腎功能不全者血中linezolid及其代謝物濃度與藥物不良反應之關聯性研究

Abstract

背景 Linezolid為oxazolidinone結構之抗生素，多用於治療具抗藥性之革蘭氏陽性球菌及非結核分枝桿菌感染，目前仿單建議劑量除較輕微之皮膚與軟組織感染外皆為600 mg IV/PO q12h，然而對於腎功能不全者沒有劑量調整建議。Linezolid有兩個主要肝臟代謝物：PNU142586及PNU142300，根據臨床前期試驗及單次劑量藥動學研究顯示，兩代謝物在腎功能不全（CLcr < 30 mL/min）者有明顯累積現象，然而後續並無文獻探討其血中濃度與腎功能，或linezolid相關不良反應之間的關聯性。因上述原因，促成本研究欲探討腎功能不全者血中linezolid及其代謝物濃度與藥物不良反應關聯性之動機。 研究目的 本研究主要探討linezolid與其代謝物之血中濃度在不同程度腎功能不全族群中的分佈、建立肌酸酐廓清率與藥物廓清率關係式，以及對於linezolid相關藥物不良反應之影響；此外，利用本族群在藥物不良反應之結果，找出linezolid與代謝物之血中濃度最適範圍。 研究方法 本研究延續前驅研究之前瞻（prospective）、觀察性（observational）設計，收案族群為2016年12月1日至2019年 5月7日間使用linezolid並進行療劑監測的成年病人。以高效能液相層析儀測量linezolid之峰（Cpeak）與谷（Ctrough）濃度，進一步計算最低（Cmin）與最高（Cmax）濃度、24小時濃度經時曲線下面積（AUC24）。在主要分析中僅納入第一次療程之第一次TDM結果，病人資料會以回顧電子病例之方式蒐集，安全性與臨床療效則會追蹤至該episode結束後之7天內，並記錄30天內之死亡率。Linezolid與其代謝物之血中濃度會進一步依照抽血日之肌酸酐廓清率（creatinine clearance, CLcr）分層分析，而不良反應則會依照Naranjo scale分數分類，並將確定與極有可能之不良反應視為linezolid 「相關」不良反應進行後續分析。 在描述性統計部份，會根據資料分佈型態決定以平均值 ± 標準差，或中位數及四分位距（interquartile；IQR）呈現；類別變數則以數量與百分比表示。比較各程度腎功能不全、有或無不良反應之族群與不同臨床療效間的血中濃度指標差異時，會以Mann-Whitney U test進行檢定。在分析腎功能與血中濃度關聯性時，濃度指標（應變數）若為類別變數則使用羅吉斯迴歸進行校正；若濃度指標為連續變數則使用簡單線性迴歸。此外，在不良反應與血中濃度關聯性則使用Cox 迴歸模型進行校正，若發現腎功能、原型藥或代謝物血中濃度為危險因子，則使用ROC分析法及自助抽樣法（boostrap）進行切點分析。 研究結果 本研究共納入117位病人，平均年齡為64.4歲、男女數量大約相等（男性占49.6%），且以住院病人為主（88.9%），約85%的病人於用藥後14天內進行療劑監測，而用藥開始前CLcr ≥ 60 mL/min者共51%、30 ≤ CLcr ≤ 59 mL/min者共25%，CLcr < 30 mL/min者則有24%。 在標準劑量（600 mg q12h）下，約有7成病人使用注射劑型，其linezolid Cmin、 Cmax及AUC24濃度平均分別為12.9 μg/mL、28.8 μg/mL及470.9 mg/L-h；代謝物PNU142586高於LLOQ的比例、平均給藥前血中濃度（僅納入高於LLOQ之濃度）及平均PNU142586/linezolid之比值（M/L ratio）分別為94.4%、13.8 μg/mL及2.2，而化合物 X則為74.7%、7.4 μg/mL及0.8。 依腎功能分層後發現嚴重腎功能不全（CLcr < 30 mL/min）族群相較於腎功能正常（CLcr ≥ 60 mL/min）者，代謝物給藥前濃度累積程度較原型藥劇烈，linezolid Cmin 、給藥前PNU142586濃度與化合物 X濃度約上升至3倍、12倍與8倍。若比較M/L ratio，同樣發現腎功能越差者代謝物累積情形有顯著上升。在校正其他因子後顯示腎功能分類在linezolid Cmin、Cmax、AUC24及 PNU142586與化合物 X之給藥前血中濃度、化合物 X高於LLOQ之比例、PNU142586之對數化M/L ratio及M/L ratio之比例為正向之顯著預測因子，其餘濃度指標無顯著結果。血小板低下、乳酸性酸中毒與血中乳酸上升之不良反應，結果同樣為腎功能越差者，發生比例顯著越高；校正其他因子後，顯示M/L ratio142586 >1者，發生「相關」之血小板低下風險顯著為M/L ratio142586 < 1時的8倍；乳酸性酸中毒與血中乳酸上升之正向顯著風險因子則包括linezolid Cmin >10 μg/mL。 血小板低下之血中濃度與腎功能切點分析則發現，Cmin切點平均為8.54 μg/mL，與pilot study之結果（Cmin平均＝9.14 μg/mL）相似1；PNU142586給藥前濃度切點平均則為2.71 μg/mL，且以PNU142586給藥前濃度＝3 μg/mL當切點時，其sensitivity及specificity相較於以Cmin＝9 μg/mL做為濃度切點時來得好（sensitivity: 62.8 % vs. 57.4%、specificity: 77.3% vs. 65.3%），但以腎功能（CLcr）做為切點時的表現皆劣於上述結果。 結論 本研究首次探討腎功能與linezolid代謝物血中濃度以及藥物不良反應之關聯性。結果與過去文獻相似，顯示相較於正常腎功能者（CLcr ≥ 60 mL/min），嚴重腎功能不全族群（CLcr <30 mL/min）在兩個linezolid肝臟代謝物的血中濃度皆有大幅累積。不良反應分析結果也發現linezolid Cmin及 M/L ratio142586 >1為「相關」之血小板低下的風險因子。在切點分析得PNU142586給藥前血中濃度之最佳切點約為3 μg/mL，且sensitivity及specifcity之表現皆優於以Cmin = 9 μg/mL或CLcr = 70 mL/min做為切點時之結果。本研究結果將應用於後續TDM、劑量調整與藥物不良反應之監測，以提升用藥安全。

Background Linezolid is an oxazolidinone antibiotic, which is widely used in Gram-positive cocci (GPC) and Mycobacterium spp. According to the U.S. food and drug administration (FDA) labeling, 600 mg PO/IV q12h was recommended to almost all of the GPC infection. In the clinical practice, the challenge of treatment is the side effects of linezolid. Moreover, linezolid-induced thrombocytopenia is considered as a concentration-dependent side effect in most of studies. Meanwhile, the incidence of linezolid-induced thrombocytopenia was significantly higher in patients with impaired renal function. Its two metabolites, PNU142300 and PNU142586, has a 7 -8 fold increase in exposure in severe renal impairment (CLcr < 30 mL/min), but there is no dose adjustment suggestion for this group, and the clinical significance of metabolite accumulation is not yet fully elucidated. Objective This study is aimed to investigate the association between plasma levels of linezolid and its metabolites in renal impairment. Meanwhile, confirming the effects of accumulation on linezolid-induced adverse effects. Finally, we want to identity the optimal range of creatinine clearance, plasma levels of linezolid and its metabolites. Methods We conducted a prospective observational study from December 1, 2016 to May 7, 2019 at National Taiwan University Hospital (NTUH). Linezolid adult users with at least 1 therapeutic drug monitoring (TDM) of linezolid were enrolled. High-performance liquid chromatography (HPLC) was appiled for plasma trough and peak level measurement. The minimum concentration (Cmin), maximum concentration (Cmax) and area under the 24-hour concentration-time curve (AUC24) was calculated at same time. Data was collected via electronic medical charts. Only the first TDM of the first course in each patient would be assessed. In concentration analysis, we calculated the renal function at the day blood was drawn by Cockcroft and Gault equation. Linear regression and logistic regression was performed to find out the correlation between plasma levels and renal function. Safety and clinical outcomes were assessed until 7 days after the completion of each episodes, and mortality at day 30 was also recorded. Furthermore, the correlation between plasma levels and ADRs was adjusted by using Cox proportional hazard model. Cut-off points were analyzed through application of receiver operating characteristic (ROC) curve and of bootstrap methods. In general, we comparing the difference of outcomes between two groups in this study by using Mann-Whitney U test. Significance level was 0.05. Results A total of 117 patients were enrolled, mean age were 64.4 years old. Most of the patients were inpatient (88.9%), and around 85% of them performed TDM within 14 days after linezolid use. Percentage of the patients with baseline CLcr ≥ 60 mL/min was 51%. In the end, a total of 115 episodes were included for analysis. 73% of the episodes used injection form linezolid with standard dosage (600 mg q12h), and the mean linezolid Cmin, Cmax and AUC24 were 12.9 μg/mL, 28.8 μg/mL and 470.9 mg/L-h, respectively. Mean plasma level before the next dose and metabolite/linezolid ratio (M/L ratio) of PNU142586 were 13.8 μg/mL and 2.2. Compared with patients with normal renal function, patients with severe renal impairment experienced more accumulations of metabolites: the plasma level of linezolid and PNU142586 before the next dose showed 3 and 12-folds. M/L ratio was also higher. In multivariable regression, a worse renal function was found to be a significant predict factor for linezolid Cmin, Cmax and AUC24. Similar results were found in M/L ratio with log transformation or M/L ratio >1 of PNU142586. In safety outcome analysis, we also found that the higher severity of renal impairment, the higher incidence rate of thrombocytopenia, lactic acidosis and elevated serum lactate level (p<0.05). The results from multivariable Cox regression showed when M/L ratio142586 >1, the hazard of “related” thrombocytopenia was significantly increased. And there was a positive finding between linezolid Cmin >10 μg/mL and the hazard of lactic acidosis and elevated serum lactate level. The cut-off point for the upper limit of Cmin in terms of thrombocytopenia in this study was similar to it was in pilot study (8.54 μg/mL vs. 9.14 μg/mL), but the sensitivity and specificity were improved when using 3 μg/mL as the cut-off point for the upper limit of plasma level of PNU142586 (sensitivity= 62.8 %, specificity= 77.3%). Lowest sensitivity and specificity were found when using CLcr = 70 mL/min as the cut-off point. Conclusion This study was the first study which aimed to find the association between ADRs and plasma concentrations of two metabolites of linezolid in patients with renal impairment. Overall, our results were consisted with previously published literatures, which showed plasma level of its metabolites has increase in exposure in patients with severe renal impairment. In addition, we found a positive correlation between M/L ratio142586 and linezolid-related thrombocytopenia, and the cut-off point for the upper limit of plasma level before the next dose of PNU142586 was 3 μg/mL. These results indicated that performing TDM may prevent ADRs.