,6]. 그러나 ADA치는낮지만 (25.6 IU/L) 심낭생검에서만성육아종성병변이발견되어결핵성심낭염으로진단된경우도있고 [1], ADA 가 40 IU/L 으로결핵성심낭삼출로의심되었지만심낭중피종 (malignant pericardial mesothelioma) 이진단된경우도있어 [7], 전적으로 ADA 수치만으로결핵성심낭염을진단하는데는제한이있을것으로판단된다 [1,5]. 악성심낭액 (malignant effusion) 과갑상선기능저하증 (hypothyroidism) 에서단핵구 (monocyte) 백분율이, 류마티스심낭액과세균성심낭액에서는호중구백혈구 (neutrophil leukocyte) 백분율이높은것 [5] 과달리결핵성심낭염에서는심낭액림프구백분율이높았다 [8,9]. 흉막액 (pleural effusion) 의검사에서포도당수치가낮을수록결핵성흉막액가능성이높았다고보고하였다 [10]. 결핵성심낭염의심낭액의조사에서는포도당수치는통계학적유의성은없었으나, 다른군에비하여가장낮았다고보고되었다 [1]. 결핵성심낭염의네가지병리학적단계중초기에는과립백혈구증가증 (polymorphonuclear leukocytosis) 이발생되며 [8] 이에심낭액림프구백분율의진단적가치가감소될수있어, 림프구백분율단독보다는심낭액의다른기본적생화학검사 ( 예를들어포도당수치 ) 와의조합을고려할수있다. 이에저자들은심낭천자를시행하여심낭액을얻을수있었던환자를대상으로결핵성심낭염진단유무에따라결핵성심낭염군 (Tuberculous Pericarditis group) 과비결핵성심낭염군 (Non-tuberculous Pericarditis group) 두군으로분류하고, 결핵성심낭염진단에있어심낭액림프구백분율을포함한기본심낭액생화학 (basic pericardial biochemical) 검사지표 (parameter) 의진단가치유무를파악하고, 심낭액에서림프구백분율과포도당수치를조합한새로운지표를제시하고자한다. 대상및방법대상본연구는 2004년 6월 19일부터 2011년 11월 23일까지본원을내원하여임상증상, 흉부 X-선검사, 심전도소견등에서심낭액이의심되어시행한심장초음파검사또는흉부전산화단층촬영상에서중등도이상의심낭액소견을보이거나소량이라도호흡곤란악화되어심낭천자 (pericardiocentesis) 를시행한환자 49명중에서유효한심낭액의생화학적검사를시행하지않는 7명을제외하고, 42명의환자를대상으로의무기록을통해후향적으로조사하였다. 연구방법심낭액은심장초음파을이용하여흉골하연및심첨부에서심낭천자를시행하여얻었고심낭삼출의원인은 2004년 ESC (European Society of Cardiology) 지침에따라분류하였으며 [5], 결핵성심낭염의진단은심낭액또는그이외의검체 ( 예를들어흉수 ) 에서결핵균 (Mycobacterium tuberculosis, MTB) 의존재가입증 (Definite diagnosis) 되거나 [5,11], 원인불명이나임상적으로의심되어투여된항결핵제에지속적인임상호전이있는경우 (Probable diagnosis) 를결핵성심낭액으로진단하였다 [8,11-13]. 심낭액에서결핵균의존재가입증된경우가 2명, 경험적항결핵제치료에반응이있는경우가 8명이었다. 이에따라결핵성심낭염으로진단된군 (Tbc., Tuberculous Pericarditis group) 과진단되지않은군 (Non-Tbc., Non-Tuberculous Pericarditis group) 으로나누었다. 두군에서 ADA (p) ( 심낭액 ADA, 단위 : IU/L) 와함께 %Lymph (p) ( 심낭액림프구백분율 : 심낭액백혈구중림프구백분율 ), 백혈구, 적혈구, 포도당, 총단백질, 알부민, 젖산탈수소효소 (LDH, Lactate Dehydrogenase), 수소이온농도지수 (ph) 의기본적인생화학검사의진단적가치유무를평가하고, 이로부터도출된 %Lymph (p)/glucose (p) ( 심낭액림프구-포도당비 : 심낭액백혈구중림프구백분율을심낭액포도당수 - 459 -
- 대한내과학회지 : 제 83 권제 4 호통권제 626 호 2012 - 치로나눈값, 단위 : % ㆍdL/mg) 에대한통계적차이를분석하고감별진단에사용될수있는지표로이용될수있는지여부를검정하였다. 통계처리모든결과의통계처리는 SPSS (ver18.0k for Windows, SPSS Inc, Chicago, IL, USA) 를이용하여분석하였다. 연속변수는평균 ± 표준편차 (mean ± SD) 로, 범주형변수는백분율 (percent, %) 로표시하였다. 두군간의기본적특징, 심낭액생화학적검사결과를비교하여연속변수는 Mann-Whitney test, 범주형변수는 Chi-squared test를시행하였다. 표본의수가충분히많은경우 Pearson 카이제곱중에서점근유의확률 ( 양쪽검정 ) 값을선택하였고, 5보다작은기대빈도가있을경우에는 Fisher의 정확한검정 등의직접확률법을이용하였다 [14]. 두군간의심낭액기본생화학적검사는 ROC (receiver operation characteristics) 곡선을시행하여진단적가치가있는지표를추출해내어, 각각에대한제한치 (cut-off value) 를정하였고, 교차분석 (Crosstab) 을시행하였다. ADA (p) 에대한 %Lymph (p)/glucose (p), Glucose (p) 의단순상관분석 (Bivariate correlation analysis) 을실시하였다. 두군간의단순로지스틱회귀분석 (simple logistic regression analysis) 에서통계적으로유의한지표 (parameter) 인 ADA (p), %Lymph (p)/ Glucose (p), 악성과거력유무에대하여다중로지스틱회귀분석 (multiple logistic regression analysis) 을시행하여세가지변수가서로에게주는영향을감안하였다. 앞서언급하였던 ROC 곡선분석결과에따라 %Lymph (p)/glucose (p) 와 ADA (p) 제한치 (cut-off value) 를모두만족하는지유무에대한단순로지스틱회귀분석 (simple logistic regression analysis) 을시행하였다. %Lymph (p)/glucose (p), ADA (p), %Lymph (p)/glucose (p) & ADA (p) ( 상기두지표를병합하여적용 ) 변수에대한 Sp (Specificity, 특이도 ), Se (Sensitivity, 민감도 ), PPV (positive predictive value, 양성예측도 ), NPV (negative predictive value, 음성예측도 ) 를구하여서로비교하였다. p는 0.05 이하를통계적으로유의한것으로정의하였다. 결과임상양상및원인대상환자의연령분포는 14세에서 87세, 평균연령은 68.7 ± 15.6세였고성별은남자 23명 (54.8%) 과여자 19명 (45.2%) 이었다. 내원시환자의기저질환은표 1과같다 (Table 1). 결핵이진단된군 (Tbc.) 과결핵이진단되지않은군 (Non-Tbc.) 사이의기본특성비교에서악성종양과거력유무 (p = 0.031) 이외에나이, 성별, 연령, 과거질환, 사회력에따른통계학적으로의미있는차이는없었다 (Table 1). 심낭액의원인으로는전체 42명중악성종양 12명 (28.6%), 결핵성심낭염 10명 (23.8%), 원인미상 10명 (23.8%) 등이었다 (Table 2). 원인미상에대한추정질환은악성종양 3명 (30.0%), 급성바이러스성심낭염 1명 (10.0%), 방사선조사 1명 (10.0%), 외상 1명 (10.0%), 원인불명 4명 (40.0%) 이었다. 심낭액의생화학적분석흉수에서적용되는 Light s criteria로분류하면삼출액 (exudate) 을보인 40명 (95.2%), 여출액 (transudate) 2명 (4.8%) 이었다. 결핵이진단된군 (Tbc.) 에서는 %Lymph (p) 가 76.6 ± 18.9% 으로결핵성심낭염이진단되지않은군 (Non-Tbc.) 55.2 ± 25.5% 보다높았고 (p = 0.013), ADA (p) 도 96.3 ± 54.1 IU/L, 27.8 ± 22.0 IU/L (p < 0.001) 로결핵성심낭염으로확진된군 (Tbc.) 에서의미있게높았다 (Table 3). 악성종양과거력유무두군에대한단순로지스틱회귀분석 (OR 0.11; 95% CI 0.01-0.98; p = 0.048) 에서는통계적유의성이있었으나, 다중로지스틱회귀분석 (OR 0.10; 95% CI 0.00-2.27; p = 0.147) 에서는없었다 (Table 5). 폐결핵과거력유무두군에대한단순로지스틱분석결과통계적유의성이없었다 (p = 0.219), 폐결핵과거력유무에대한생화학검사의 Mann-Whiteny 검정결과 %Lymph (p) 에서만폐결핵과거력이있는경우 84.5 ± 9.5%, 폐결핵과거력이없는경우 57.8 ± 25.5% 로통계학적유의한차이를보였다 (p = 0.022). ADA (p) ADA (p) 의제한치 (cut-off value) 를 48.5 IU/L 로정하였을때 ROC curve 에서 AUC (area under curve) 는 0.89 (p < 0.001), 민감도 (Se) 는 90.0%, 특이도 (Sp) 는 87.5% 이었고 (Fig. 1), ADA (p) 48.5 IU/L 는두군간의단순로지스틱회귀분석 (OR 63.00; 95% CI 6.21-638.78; p < 0.001) 에서는통계적유의성이있었다. - 460 -
- 양경호외 7 인. 심낭액림프구 - 포도당비 - Table 1. Baseline characteristics between the two groups Etiology Subgroup Total (n = 42) Parameters Tbc (n = 10) Non-Tbc (n = 32) p Age, yr 68.7 ± 15.6 74.0 ± 5.6 67.0 ± 17.4 0.315 Sex, M/F 23/19 (54.8/45.2) 5/5 (50.0/50.0) 18/14 (56.3/43.8) 0.729 Underlying disease HTN 20 (47.6) 5 (50.0) 17 (53.1) 0.863 DM 10 (23.8) 0 (0.0) 10 (31.2) 0.084 Pul. Tbc 4 (9.5) 2 (20.0) 2 (6.3) 0.236 IHD 4 (9.5) 0 (0.0) 4 (12.5) 0.557 CHF 5 (11.9) 0 (0.0) 5 (15.6) 0.315 A. fib 6 (14.3) 3 (30.0) 3 (9.4) 0.135 Hypothyroidism 3 (7.1) 0 (0.0) 3 (9.4) 0.315 CKD 2 (4.8) 0 (0.0) 2 (6.3) > 0.999 Malignancy a 17 (40.5) 1 (10.0) 16 (50.0) 0.031 Social History Alcohol 9 (21.4) 0 (0.0) 9 (28.1) 0.086 Smoking 3 (7.1) 0 (0.0) 3 (9.4) > 0.999 Statistical analyses were performed between the tuberculous and non-tuberculous groups. Age was analyzed by the Mann-Whitney U-test. Other parameters were analyzed by chi-square tests. Data are mean ± standard deviation or n (%). Tbc., tuberculous group; Non-Tbc, non-tuberculous group; HTN, hypertension; DM, diabetes mellitus; Pul. Tbc, pulmonary tuberculosis; IHD, ischemic heart disease; A. fib, atrial fibrillation; CKD, chronic kidney disease; NS, not significant. a Malignancy had been diagnosed prior to the detection of pericardial effusion. Table 2. Cause of pericardial effusion Cause Number (%) Malignancy 12 (28.6) Tuberculous 10 (23.8) Heart failure 3(7.1) Renal failure 2(4.8) Hypothyroidism 2(4.8) Traumatic 1(2.4) Radiation 1(2.4) Autoimmune disease 1(2.4) Undetermined 10 (23.8) Total 42 (100.0) The expected frequency (p = 0.01) was analyzed by a chi-square test. Data are reported as n (%). %Lymph (p) ( 심낭액백혈구중림프구백분율 ) %Lymph (p) 의제한치 (cut-off value) 를 73.5% 로정하였을때 ROC curve 에서 AUC는 0.76 (p = 0.015), 민감도 (Se) 는 70.0%, 특이도 (Sp) 는 71.9% 이었다 (Fig. 1). 상기제한치에대한 %Lymph (p) 73.5% 는두군간의단순로지스틱회귀분석 (OR 5.96; 95% CI 1.26-28.28; p = 0.025) 에서는통계적유의성이있었으나, 다중로지스틱회귀분석 (OR 4.02; 95% CI 0.45-35.90; p = 0.212) 에서는없었다 (Table 4). Glucose (p) Glucose (p) 는결핵군 (Tbc.), 비결핵군 (Non-Tbc.) 각각에서 91.9 ± 35.5 mg/dl, 114.5 ± 50.1 mg/dl (p = 0.122) 로, ROC curve 분석결과 AUC (Area Under Curve) = 0.336, p = 0.121 로통계적인유의성이없었다. ADA (p) 에대한단순상관분석 ( 이변량상관계수, Person s correlation coefficient) 결과 r = -0.370, p = 0.016으로음적선형관계를보였다 (Fig. 2). %Lymph (p)/glucose (p) %Lymph (p)/glucose (p) 은두군간에서각각 1.6 ± 2.5%, 1.0 ± 2.7% (p = 0.006) 로통계적인유의성이있었고 (Table 1), ROC curve 분석결과 %Lymph (p)/glucose (p) 의제한치 - 461 -
- The Korean Journal of Medicine: Vol. 83, No. 4, 2012 - Table 3. Blood chemistry findings of patients with pericardial effusion Etiology Subgroup Total (n = 42) Parameters Tbc (n = 10) Non-Tbc (n = 32) p-value WBC (p), 10 3 /ul 3.2 ± 4.1 4.1 ± 3.3 3.0 ± 4.3 0.108 RBC (p), 10 3 /ul 696.8 ± 1441.9 816.9 ± 2316.4 659.3 ± 1089.0 0.716 %Lymph (p), % 60.3 ± 25.6 76.6 ± 18.9 55.2 ± 25.5 0.013 %Lymph (p)/glucose (p), % ㆍdL/mg 1.2 ± 2.6 1.6 ± 2.5 1.0 ± 2.7 0.006 Glucose (p), mg/dl 109.1 ± 47.7 91.9 ± 35.5 114.5 ± 50.1 0.122 ADA (p), IU/L 44.1 ± 43.4 96.3 ± 54.1 27.8 ± 22.0 0.000 TP (p) 5.1 ± 1.2 5.6 ± 0.8 5.0 ± 1.2 0.475 Albumin (p), g/dl 3.0 ± 0.6 3.0 ± 0.4 3.0 ± 0.6 0.440 LDH (p), IU/L 862.0 ± 886.0 1096.8 ± 1032.3 788.6 ± 839.9 0.122 ph 7.9 ± 0.2 7.9 ± 0.2 7.9 ± 0.2 0.975 Exudate/Transudate 40/2 (95.2/4.8) 10/0 (100.0/0.0) 30/2 (93.8/6.1) > 0.999 Statistical analyses were performed between the tuberculous and non-tuberculous groups. Exudate/transudate was analyzed by Fisher s exact test. Other parameters were analyzed by a Mann-Whitney U-test. Data are mean ± standard deviation or n (%). (p), pericardial fluid; %Lymph (p)/glucose (p), the ratio between the percentage of lymphocytes and glucose levels in pericardial fluid; %Lymph (p), percentage of lymphocytes in pericardial fluid; TP, total protein; Tbc, tuberculous group; Non-Tbc, non-tuberculous group; NS, not significant. Table 4. Logistic model coefficients for %Lymph (p) between the Tbc and Non-Tbc groups Univariate Analysis Multivariate Analysis OR 95% CI p OR 95% CI p %Lymph (p) 73.5% 5.96 1.26-28.28 0.025 4.02 0.45-35.90 0.212 Malignancy 0.11 0.01-0.98 0.048 0.22 0.01-3.34 0.272 ADA (p) 48.5 IU/L 63.00 6.21-638.78 < 0.001 35.52 3.37-417.72 0.003 The significance of the univariate or multivariate analysis was analyzed by simple or multiple logistic regression analyses, respectively. (p), pericardial; %Lymph (p)/glucose (p), the ratio between the percentage of lymphocytes and glucose levels in pericardial fluid; OR, odds ratio; 95% CI, 95% confidence interval; Tbc, tuberculous group; Non-Tbc, non-tuberculous group. Table 5. Logistic model coefficients for %Lymph (p)/glucose (p) between the Tbc and non-tbc groups Univariate Analysis Multivariate Analysis OR 95% CI p OR 95% CI p %Lymph (p)/glucose (p) 0.678% ㆍdL/mg 19.80 2.20-178.17 0.008 20.39 1.06-392.93 0.046 Malignancy 0.11 0.01-0.98 0.048 0.10 0.00-2.27 0.147 ADA (p) 48.5 IU/L 63.00 6.21-638.78 < 0.001 44.24 2.85-686.97 0.007 The significance of the univariate or multivariate analysis was analyzed by simple or multiple logistic regression analyses, respectively. (p), pericardial; %Lymph (p)/glucose (p), the ratio between the percentage of lymphocytes and glucose levels in pericardial fluid; OR, odds ratio; 95% CI, 95% confidence interval; Tbc, tuberculous group; Non-Tbc, non-tuberculous group. (cut-off value) 를 0.678% ㆍdL/mg 로정하였을때 AUC 0.79 (p = 0.006), Se 90.0%, Sp 71.9% 로통계적인유의성이있었다. 상기제한치에대한 %Lymph (p)/glucose (p) 0.678 % ㆍ dl/mg 는두군간의단순로지스틱회귀분석결과 OR 19.80, 95% CI 2.20-178.17, p = 0.008, 다중로지스틱회귀분석결과 OR 20.39, 95% CI 1.06-392.93, p = 0.046로통계적인유의성이있었다 (Table 5). 전체 42명을대상으로 ADA (p) 에대한 %Lymph (p)/glucose - 462 -
- Kyung Ho Yang, et al. The ratio between lymphocyte and glucose - Figure 1. The ROC curves for discriminating the Tbc from the Non-Tbc groups using cut-off levels for ADA (p), %Lymph (p)/glucose (p), and %Lymph (p). ADA (p), pericardial ADA; %Lymph (p)/glucose (p), ratio between the percentage of lymphocytes and glucose levels in pericardial fluid; ROC, receiver operating characteristic; AUC, area under the curve; Tbc, tuberculous group; Non-Tbc, non- tuberculous group. Figure 3. The correlation between %Lymph (p)/glucose (p) and ADA (p). (p), pericardial; %Lymph (p)/glucose (p), ratio between the percentage of lymphocytes and glucose levels in pericardial fluid. The correlation was analyzed by Pearson s correlation analysis (r = 0.412, p = 0.007). 양성예측도 (PPV, positive predictive value) 양성예측도 (PPV), 특이도 (Sp) 는 %Lymph (p)/glucose (p) 0.678% ㆍdL/mg 단독, ADA (p) 48.5 IU/L 단독, %Lymph (p)/glucose (p) 0.678% ㆍdL/mg과 ADA (p) 48.5 IU/L를병합하였을때각각 PPV 47.4% Sp 68.8%, PPV 69.2% Sp 87.5%, PPV 80.0% Sp 93.8% 였다 (Table 6). 고 찰 Figure 2. Correlation between Glucose (p) and ADA (p). (p), pericardial; Glucose (p), the level of glucose in the pericardial fluid. The correlation was analyzed by Pearson s correlation analysis (r = -0.370, p = 0.016). (p) 의단순상관분석결과상관계수 r = 0.412 (p = 0.007) 로양적선형관계 (positive correlation) 을보였다 (Fig. 3). ADA (p) 48.5 IU/L, %Lymph (p)/glucose (p) 0.678% ㆍ dl/mg 두지표를동시에만족하는지유무에대한교차분석결과 OR 60.00; 95% CI 7.28-494.61; p < 0.001로통계학적유의성이있었다. 결핵성심낭염 (tuberculous pericarditis, TPE) 은서구에서는 0-2% 로드문질환이나 [4], 우리나라에서는상대적으로흔한질환이다 [1-4]. 본원을방문하여결핵성심낭염으로진단하고치료한환자도 23.8% 로국내의다른보고들과비슷하였다. ESC guideline에따르면결핵성심낭염 (TPE) 이의심되는경우진단을위해항산성염색 (acid fast staining), 결핵균배양 (mycobacterial culture) 또는형광성장감지 (radiometric growth detection), ADA (adenosine deaminase), 인터페론감마 (interferongamma), 심낭리소자임 (pericardial lysozyme), 중합효소연쇄반응 (PCR) 분석등을할수있다고했다 [5,13]. Levy 등 [15] 은비결핵성심낭염감별을위하여바이러스및세균중합효소연쇄반응를시행하여총 106명의심낭염환자에서 7명의바이러스성심낭염, 4명의세균성심낭염을진단하였다. Park 등 [16] 은결핵성심낭염 (TPE) 의경우실제심낭삼출액 - 463 -
- 대한내과학회지 : 제 83 권제 4 호통권제 626 호 2012 - Table 6. Diagnostic properties of %Lymph (p)/glucose (p), ADA (p), and %Lymph (p)/glucose (p) & ADA (p) between the Tbc and Non-Tbc groups Sensitivity Specificity PPV NPV %Lymph (p)/glucose (p) 90.0% 68.8% 47.4% 95.7% ADA (p) 90.0% 87.5% 69.2% 96.6% %Lymph (p)/glucose (p) & ADA (p) 80.0% 93.8% 80.0% 93.8% (p), pericardial; %Lymph (p)/glucose (p), ratio between the percentage of lymphocytes and glucose levels in pericardial fluid; %Lymph (p)/glucose (p) & ADA (p), the combination of %Lymph (p)/glucose (p) and ADA (p); PPV, positive predictive value; NPV, negative predictive value; Tbc, tuberculous group; Non-Tbc, non-tuberculous group. 에서의결핵균배양률도낮아심막조직검사 (pericardial biopsy) 을이용하여진단율을높일수있는지를비교하였으나심낭액으로시행한다른실험실검사보다심막조직검사 (pericardial biopsy) 의우월성을언급하기어렵다고하였다. Kim 등 [1] 은도말검사와배양검사의민감도가낮은현실에서심낭액 ADA를기준으로한결핵성심낭염의진단은매우유용한방법으로사료된다고하였다 [6]. 저자들의경우도결핵성심낭염 (TPE) 진단에심낭액인터페론감마, 심낭리소자임 (pericardial lysozyme), 심막조직검사는시행하지못하였으나, 심낭액 ADA를진단에유용한지표로사용하였다. 결핵균핵산증폭검사 (TB-PCR) 는일부환자들에대하여시행되었고, 양성자는없었다. 바이러스및세균성심낭염감별을위한바이러스및세균중합효소연쇄반응은시행하지못하였으나, 세균배양검사는시행되었고양성자는없었다. Adenosine deaminase (ADA) 는퓨린대사 (purine metabolism) 에관여하는효소로주로 T임파구에서검출되며한연구에따르면결핵성심낭염 (TPE) 에약 10-303 IU/L 정도의농도를보인다고보고하였다 [11,13]. 연구에따라제한치는 40-60 IU/L 로다양하다 [1,6,13]. 본연구 42명을대상으로시행한 ROC 곡선결과심낭액 ADA 제한치는 48.5 IU/L 로민감도 90.0%, 특이도 87.5% 이며 p 값은 0.001 미만으로이전연구 [1,6,13] 결과들과유사하였다. 결핵성심낭염군 (Tbc.) 에서심낭액 ADA 의평균은 96.3 IU/L, 표준편차는 ± 54.1 IU/L 로 10-303 IU/L 를벗어나지않았다 [11,13]. 그러나결핵성심낭염 (TPE) 과관계없이특정질환 ( 예, 림프종 ) 에 ADA 상승이나타나기도하는데 [6], 이는 ADA치가 T-임파구의활성에의존하므로면역저하상태나면역자극상태에따라높거나낮게나올수있기때문이다 [1]. 이로인해전적으로심낭액 ADA치만으로결핵성심낭염 (TPE) 을진 단하는데는제한이있을것으로여겨진다 [1]. ADA 치에대한인간면역억제바이러스 (HIV, human immunodeficiency virus) 의영향에대해서논란이있지만 [12], 본연구에서이에대한고려를하지못하였다. Reuter 등 [12] 은결핵성심낭염군 (Tbc., n = 142) 과비결핵성심낭염군 (Non-Tbc., n = 61) 을비교하여결핵성심낭염군 (Tbc.) 에서상대적으로심낭단백질 (pericardial protein), 심낭림프구백분율 (percentage of lymphocyte in pericardial fluid) 이통계적으로유의하게높다고하였다. 본연구에서 Mann- Whitney test 결과결핵성심낭염군 (Tbc.) 에서상대적으로심낭액림프구백분율수치가통계학적으로높았으나 (p = 0.013), 두군간에심낭액단백질수치의차이는없었다 (p = 0.475) (Table 3). ROC 곡선시행결과심낭액림프구백분율의제한치는 73.5% 로민감도 70.0%, 특이도 71.8% 였다. 단순로지스틱회귀분석결과심낭액림프구백분율 73.5% (OR 5.96; 95% CI 1.26-28.28; p = 0.025) 는통계적유의한연관성이있었다. 다중로지스틱회귀분석결과심낭액림프구백분율 73.5% (OR 4.02; 95% CI 0.45-35.90; p = 0.212) 은통계적유의성이없었다 (Table 4). 흉수 (pleural effusion) 의화학적검사에서는흉수포도당수치가낮을수록원인이결핵가능성이높다고보고하였으나 [10], 심낭액의포도당수치에관해서는언급이없었다. 본원 42명을대상으로한심낭액포도당수치도 ROC 곡선, 교차분석에서통계적으로의미가없었다. 심낭액림프구백분율 (%) 을분자로, 심낭액포도당 (mg/dl) 을분모로한새로운지표 %Lymph (p)/glucose (p) ( 심낭액림프구 -포도당비, 심낭액백혈구중림프구백분율을심낭액포도당수치로나눈값 ; 단위, % ㆍdL/mg) 에대하여분석해보았다. 단순로지스틱회귀분석결과 %Lymph (p)/glucose (p) 0.678% ㆍdL/mg의 OR 19.80, p = 0.008로 (Table 5) - 464 -
- 양경호외 7 인. 심낭액림프구 - 포도당비 - %Lymph (p) 의 OR 5.96, p = 0.025보다상대적인통계적유의성이높았다 (Table 4). 심낭액 ADA (Table 3), 악성종양과거력유무 (Table 1) 를포함한다중로지스틱회귀분석결과 %Lymph (p)/glucose (p) 0.678% ㆍdL/mg (OR 20.39; 95% CI 1.06-392.93; p = 0.046), ADA (p) 48.5 IU/L (OR 44.24; 95% CI 2.85-686.97; p = 0.007) 두지표가통계학적유의하게진단적가치를가지는것으로나타났다 (Table 5). 실제로 ADA (p) 48.5 IU/L, %Lymph (p)/glucose (p) 0.678% ㆍ dl/mg 두지표를동시에만족하는지유무에대한교차분석결과통계적으로유의하였다. 양성예측도 (PPV), 특이도 (Sp) 도 %Lymph (p)/glucose (p) 0.678% ㆍdL/mg 단독, ADA (p) 48.5 IU/L 단독으로적용하였을때보다두지표를병합하여적용하였을때가장높았다 (Table 6). %Lymph (p)/glucose (p) 0.678% ㆍdL/mg 그리고 ADA (p) 48.5 IU/L 인환자 10 명중 8명이결핵성심낭염 (TPE) 으로 ADA (p) 48.5 IU/L 단독, %Lymph (p)/glucose (p) 0.678% ㆍdL/mg 단독인경우각각 13명중 9명, 19명중 9명보다결핵성심낭염 (TPE) 진단비율이높았다. 비결핵군 (Non-Tbc.) 32명에대하여 %Lymph (p)/glucose (p) < 0.678% ㆍdL/mg 또는 ADA (p) < 48.5 IU/L 인 30명이비결핵성심낭염 (NTPE, Non-Tuberculous Pericarditis) 으로 ADA (p) 48.5 IU/L 단독, %Lymph (p)/glucose (p) 0.678% ㆍdL/mg 단독인경우각각 28명, 22명보다비결핵성심낭염 (NTPE) 진단비율이높았다. 이와같이두기준을같이사용하였을때가단일기준 ( 심낭액 ADA) 일때보다검사양성시결핵성심낭염 (TPE) 가능성, 비결핵성심낭염시검사음성일가능성이높았다. 최근심낭액 LDH (lactate dehydrogenase) 와혈청 LDH과의비율 (P/S LDH ratio) 의진단적가치가언급되었으나 [17], 본원환자들의심낭액 LDH와혈청 LDH의측정시기가 24시간이상차이있는경우가많아분석을시행할수없었다. 그러나심낭액 LDH 수치와결핵진단유무에대한교차분석을시행하였다. 결과는 p = 0.118로통계학적의미는없었다. 본연구는본원을방문하여심낭천자술을받은경우로제한하였기때문에심낭액환자모두를대표할수는없다. 결핵성심낭염 (TPE) 의환자수가 10명으로통계학적의미가약해질수있다. 또한결핵성심낭염 (TPE) 과비결핵성심낭염 (NTPE) 을구분하기위한검사가충분히이루어지지않아정보편견 (information bias) 의발생을생각할수도있다 [18]. ADA (p) 에대한 Glucose (p), %Lymph (p)/glucose (p) 의상관 관계에서각분석마다 1-2명의동떨어진수치들로인해결과값이왜곡되었을가능성도배제할수없다 (Figs. 2 and 3). 단변량분석결과에따라다변량분석시결과에영향을미칠수있는 ADA (p), %Lymph (p)/glucose (p), 악성과거력유무이외의다른요인들은고려하지않았다. 나이, 성별, 악성과거력유무이외의기저질환은단변량분석결과통계적으로유의하지않았다 (Table 1). 본원에서는결핵성심낭염진단기준에앞서서론에서언급하였던심낭천자에서얻은심낭액의결핵세포진검사를포함하였으나, 본원에서이루어지지않은심낭생검을제외하였고, 확진 (definite diagnosis) 를위한충분한검사도이루어지지않았다 [5,13,15]. Mayosi 등 [8] 은심낭액에서결핵균 (mycobacterium tuberculosis, MTB) 의존재가입증되는경우확진 (definite tuberculous pericarditis) 이라정하였고, 그이외의검체 ( 예를들어흉수 ) 에서결핵균의존재하거나임상적으로의심되어투여된항결핵제에지속적인임상호전이있는경우결핵성심낭염 (TPE) 가능성이있어 (probable tuberculous pericarditis) 진단기준에포함시켰었다. 그이외의논문에서도임상적판단하에투여된항결핵제에지속적인임상호전이있을경우를결핵성심낭염 (TPE) 으로진단하였다 [11-13,19]. Park 등 [16] 의연구에서조직학적으로결핵이확진된 15예중 10예 (66.7%) 에서평균 23개월간추적검사상임상증상의호전과흉부방사선촬영이나심초음파도상심낭삼출의완해를볼수있었고, 조직검사상결핵이확진되지는않았으나경험적으로항결핵요볍을시행한경우를살펴보면조직검사상만성염증소견을보였던 19예와섬유화소견을보였던 7예의 26예중 18예 (69.2%) 에서완해를보여실제적으로결핵을확진하고나서치료를한것과그렇지않은경우를비교했을때그경과에있어서별다른차이가없음을알수있다고하였다. 한연구에서심낭액 LDH의진단적가치 [17] 를언급한것과같이본논문은결핵성심낭염 (TPE) 진단에있어심낭액림프구- 포도당비라는새로운지표를제시하였다는데에의미가있을것으로사료된다. 따라서결핵성심낭염 (TPE) 의진단을결정하는데있어, 이심낭액림프구- 포도당비라는지표가도움이되는지유무를재평가해야한다. 향후의대규모연구에서도심낭액 ADA, 심낭액림프구-포도당비를병합하여적용하였을때가각각단독으로사용하였을때보다상대적으로높은양성예측도, 특이도를보이고통계적유 - 465 -
- The Korean Journal of Medicine: Vol. 83, No. 4, 2012 - 의성이있다면항결핵약제투여전결핵성심낭염을확인하는데심낭액림프구- 포도당비가참고될수있다. 요약목적 : 결핵성심낭염진단에심낭액의검사을통해간편하고유용하게진단에도움을받을수있는생화학적지표로 ADA가이용되고있다. 그러나심낭액 ADA로결핵성심낭염과그외다른심낭염을감별해내기가어려운경우도있다. 이에심낭액 ADA 같이결핵성심낭염진단에참고될수있는심낭액생화학검사유무를확인하고자하였다. 방법 : 인산의료재단선린병원에서 2004년 6월 19일부터 2011년 11월 23일까지 42명의심낭천자술시행후생화학검사가결핵진단유무와관련있는지를분석하였다. 진단기준을정하여이에해당되었을때에결핵군, 해당되지않을때에비결핵군으로분류하였다. 분석은 ROC 곡선, 교차분석, 단순및다중로지스틱회귀분석, 단순상관분석을시행하였다. 결과 : ROC 곡선시행결과심낭액 ADA와 %Lymph (p)/ Glucose (p) ( 심낭액림프구- 포도당비 : 심낭액백혈구중림프구백분율을심낭액포도당수치로나눈값 ; 단위 : % ㆍdL/mg) 의제한치를각각 48.5 IU/L ( 민감도 90.0%, 특이도 87.5%, 유의확률 < 0.001) 와 0.678% ㆍdL/mg ( 민감도 90.0%, 특이도 71.9%, 유의확률 = 0.006) 로정하였다. 다중로지스틱회귀분석결과심낭액 ADA 48.5 IU/L ( 교차비 44.24; 95% 신뢰구간 2.85-686.97; 유의확률 = 0.07) 이었고, 심낭액림프구- 포도당비 0.678% ㆍdL/mg ( 교차비 20.39; 95% 신뢰구간 1.06-392.93; 유의확률 = 0.046) 는통계적유의성이있었다. %Lymph (p)/ Glucose (p) 0.678% ㆍdL/mg 과 ADA (p) 48.5 IU/L 를병합 (PPV 80.0% Sp 93.8%) 하였을때 %Lymph (p)/glucose (p) 0.678% ㆍdL/mg 단독 (PPV 47.4% Sp 68.8%), ADA (p) 48.5 IU/L (PPV 69.2% Sp 87.5%) 단독일때보다양성예측도, 특이도가상대적으로높았다. 결론 : 향후의대규모연구에서도동일한결과가나온다면심낭액림프구- 포도당비를심낭액 ADA와병합하여적용하는것이단일기준 ( 심낭액 ADA, 심낭액림프구- 포도당비 ) 일때보다비결핵성심낭염 (NTPE, non-tuberculous pericarditis) 을배제하는데더도움이된다. 중심단어 : 결핵성심낭염 ; 림프구 ; 포도당 REFERENCES 1. Kim DY, Park JH, Shin JD, et al. Long-term follow-up results and clinical manifestations of patients with a moderate to large amount of pericardial effusion. Korean J Med 2008; 74:154-161. 2. Park KS, Kim CH, Min BC, Choe KH. Diagnostic value of adenosine deaminase activity in tuberculous pericardial effusion. Korean Circ J 1990;20:141-147. 3. Eo WK, Lee SK, Choi CJ, et al. Clinical studies on the etiology and clinical course of pericardial effusions. Korean Circ J 1990;20:211-219. 4. Sagristà-Sauleda J, Mercé AS, Soler-Soler J. Diagnosis and management of pericardial effusion. World J Cardiol 2011; 26:135-143. 5. Maisch B, Seferović PM, Ristić AD, et al. Guidelines on the diagnosis and management of pericardial diseases executive summary; the Task Force on the diagnosis and management of pericardial diseases of the European society of cardiology. Eur Heart J 2004;25:587-610. 6. Tuon FF, Litvoc MN, Lopes MI. Adenosine deaminase and tuberculous pericarditis: a systematic review with metaanalysis. Acta Trop 2006;99:67-74. 7. Cho I, Chun EJ, Jeon KH, et al. A case of malignant pericardial mesothelioma misdiagnosed as tuberculosis pericarditis. Korean J Med 2009;76(Suppl 1):S81-S85. 8. Mayosi BM, Burgess LJ, Doubell AF. Tuberculous pericarditis. Circulation 2005;112:3608-3616. 9. Augustin P, Desmard M, Mordant P, et al. Clinical review: intrapericardial fibrinolysis in management of purulent pericarditis. Crit Care 2011;15:220. 10. Barber LM, Mazzadi L, Deakins DD, Reese CN, Rogers WL. Glucose level in pleural fluid as a diagnostic aid. Dis Chest 1957;31:680-687. 11. Cherian G. Diagnosis of tuberculous aetiology in pericardial effusions. Postgrad Med J 2004;80:262-266. 12. Reuter H, Burgess L, van Vuuren W, Doubell A. Diagnosing tuberculous pericarditis. QJM 2006;99:827-839. 13. Burgess LJ, Reuter H, Carstens ME, Taljaard JJ, Doubell AF. The use of adenosine deaminase and interferon-gamma as diagnostic tools for tuberculous pericarditis. Chest 2002; 122:900-905. 14. Ahn JO. Statistical Analysis of Biomedical Data Using SPSS 18.0. 1st ed., rev. 3, Seoul: Hannarae Publishing Co., 2011. 15. Levy PY, Fournier PE, Charrel R, Metras D, Habib G, Raoult D. Molecular analysis of pericardial fluid: a 7-year experience. Eur Heart J 2006;27:1942-1946. - 466 -
- Kyung Ho Yang, et al. The ratio between lymphocyte and glucose - 16. Park SY, Kim KS, Bae JH, Kim YH. The usefulness of pericardial biopsy to evaluate the causes of pericardial disease. Korean Circ J 1999;29:517-522. 17. Kim SE, Park DG. Diagnostic value of lactate dehydrogenase in pericardial effusions in patients undergoing pericardiocentesis. Korean J Med 2012;82:194-199. 18. Tripepi G, Jager KJ, Dekker FW, Wanner C, Zoccali C. Bias in clinical research. Kidney Int 2008;73:148-153. 19. Syed FF, Mayosi BM. A modern approach to tuberculous pericarditis. Prog Cardiovasc Dis 2007;50:218-236. - 467 -