N端前B型利納利尿胜肽與B型利納利尿胜肽檢測在診斷心臟衰竭患者一致性的調查總量相關性之探討

Investigation of diagnostic concordance between NT-proBNP and BNP in heart failure patients

相較於歐美，台灣目前對於BNP與NT-proBNP之實際檢測模式與判讀一致性仍缺乏系統性驗證與規範。B型利納利尿胜肽（BNP）與N端前B型利納利尿胜肽（NT-proBNP）雖同源自preproBNP，因其生理特性、半衰期及清除機制不同，可能導致心衰竭判讀結果產生差異。BNP具生物活性，主要透過受體結合與酶促作用代謝清除；NT-proBNP則為無活性片段，主要由腎臟排除。本研究採回溯性分析，納入275位同時檢測BNP（以100 pg/mL為臨界值）與NT-proBNP（依年齡分別為450、900、1800 pg/mL）的疑似心衰竭個案，探討兩者結果的一致性與影響因子。整體結果顯示，161例兩項皆異常，61例皆正常，判讀一致性達81%（Cohen’s Kappa = 0.62，p = 0.038），高於Farnsworth等人（2018）報導的約70%。在53例不一致個案中，16例被診斷為心衰竭，其中4例使用Entresto，6例年齡超過75歲；其餘37例非心衰竭個案中，亦有三分之二年齡超過75歲，顯示年齡與藥物使用可能為影響判讀結果的主要干擾因子。統計分析顯示，Entresto使用與BNP偏高具有顯著關聯（OR = 12.3，p = 0.021）。此外，eGFR≤60 mL/min/1.73m²組之NT-proBNP/BNP比值明顯高於> 60組（10.20 vs. 3.16，p < 0.001），反映腎功能不全對NT-proBNP清除效率之影響。本研究結果顯示，BNP與NT-proBNP整體一致性良好，惟在高齡、腎功能不全或接受Neprilysin抑制劑治療者中，兩者檢測結果可能出現偏差。NT-proBNP在上述情境中可能更具穩定性與臨床參考價值。本研究透過回溯性資料分析，評估兩者在心衰竭診斷中的一致性及差異來源，進一步分析不一致結果的臨床特徵與潛在解釋因子。本研究非為比較兩者優劣，而是透過在地實證資料，協助臨床人員更全面理解其變異性，並提供更充分的臨床判讀與決策依據。

Background: Taiwan lacks standardized validation and interpretation guidelines for BNP and NT-proBNP testing in heart failure, unlike Europe and the United States. Although both biomarkers originate from preproBNP, their differences in biological activity, half-life, and clearance mechanisms—BNP undergoing enzymatic degradation and NT-proBNP being primarily cleared by the kidneys—may lead to inconsistent diagnostic outcomes.
Objective: This study aimed to evaluate the concordance between BNP and NT-proBNP results in patients suspected of heart failure and to identify potential clinical factors contributing to discrepancies.
Methods: This retrospective analysis included 275 patients who underwent both BNP (cutoff: 100 pg/mL) and NT-proBNP testing (cutoffs: 450, 900, and 1800 pg/mL according to age). Concordance between the two tests was assessed using Cohen’s Kappa, and clinical characteristics of discordant cases were further analyzed.
Results: Among the participants, 161 had both biomarkers elevated, and 61 had both within normal limits, yielding an agreement rate of 81% (Cohen’s Kappa = 0.62, p = 0.038). Among 53 discordant cases, 16 were diagnosed with heart failure; 4 were using Entresto, and 6 were aged over 75. Of the 37 non-heart failure cases, about two-thirds were also aged over 75, suggesting age and medication may influence discordance. Entresto use was associated with a significantly increased likelihood of elevated BNP (OR = 12.3, p = 0.021). Moreover, patients with impaired renal function (eGFR≤60) had a significantly higher NT-proBNP/BNP ratio than those with normal function (10.20 vs. 3.16, p < 0.001), reflecting the impact of renal clearance.
Conclusion: BNP and NT-proBNP results showed good overall agreement in diagnosing heart failure. However, discrepancies were more common among elderly patients, those with renal impairment, and those receiving neprilysin inhibitors. In such cases, NT-proBNP may serve as a more stable biomarker. This study provides real-world evidence to support clinical interpretation and decision-making without aiming to rank one marker above the other.