Premium
Development and validation of a prediction model (AHC) for early identification of refractory thrombotic thrombocytopenic purpura using nationally representative data
Author(s) -
Gui RuoYun,
Huang QiuSha,
Cai Xuan,
Wu Jin,
Liu HuiXin,
Liu Yi,
Yang LinHua,
Zhang JingYu,
Cheng YunFeng,
Jiang Ming,
Mao Min,
Fang MeiYun,
Liu Hui,
Wang LiRu,
Wang Zhao,
Zhou HeBing,
Lan Hai,
Jiang ZhongXing,
Shen XuLiang,
Zhang Lei,
Fan ShengJin,
Li Yueying,
Wang QianFei,
Huang XiaoJun,
Zhang XiaoHui
Publication year - 2020
Publication title -
british journal of haematology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.907
H-Index - 186
eISSN - 1365-2141
pISSN - 0007-1048
DOI - 10.1111/bjh.16767
Subject(s) - medicine , cohort , logistic regression , refractory (planetary science) , creatinine , thrombotic thrombocytopenic purpura , retrospective cohort study , pediatrics , platelet , physics , astrobiology
Summary Immune‐mediated thrombotic thrombocytopenic purpura (iTTP) is a rare and life‐threatening haematological emergency. Although therapeutic plasma exchange together with corticosteroids achieve successful outcomes, a considerable number of patients remain refractory to this treatment and require early initiation of intensive therapy. However, a method for the early identification of refractory iTTP is not available. To develop and validate a model for predicting the probability of refractory iTTP, a cohort of 265 consecutive iTTP patients from 17 large medical centres was retrospectively identified. The derivation cohort included 94 patients from 11 medical centres. For the validation cohort, we included 40 patients from the other six medical centres using geographical validation. An easy‐to‐use risk score system was generated, and its performance was assessed using internal and external validation cohorts. In the multivariable logistic analysis of the derivation cohort, three candidate predictors were entered into the final prediction model: age, haemoglobin and creatinine. The prediction model had an area under the curve of 0.886 (95% CI: 0.679–0.974) in the internal validation cohort and 0.862 (95% CI: 0.625–0.999) in the external validation cohort. The calibration plots showed a high agreement between the predicted and observed outcomes. In conclusion, we developed and validated a highly accurate prediction model for the early identification of refractory iTTP. It has the potential to guide tailored therapy and is a step towards more personalized medicine.