Background
The prevalence of chronic kidney disease (CKD) is increasing globally. Early diagnosis in primary care may have a role in ensuring proper intervention. We aimed to determine the prevalence and outcome of CKD in primary care.

Methods
We performed an observational cohort study in primary care in Copenhagen (2001–2015). Outcomes were stroke, myocardial infarction (MI), heart failure (HF), peripheral artery disease (PAD), all-cause- and cardiovascular mortality. We combined individuals with normal kidney function and CKD stage 2 as reference. We conducted cause-specific Cox proportional regressions to calculate the hazard ratios for outcomes according to CKD group. We explored the associations between kidney function and the outcomes examined using eGFR as a continuous variable modelled with penalised splines. All models were adjusted for age, gender, diabetes, hypertension, existing CVD, heart failure, LDL cholesterol and use of antihypertensive treatment.

Results
We included 171,133 individuals with at least two eGFR measurements of which the majority (n = 157,002) had eGFR > 60 ml/min/1.73m2 at index date, and 0.05% were in CKD stage 5. Event rates were low in eGFR > 60 ml/min/1.73m2 but increased in those with higher stages of CKD. In adjusted analyses we observed an increase in hazard rates for every outcome with every increment in CKD stage. Compared to the reference group, individuals in CKD stage 4 had double the hazard rate of PAD, MI, cardiovascular and all-cause mortality.

Conclusions
Our data from a large primary care cohort demonstrate an early increase in the risk of adverse outcomes already at CKD stage 3. This underlines the importance of studying early intervention in primary care.