Background About 2/3 patients of coronary heart disease (CHD) are complicated with disorder of carbohydrate metabolism which results in hyperglycemia and subsequent abnormality of coagulation system and inflammation. These patients have serious coronary artery pathology, multiple complications and poor prognosis. Platelets and lymphocytes play important roles in the occurrence and progression of atherosclerosis. The platelet/lymphocyte rate (PLR) is one simple hematological index. Previous studies confirmed that PLR could predict the long-term mortality of non-ST elevated myocardial infarction (NSTEMI). If simple hematological index could predict the prognosis of such kind of patients, it will provide new thought for early diagnosis and treatment in future. Therefore, the present study try to investigate if PLR could predict the poor prognosis of CHD patients complicated with impaired glucose tolerance (IGT) through calculating PLR. Methods/design The present study is performed with strategy of an observational and prospective single-centre cohort. These patients are recruited from August 2013 to August 2014, according to the inclusion criteria of CHD complicated with IGT. CHD is confirmed with coronary angiography while IGT is determined according to the WHO criteria (1999). Routine blood test and serum glucose data of patients are acquired before hospitalization and surgery. According to the median of PLR after admission, the patients are divided into 3 groups. The patients are followed up for half, 1 and 3 years, respectively. The major clinical endpoint is mortality. The minor clinical endpoint indices are the correlations of PLR with MACE (including mortality, recurrent rate of infarction and reperfusion rate of target vessels), recurrent infarction, re-perfusion rate of target vessel, intra-stand thrombogenesis, stroke and acute onset of heart failure. The correlations are analyzed with receiver operating characteristics (ROC) survival curve and Kaplan-Meier survival analysis to find optimal prognosis index. Summary Through regression analysis of long-term follow-up of patients, it is expected to find optimal predicting index of prognosis. While judging whether PLR is effective, other possible factors for new predictor are sought in order to provide help for future study.