Linezolid 藥品血中濃度與不良反應之關聯性研究

Abstract

背景 Linezolid 目前是治療革蘭氏陽性球菌及分枝桿菌感染的後線抗生素，嚴重的不良反應包括骨髓抑制、乳酸性酸中毒、周邊神經病變等。過去文獻提到linezolid 較高的血中濃度可能跟血小板低下有關，然而只有少數研究探討臨床療效及其他不良反應跟血中濃度的關係。同時劑量調整方式目前尚未有一致的共識；因此有進一步研究之必要性。 研究目的 此研究主要目的為分析linezolid 血中濃度的分布情形及劑量的關係、血中濃度與副作用及臨床療效之間的關係，以及尋找血中濃度的最佳切點。 研究方法 本研究為一前瞻性觀察性研究，於2016 年12 月1 日至2017 年5 月15 日間在國立臺灣大學醫學院附設醫院收案。收案對象為接受linezolid 治療且有進行療劑監測之成年病人，並使用高效能液相層析儀測量linezolid 的峰（Cpeak）、谷（Ctrough）濃度，再利用該濃度計算出最低（Cmin）、最高血中濃度（Cmax）以及24 小時曲線下面積（area-under-the-curve, AUC）。病人所有的資料都經由回顧電子病歷來記錄，同時會追蹤安全性及臨床療效直到療程結束後7 天，並記錄療程結束後30 天內死亡率。所有的不良反應都會由Naranjo scale 做進一步的評估，另外會將確定和極有可能的不良反應定義為linezolid 相關的不良反應納入分析。統計方法會使用Mann-Whitney U test 來比較各個不良反應（有、無）以及臨床療效（改善、失敗）兩組間的濃度差異；同時也會使用羅吉斯回歸進行校正。濃度的切點則是由ROC 曲線搭配自助抽樣法（bootstrap）進行分析，另外在劑量調整的部分則會以描述性統計的方式呈現。最後，在代謝物可能毒性的部分則使用Mann-Whitney U test 作初步的評估。 研究結果 在納入的49 位病人中，平均年齡為61.2 歲且有49 %的病人為男性，並於研究期間進行了66 次療劑監測，大部分（80.4 %）監測時間皆在14 天內。Linezolid 標準劑量（600 mg q12h）及減半劑量（600 mg qd）的Cmin, Cmax 及AUC24 濃度中位數分別為9.19 μg/mL、24.33 μg/mL、360.98 mg/L-hr 及3.48μg/mL、21.59 μg/mL、199.97 mg/L-hr。本研究血中濃度約比仿單高1.5 至2 倍之多，然而跟過去日本文獻相比則類似。在發生的副作用中最多的為血小板低下（62 %），最少為周邊神經病變（6 %）。發生血小板低下的病人的Cmin 顯著較無 發生者高（所有不良反應：13.06 vs. 6.97 μg/mL, p=0.0292；相關之不良反應：17.55 vs. 6.97 μg/mL, p=0.0007）；同時Cmin 在發生白血球低下的病人族群也顯著比無發生者高（相關之不良反應：22.54 vs. 9.62 μg/mL, p=0.0291）。在其他副作用（貧血、乳酸性酸中毒、周邊神經病變）以及臨床療效（包括死亡）的部分則未達顯著差異。在濃度切點的部分，若要維持linezolid 安全性應控制Cmin 低於9 μg/mL；此外，降低劑量且仍持續使用7 天以上似乎能改善血小板低下副作用。最後，初步評估的結果顯示代謝物毒性可能跟血小板低下、貧血、乳酸性酸中毒有關，然而需要更多研究來進一步探討其中的關聯性。 結論 Linezolid 標準劑量的血中濃度顯著高出仿單記載1.5 到2 倍，但跟日本族群類似。較高的Cmin 跟血小板低下及白血球低下有關，然而在其他副作用及臨床療效的部分則未達顯著差異。早期使用療劑監測（濃度切點：9 μg/mL）來調整linezolid劑量或許能預防相關之血小板低下及白血球低下的發生。

Background Linezolid is the second line in the treatment of gram-positive cocci and Mycobacterium spp. Severe adverse drug reactions (ADRs) of linezolid include myelosuppression, lactic acidosis, and peripheral neuropathy. Previous studies also showed that high plasma concentration of linezolid may induce thrombocytopenia. Only a few studies demonstrated the relationship between plasma concentration and clinical outcomes, and the appropriate method of dose adjustment still remained controversial. Further investigation is warranted. Objectives This study primarily aimed to investigate the association between concentration of linezolid, ADRs and clinical outcome. The cut-off point of linezolid concentration for therapeutic drug monitoring (TDM) was also explored. Methods A prospective observational study was performed at National Taiwan University Hospital (NTUH) between December 1, 2016, and May 15, 2017. Adult patients with TDM of linezolid were included. Linezolid peak and trough concentrations were measured using HPLC and then the concentrations used to calculate the minimum concentration (Cmin), maximum concentration (Cmax) and 24-hour area-under-the-curve (AUC). All patients’ data were collected via electronic medical charts. The safety and clinical outcomes were assessed throughout the treatment course until 7 days after the completeness of linezolid and 30-day mortality was also recorded. All ADRs were evaluated by Naranjo scale, and the definite/probable ADRs were defined as linezolid-associated adverse drug reactions in the analysis. Mann-Whitney U test was used to compare the difference of concentration between two groups of ADRs or clinical outcomes and the logistic regression was applied to adjust the potential confounders. Receiver operating characteristic (ROC) curve and bootstrap methods were used to calculate the cut-off points. The effect of dose adjustment was performed by descriptive analysis. Finally, the possible toxicity of metabolites was preliminarily evaluated by Mann-Whitney U test. Results Among 49 patients enrolled in the study, mean age were 61.2 years old and 49 % were male. There were 66 linezolid TDMs collected, and most (80.4%) were performed within 14 days. The median Cmin, Cmax and AUC24 of standard dose (600 mg twice daily) and reduced dose (600 mg daily) were 9.19 μg/mL, 24.33 μg/mL, 360.98 mg/L-hr and 3.48 μg/mL, 21.59 μg/mL, 199.97 mg/L-hr, respectively. Plasma concentrations were 1.5- to 2 folds higher than which in the prescribing information but similar to the results in Japanese studies. Thrombocytopenia was the most common adverse drug reactions (62 %) and peripheral neuropathy appeared infrequent (6 %). Median Cmin concentration was significantly higher in patients with thrombocytopenia compared with patients without thrombocytopenia (total ADRs: 13.06 vs. 6.97 μg/mL, p=0.0292; related ADRs: 17.55 vs. 6.97 μg/mL, p=0.0007). Median Cmin level was also significantly higher in patients with related leukopenia ADR than patients without leukopenia (22.54 vs. 9.62 μg/mL, p=0.0291). No association was identified between plasma concentration and anemia, lactic acidosis, peripheral neuropathy or clinical outcomes including mortality. The cut-off point for the upper limit of Cmin in terms of thrombocytopenia was 9 μg/mL. Dose adjustment might be useful for improving thrombocytopenia if that continued linezolid for at least 7 days or more. Preliminary results demonstrated that metabolites-associated possible toxicities were thrombocytopenia, anemia and lactic acidosis and further studies to explore the association was recommended. Conclusions Plasma concentrations of standard dose of linezolid were 1.5- to 2 folds higher than which in the prescribing information, but similar to the results in Japanese studies. Higher Cmin was associated with thrombocytopenia and leukopenia; however, no significant difference was observed between linezolid concentration and other ADRs or clinical outcome. Dose adjustment tailored by early TDM (cut-off point: 9 μg/mL) may be considered to prevent linezolid associated thrombocytopenia and leukopenia.