併用醣蛋白類抗生素與β-內醯胺類抗生素和急性腎損傷的風險

Abstract

引言: 在臨床上，vancomycin (VAN)常常與piperacillin/tazobactam (TZP)一同並用來治療嚴重的感染症。在2011年，有幾篇研究發現VAN和TZP的並用會增加急性腎損傷的風險。此後，越來越多研究也證實此項發現。在台灣，除了VAN外，teicoplanin (TA)也是另外一種常用的醣蛋白類抗生素。雖然TA相較於VAN有較少的腎毒性風險，然而目前臨床上並沒有充足的證據支持當並用TZP時，以TA取代VAN的使用。宗旨: 在並用不同醣蛋白類抗生素和β-內醯胺類抗生素(β-lactam)的病人中，欲比較各組間急性腎損傷的風險，主要分成3個目標研究: 1. TA與TZP並用組和VAN與TZP並用組(也就是: TA+TZP vs. VAN+TZP)。2. TA與TZP並用組和TA與β-lactam並用組(也就是: TA+TZP vs. TA+β-lactam)。3. VAN與TZP並用組和VAN與β-lactam並用組(也就是: VAN+TZP vs. VAN+β-lactam)。本研究所選取的β-lactam為與TZP具有相似抗菌譜範圍的抗生素。方法:本研究為回溯性的世代研究，使用台灣大學醫學院附設醫院的電子健康紀錄資料。病人如果在住院期間有並用任一種醣蛋白類抗生素與TZP或是特定的β-lactam會被納入為本研究的研究對象。之後再分類為4個研究族群，分別為TA與TZP並用組、VAN與TZP並用組、TA與β-lactam並用組和VAN與β-lactam並用組。在每一個研究族群中，並用之二種抗生素須並用至少3天，病人若在並用抗生素前符合以下條件會被排除: 估算腎絲球過濾率小於15 mL/min/1.73m2、使用腎臟替代療法或近期發生急性腎損傷。本研究使用傾向分數匹配的方式來控制潛在的干擾因子，並分別計算3個目標研究中二組間的風險比值。結果: 於第1個目標研究中，共有211組配對病人(TA與TZP並用組和VAN與TZP並用組以1:1的方式配對)，TA和VAN劑量的中位數分別為10.3 mg/kg/day和26.7 mg/kg/day。VAN的低谷濃度中位數為12.3 mg/L。TA和TZP並用組和VAN與TZP並用組具有相近的急性腎損傷風險 (12.3% vs. 11.4%; HR=1.23, 95% CI=0.71-2.15, P=0.464)。於第二個目標研究中，共有243組配對病人 (TA與TZP並用組和TA與β-lactam並用組以1:3的方式配對)，二組的TA劑量中位數皆為10.7mg/kg/day。TA與TZP並用組和TA與β-lactam並用組也具有相近急性腎損傷風險(14.8% vs. 14.2%; HR=1.26, 95% CI=0.86-1.85, P=0.238)。於第三個目標研究中，共有205組配對病人(VAN與TZP並用組和VAN與β-lactam並用組以1:4的方式配對)，二組VAN劑量的中位數約為28.0 mg/kg/day。相較於VAN與β-lactam並用組，VAN與TZP並用組在統計上有顯著的急性腎損傷風險(13.2% vs. 9.6%;HR=1.72, 95% CI=1.10-2.68, P=0.02)。總結:本研究結果並不支持臨床上以TA與TZP的並用來取代VAN與TZP的並用組合；亦無法支持使用TA與β-lactam的並用來取代TA與TZP的並用組合。當考慮到急性腎損傷的風險時，臨床上建議以VAN與其他β-lactam的並用來取代VAN與TZP的並用組合。我們也建議密集的監控血液中的肌酸酐值，當病人同時並用醣蛋白類抗生素以及TZP或是其他β-lactam類抗生素。最後，本研究的病人使用較低劑量的VAN且有較低的VAN 谷底濃度，我們建議更多的相關研究來探討急性腎損傷和TA與TZP並用和VAN與TZP並用的關係，尤其當病人的VAN谷底濃度介於15至20 mg/L之間時。

Introduction: Vancomycin (VAN) and piperacillin/tazobactam (TZP) are commonly combined antibiotics used to treat some severe infection. Since 2011, a signal of increased risk of acute kidney injury (AKI) was detected when VAN and TZP were administered concurrently. More evidence came one after another to prove this phenomenon. Besides VAN, teicoplanin (TA) is another glycopeptide widely used in Taiwan. Although TA has less nephrotoxicity than VAN, there is no evidence to support it to replace VAN with TA both in combination with TZP clinically.Aim:To compare the risk of acute kidney injury (AKI) among patients receiving the following glycopeptide and β-lactam combinations: 1) TA plus TZP versus VAN plus TZP (i.e., TA+TZP vs. VAN+TZP), 2) TA plus TZP versus TA plus other β-lactams with similar antibacterial spectrum to TZP (i.e., TA+TZP vs. TA+β-lactam) and 3) VAN plus TZP versus VAN plus β-lactam (i.e., VAN+TZP vs. VAN+β-lactam).Methods: This was a retrospective cohort study using electronic health records from National Taiwan University Hospital (NTUH). Patients were included if a combination of glycopeptide and TZP or other selected β-lactam were used during hospitalization. Four study groups were identified: TA+TZP, VAN+TZP, TA+β-lactam and VAN+β-lactam. In each group, two antibiotics of interest were required to overlap for at least three days. Patients with estimated glomerular filtration rate <15 mL/min/1.73m2, renal replacement therapy, or recent AKI history were excluded. We used propensity score matching to control for potential cofounders, and hazard ratio (HR) of AKI between study groups was calculated.Results:The final sample contained 211 TA+TZP and VAN+TZP pairs (1:1 match). The median dose of TA and VAN was 10.3 and 26.7 mg/kg/day, respectively. The median trough level of VAN was 12.3 mg/L. AKI risk was similar in the TA+TZP group compared to that in the VAN+TZP group (12.3% vs. 11.4%; HR=1.23, 95% confidence interval [CI]=0.71-2.15, P=0.464). For the second comparison, there were 243 TA+TZP and TA+β-lactams pairs (1:3 match). The median dose of TA was both 10.7 mg/kg/day. There was similar AKI risk between the use of TA+TZP and TA+β-lactam (14.8% vs. 14.2%; HR=1.26, 95% CI=0.86-1.85, P=0.238). For the final comparison, there were 205 VAN +TZP and VAN +β-lactams pairs (1:4 match), and the median dose of VAN was both around 28.0 mg/kg/day. There was higher risk of AKI in the VAN+TZP group (13.2% vs. 9.6%; HR=1.72, 95% CI=1.10-2.68, P=0.02).Conclusion:Our study does not support TA+TZP as a good substitute of VAN+TZP when AKI is of concern. Neither is TA combined with TZP and TA with β-lactams. The result of higher AKI risk among patients using VAN+TZP suggests that when combination treatment is needed, VAN and other β-lactam might be abetter choice.We also suggest a close monitoring of serum creatinine when patients use combination of glycopeptide and either TZP or otherβ-lactam. Finally, due to the low VAN dose and trough level in this study, more evidence is required to further evaluate risk of AKI among patients receivingTA+TZP andVAN+TZP, especially when VAN trough is between 15 and 20 mg/L.