A score to predict the occurrence of very late stent thrombosis in patients treated by percutaneous coronary intervention for acute coronary syndrome

Study population

From
January 1, 2014 to June 1, 2015, a total of 6,038 consecutive ACS patients who were admitted to Jilin University 1st
hospital for PCI treatment and met the basic inclusion and exclusion criteria
were enrolled . The
primary endpoint was definite stent thrombosis. We exclude patients who died
prior to discharge (n=236), did not agree to take part in the follow-up
interview (n=413), or were lost to follow-up after discharge (n=204). The
median length of follow-up was 1,502 days.

The
external validation cohort included 2,438 consecutive patients admitted to the first affiliated hospital of Shantou University for PCI from
January 1, 2014 to January 1, 2015 with the same inclusion and exclusion
criteria. We exclude patients who died prior to discharge (n=88), did not agree
to take part in the follow-up interview (n=186), or were lost to follow-up
after discharge (n=106). The median length of follow-up was 1422 days.

Ethics
committee approval was obtained for all investigational sites, and the study
was performed in accordance with the ethical standards laid down in the 1964
Declaration of Helsinki and its later amendments. All patients gave written
informed consent.

Definition of VLST

VLST was
defined according to the 2007 definition of Academic Research Consortium (ARC) stent
thrombosis criteria: stent thrombosis that occured more than one year after
coronary stent implantation and was clearly confirmed by coronary angiography
(CAG).

Statistical Analysis

Data are
summarized as number (percentages) for categorical variables. Continuous
variables are expressed as mean±SD. Categorical data were compared using the Chi-square
test or Fisher exact tests while quantitative data were compared using the
Mann–Whitney U test. All tests were 2-sided at the 0.05 significance level.

The
derivation cohort was used to identify predictors of occurrence of VLST and to establish
a risk-scoring system which was verified in the validation cohort. The
significant variables (p<0.05) in univariable cox regression analysis and
independent risk factors reported in previous literature were included in the multivariate
cox regression analysis, and the optimal predictive variables of VLST were
obtained by backward stepwise method. Some variables, such as leukocyte count (there
is no sufficient evidence to completely exclude the possibility of infection
and other factors that leading to its increase) are not included in the final
model. Procedural variables such as stent diameter, length, type, and post-dilation
were based on the operator's consideration during PCI.

Multicollinearity
between the variables in the final model was assessed by calculation of the
variance inflation factor. These corrected coefficients of significant
predictors in the final model were then divided by the lowest coefficient value
and rounded to the nearest integer to assign a weight to each predictor. Total
score of each patient was calculated by adding these weights together. An
objective assessment of calibration was obtained by performing the
Hosmer–Lemeshow goodness-of-fit test and by plotting observed vs. predicted
incidence rate across deciles of risk score. The predictive performance of the
risk model was assessed by the c-statistic and was cross-validated (leave one
out method) in the development cohort. Clinical usefulness was evaluated using
decision curve analysis. Statistical analysis was performed with the use of
Statistical Package for Social Sciences version 23 (SPSS, Chicago, IL, USA) and
R version 3.5.0 (R Foundation for Statistical Computing, Vienna, Austria) using
the packages rms.