{"id":561530,"date":"2024-03-22T09:21:57","date_gmt":"2024-03-22T13:21:57","guid":{"rendered":"https:\/\/platohealth.ai\/platowire\/postdischarge-major-bleeding-myocardial-infarction-and-mortality-risk-after-coronary-artery-bypass-grafting-renal-platohealth-ai\/"},"modified":"2024-03-24T05:59:53","modified_gmt":"2024-03-24T09:59:53","slug":"postdischarge-major-bleeding-myocardial-infarction-and-mortality-risk-after-coronary-artery-bypass-grafting-renal-platohealth-ai","status":"publish","type":"platowire","link":"https:\/\/platohealth.ai\/platowire\/postdischarge-major-bleeding-myocardial-infarction-and-mortality-risk-after-coronary-artery-bypass-grafting-renal-platohealth-ai\/","title":{"rendered":"Postdischarge Major Bleeding, Myocardial Infarction, And Mortality Risk After Coronary Artery Bypass Grafting – Renal.PlatoHealth.ai"},"content":{"rendered":"
<\/div>\n
\n

Data sources and definitions<\/h3>\n

Individual patient data from five mandatory nationwide registries were merged and pseudonymised by the official authority Statistics Sweden<\/em>, as previously described.12<\/a> The merging of patient data was based on the personal identification number which is given to all Swedish residents at birth or shortly after immigration. The Swedish Cardiac Surgery Registry<\/em> is a part of the SWEDEHEART Registry<\/em> and contains information, including preoperative status and postoperative complications, on cardiac surgery procedures in Sweden since 1992, with a coverage of 98\u201399%.13 14<\/a> The National Patient Register<\/em> has full coverage of diagnoses on hospital admission according to the International Classification of Diseases, Ninth Revision (ICD-9) and 10th Revision (ICD-10), with a validity of 85\u201395%.15<\/a> The Swedish Prescribed Drug Register<\/em> has information on all prescriptions dispensed from Swedish pharmacies since July 2005 (according to the Anatomical Therapeutic Chemical (ATC) classification). The Swedish Cause of Death Register<\/em> comprises data on all deaths of persons registered in Sweden. The Longitudinal Integration Database for Health Insurance and Labor Market Studies (LISA) Register<\/em> contains data on social variables. The Swedish Population Register<\/em> contains basic demographic data, including date of emigration where applicable.<\/p>\n

Major bleeding and MI were defined as a new hospitalisation with a primary diagnosis according to the ICD-10 of bleeding or MI, respectively; the ICD codes used are listed in online supplemental table S1<\/a>. Bleeding events were further subclassified, according to bleeding site, as intracranial, gastrointestinal, pericardial or \u2018other bleeding\u2019 location.<\/p>\n

Supplemental material<\/h3>\n<\/p>\n

All relevant comorbidities registered in the Swedish Cardiac Surgery Registry<\/em> and National Patient Register<\/em> until start of follow-up were registered as baseline data; the ICD codes used are listed in online supplemental table S2<\/a>. Data on socioeconomic variables were retrieved from the LISA Register<\/em> at baseline. Preoperative data on left ventricular ejection fraction and kidney function were gathered from the Swedish Cardiac Surgery Registry<\/em>. The Chronic Kidney Disease Epidemiology Collaboration formula was used to calculate estimated glomerular filtration rate (eGFR) based on preoperative creatinine levels.16<\/a> Time-updated use of platelet inhibitors, anticoagulants, beta blockers, renin angiotensin system (RAS) inhibitors and statins was defined as dispensed prescriptions in the Swedish Prescribed Drug Register<\/em> and updated every third month as previously described.12<\/a> ATC codes used are listed in online supplemental table S3<\/a>.<\/p>\n

This study has been composed according to the recommendations in the Strengthening the Reporting of Observational Studies in Epidemiology reporting guideline.17<\/a> The study was performed in accordance with the Declaration of Helsinki.<\/p>\n<\/div>\n

\n

Statistical analysis<\/h3>\n

Continuous variables are presented as mean with SD, median with range or median with 25th\u201375th percentiles where appropriate. Categorical variables are presented as frequency with percentage. For comparison between two groups, Fisher\u2019s exact test was used for dichotomous variables, Mantel-Haenszel \u03c72<\/sup> trend test for ordered categorical variables, \u03c72<\/sup> test for unordered categorical variables and Mann-Whitney U test for continuous variables. Piecewise Cox proportional hazards models for the effect of major bleeding and MI, respectively, on subsequent mortality risk were developed. Major bleeding and MI were handled as a time-updated covariate in the respective models, meaning that patients contributed follow-up time to the \u2018no event\u2019 group until they had an event, at which time they started contributing follow-up time to the \u2018bleeding\u2019 or \u2018myocardial infarction\u2019 group, respectively. The model was adjusted for age, sex, year of surgery, left ventricular ejection fraction, eGFR category, time-updated use of platelet inhibitors, oral anticoagulants, beta blockers, RAS inhibitors and statins, and other baseline variables that were associated with mortality risk based on a separately performed stepwise (forward) regression model applying p<0.10. This resulted in further adjustment for use of left internal mammary artery, number of distal anastomoses, clinical heart failure, diabetes, peripheral vascular disease, smoking, history of cancer, stroke, chronic respiratory disease, renal failure, any previous bleeding requiring medical attention, atrial fibrillation, hyperlipidaemia, asthma, previous PCI, previous MI, hypertension, transient ischaemic attack, haemorrhagic stroke, subarachnoid haemorrhage, education level and marital status. To estimate the time-dependent risk of mortality after an event, HRs compared with the \u2018no event\u2019 group were calculated for <30, 30\u2013365 and >365\u2009days after first incidence of major bleeding and MI, respectively, similar to the methodology used by Marquis-Gravel et al<\/em>.11<\/a> Interaction analyses for mortality risk were studied in patients with major bleeding and MI, respectively, for the following subgroups: age (\u226575\/<75 years), gender, atrial fibrillation, kidney function (eGFR\u226560\/<60\u2009mL\/min\/1.73 m2<\/sup>) and multimorbidity (\u22653 of diabetes, hypertension, chronic respiratory disease, previous stroke, previous MI, atrial fibrillation, clinical heart failure, eGFR<60\u2009mL\/min\/1.73 m2<\/sup>).<\/p>\n

Similar piecewise Cox proportional hazards models were used to estimate the mortality risk for different bleeding locations compared with patients with no bleeding. Because of the small number of cases of pericardial bleeding (n=5), these patients were included in the \u2018other bleeding\u2019 group.<\/p>\n

Altogether, 287 patients suffered both major bleeding and an MI during follow-up. These patients were included in both the analysis investigating major bleeding and the analysis investigating MI. An analysis excluding patients suffering both major bleeding and MI was performed as a sensitivity analysis, with a similar piecewise Cox proportional hazards model as described above. Another sensitivity analysis excluding patients using anticoagulants was also performed.<\/p>\n

Missing data were handled as a separate category in the adjustments. All tests were two tailed and interpreted at the 0.05 significance level. Statistical analyses were performed using SAS software V.9.4 (SAS Institute).<\/p>\n<\/div>\n