Relative platelet reductions provide better pathophysiologic signatures of coagulopathies in sepsis
概要
主論文の要旨
Relative platelet reductions provide better
pathophysiologic signatures of coagulopathies in sepsis
血小板の相対的減少は敗血症における凝固障害の
病態生理学的特徴を反映する
名古屋大学大学院医学系研究科
生体管理医学講座
総合医学専攻
救急・集中治療医学分野
(指導:松田 直之
春日井 大介
教授)
【緒言】
敗血症は、感染症による制御不能な生体応答であり、多臓器不全を特徴とする病態
である。多くの無作為化比較試験(RCT)や治療ガイドラインにより敗血症の生存率は
改善傾向にあるものの、敗血症に合併する播種性血管内凝固(disseminated intravascular
coagulation:DIC)は死亡率を高める傾向がある。敗血症においては、血小板産生量の
減少ではなく複数の経路を介して血小板消費速度が増加することで血小板減少が起こ
ると考えられている。国際血栓止血学会等の敗血症における凝固障害の国際的基準は、
血小板数減少の閾値に注目しているが、相対的な血小板数の減少を評価するものでは
ない。このような背景から、相対的血小板減少と絶対的血小板数が敗血症の転帰に及
ぼす影響を評価する方針とした。本研究は、相対的な血小板減少率と絶対的血小板数
が敗血症の転帰に与える影響を推定することを目的とした。
【方法】
eICU Collaborative Research Database(2014-2015 年の米国内 208 病院、335 ICU のデ
ータリポジトリから層別無作為抽出した患者合計 20 万 859 人の ICU 入院患者からな
る公開データベース)を用いて、後ろ向き観察研究を実施した。ICU 入室時に敗血症の
患者を対象とし、敗血症の定義は、Sepsis-3 の定義に従った。ICU 入室時に血小板数の
データがない者、ヘパリンによる血小板減少症と診断された者、死亡した者、ICU 入
室後 2 日以内に退院した者は一次解析から除外した。相対的血小板減少率は、1 日目
の初期血小板数及び 2 日目の最小値を用いて算出した。Immortal time バイアスを避け
るため、対象期間前に死亡または退院した患者を除外した。主要転帰は院内死亡率と
し、副次的アウトカムは凝固障害関連合併症の有無とした。一次モデルに用いた共変
量は、年齢、性別、BMI、人種、合併症、感染の焦点、院内発症、APACHE IV score、
ICU 入室回数、IC U の種別などの患者レベル変数に加えて病床数、病院所在地地域、
教育病院か否かなどの病院レベルの変数を含めた。一次解析では多変量混合効果ロジ
スティック回帰分析を行い、相対的血小板減少率の各範囲の死亡オッズおよび初日の
絶対血小板数の死亡オッズを算出した。各変数の最適なカットオフ値は ROC 曲線解
析を使用して決定した。欠損データは、観察された共分散をもとに missing at random
仮定として多重代入法を用いた。感度解析では、主要解析の頑健性を検証するために、
(1)complete case 解析(2)共変量追加モデル(3)代替コホートでの解析を繰り返した。す
べての統計解析は、R バージョン 4.0.0 を使用して行った。
【結果】
敗血症患者 30,114 人のうち 26,176 人(87.0%)が一次解析に含まれ、3,938 人が除外
された(Figure1)。患者は平均年齢>60 歳で、白人が多く(77.8%)、呼吸器感染症があ
り(52%)。入院期間中央値は 7.4 日(IQR 4.7〜12.1 日)、院内死亡率は 12.3%であった。
ICU 入室 2 日目以降の凝固性合併症の新規発症は比較的まれであった(血栓性合併症:
139 例[0.5%]、血栓性合併症:139 例[0.5%])(Table1)。多変量混合効果ロジスティック
回帰分析では、院内死亡率の調整オッズ比は、血小板減少率 20%未満と比較した場合、
20~40%、40~60%、60~80%、80%以上では、それぞれ 1.28、2.99、6.05 だった(Table1)。
2 日目の絶対血小板数の推定効果は比較的小さかった(OR 1.2)。相対的な血小板減少
量とその結果としての絶対的な血小板数の間には死亡オッズに対する相乗的交互作用
が認められた(P = 0.006)。ショック の存在は相対的血小板減少の死亡オッズとの間に
負の交互作用を示した(P < 0.001)。呼吸不全の存在と凝固障害関連の合併症の発症は
相対的血小板減少の死亡オッズを修正しなかった。一般化加法モデルを用いたスプラ
イン回帰分析では、相対的血小板減少の院内死亡率に対する推定効果は最初の 4 日間
は規則的に増加したが、5~7 日目の 40%以上減少の推定効果は大幅な血小板減少の発
生がまれであるために信頼区間が拡大し不安定になった(Figure2)。院内死亡率の推定
オッズは、血小板数が初期 7 日間のどの日においても約 10 万/μL まで減少すると規則
的に増加した。凝固障害関連の合併症を予測するための多変量ロジスティック回帰分
析の結果、相対的血小板減少率と絶対的血小板数はともに凝固障害関連合併症と関連
していたが(複合転帰の OR[95%CI]
:それぞれ 1.25[1.19-1.31]、1.18[1.07-1.30])、
相対的減少率のみが血栓塞栓イベントの増加と関連していた(OR[95% CI]:1.43[1.031.98])(Table 3)。血小板減少率の推定修正オッズ比の結果は一次解析、complete case、
共変量追加モデル、代替コホートによる解析で同様であった。
【考察】
本研究は、敗血症における血小板減少率と血小板絶対数の 2 つの因子と、敗血症の
予後との関連を評価する大規模の研究となった。本研究の主な知見は(1)血小板減少
率が高いほど推定死亡率が高く(2)血小板減少は凝固障害関連の合併症と関連し、こ
れは結果としての血小板減少の絶対値とは無関係であることだった。死亡オッズの上
昇は、敗血症の 2〜7 日目の血小板減少の大きさの増加と関連しており、それは結果と
しての血小板数とは無関係であることが示された。さらに、血小板減少症とは無関係
に、血小板減少症とマクロ血栓症および出血性イベントとの関連性が確認された。一
方、凝固障害関連の合併症を伴わないサブグループ解析の結果、血小板減少は、主要
な止血機能以外の免疫学的あるいは微小血栓症のメカニズムによる死亡率の増加に関
与していることが示唆された。これらの知見より、敗血症性凝固障害における血小板
減少率の重要性が明確とされた。重症度の層別化は、集中治療分野における過 去の
RCT において介入の有効性を示すことに貢献している。臨床試験を計画する際、血小
板減少率は敗血症における凝固異常の重症度の層別化とサンプルサイズの推定に有用
である。絶対血小板数は、血栓塞栓イベントの高いオッズとの関連を示せなかったこ
とから、抗凝固治療の評価を検討する上では絶対数の基準だけで最適な患者をターゲ
ットにすることは不十分であり、相対的な血小板減少の指標活用が有用かもしれない。
さらに、この指標は凝固異常症に対する予防的介入の評価指標としても活用可能であ
る。治療の初期段階における急速な血小板減少を予測するためには、さらなる研究が
必要である。本研究では、層別ランダムサンプリングされた米国の多施設大規模デー
タベースを使用することで、堅牢な推定を提供し、特定の施設における敗血症の特性
や管理によって偏ることがない。この結果は、高所得国のさまざまな ICU における敗
血症の患者に一般化することができる。しかし、本研究にはいくつかの限界があった。
第一に、早期死亡者を除外した結果、重症度の低い患者を選択することになった可能
性がある。第二に、本データには凝固マーカーの詳細が含まれておらず、これらのマ
ーカーと比較した有用性の評価はできなかった。また、血小板減少を引き起こしてい
る因子が患者側の制御不能な応答による結果なのか、病原体の制御不全によるものか
を特定することはできなかった。血小板減少が主に何によって引き起こされるかは更
なる前向きの調査が必要である。第三に、合併症の重症度が報告されていないことで
ある。高リスク集団(血小板減少が急激な患者)における凝固性合併症の発生率と重症
度を評価する前向き研究が必要である。
【結論】
相対的な血小板減少の程度は、絶対的な血小板数と比較して、死亡率や凝固障害関
連合併症のリスクを層別化するためのより良いツールであることが示唆された。敗血
症における凝固異常の評価基準として、血小板減少率を評価する妥当性が検証された。