Death, coronary revascularization choices, and chronic kidney disease
In contrast to these very personal examples, the medical community embraces guidelines based upon ‘evidence’. In some ways, this is the other end of the decision-making spectrum. It is an impersonal approach in which the choice among alternative approaches is determined by the average response of average patients in the same situation.
Somewhere between these two poles lies a ‘sweet spot’ in which there is a merger of personalized decision-making that is also informed by evidence-based medicine. We have less understanding of how individuals approach this interaction between their personal wishes and the evidence from average patients to make a specific decision. This becomes relevant to how we, as medical providers, use evidence-based data to provide informed consent for many procedures.
In this issue of Coronary Artery Disease data from two registries in British Columbia, Canada were combined to assess the mortality and renal outcomes of individuals with baseline chronic kidney disease [estimated glomerular filtration rate (eGFR)<60 ml/min and referred to a nephrologist] and who underwent coronary artery bypass grafting (CABG), coronary angiography and percutaneous coronary interventions (PCIs), or medical therapy for coronary artery disease between 2003 and 2009. There are likely to be vast differences between these three groups of patients on objective variables such as demographics, baseline comorbidities, and therapeutics as well as unknown differences related to those spiritual, psychological and physical realms. In an effort to compare ‘oranges with oranges’, the authors used propensity matching in these three groups to compare outcomes. The variables used to match patients included: age at index, sex, BMI at index, smoking status, duration of care by nephrologists, mean eGFR within 1 year before index (also accounting for the number of eGFRs available), eGFR at index, index year, emergency status of procedure, diabetes, history of cerebrovascular disease, congestive heart failure, peripheral vascular disease, dyslipidemia, history of PCI or CABG, and number of diseased coronary vessels. The outcomes were death, dialysis, or 30% reduction in kidney function at 3 and 24 months.
It is appreciated that patients with chronic kidney disease are more likely to develop acute kidney disease and/or worsening kidney function following either PCI or CABG. Some of this association may be causal. The hemodynamic consequences of cardiopulmonary bypass, the nephrotoxic effects of iodinated contrast media, or the occurrence of cholesterol emboli may all lead to acute kidney injury (AKI) with incomplete recovery of function and a further decrease in kidney function in the long term. Some of the association however represents an ascertainment bias. That is, a much smaller loss of glomerular filtration rate (GFR) is necessary to raise the serum creatinine to the threshold for AKI definition when GFR is already low 1. This is based partly on mathematics. Serum creatinine does not change linearly with GFR. To achieve a 0.3 mg/dl increase in creatinine (the threshold for AKI based upon the KDIGO criteria) requires a greater reduction in absolute GFR when you start with a creatinine of 1.0 mg/dl than when you start with a creatinine of 2.0 mg/dl.