Early referral strategies for management of people with markers of renal disease: a systematic review of the evidence of clinical effectiveness, costeffectiveness and economic analysis

Background Chronic kidney disease (CKD) is a long-term condition and has been described as the gradual loss of kidney function over time. Early in the disease process, people with CKD often experience no symptoms. For a long time, CKD has been an underdiagnosed condition. Even in the absence of symptoms, CKD appears to add significantly to the burden of cardiovascular disease and death and, for an important minority, can progress to kidney failure. Objective To systematically review the evidence of the clinical effectiveness and cost-effectiveness of early referral strategies for management of people with markers of renal disease. Data sources Electronic searches of 12 major databases (such as MEDLINE, EMBASE, CINAHL, etc.) were conducted for the time period of 1990 to April 2008 to identify studies comparing early referral to other care options for people with CKD. Additional searching was performed in the NHS Economic Evaluation Database to support the cost-effectiveness literature review. Review methods Two authors reviewed all titles, abstracts and full papers to select relevant literature. A Markov model was constructed to represent the natural history of CKD. The model allowed cohorts to be tracked according to estimated glomerular filtration rate (eGFR) status and the presence of other complications known to influence CKD progression and the incidence of cardiovascular events. Results From 36 relevant natural history studies, CKD was found to be, despite marked heterogeneity between studies, a marker of increased risk of mortality, renal progression and end-stage renal disease. Mortality was generally high and increased with stage of CKD. After adjustment for comorbidities, the relative risk of mortality among those with CKD identified from the general population increased with stage. For clinical populations, the relative risk was higher. All three outcomes increased as eGFR fell. Only seven studies, and no randomised controlled trials, were identified as relevant to assessing the clinical effectiveness of early referral strategies for CKD. In the five retrospective studies constructed from cohorts starting on renal replacement therapy (RRT), mortality was reduced in the early referral group (more than 12 months prior to RRT) even as late as 5 years after initiation of RRT. Only two studies included predialysis participants. One study, in people screened for diabetic nephropathy, reported a reduction in the decline in renal function associated with early referral to nephrology specialists (eGFR decline 3.4 ml/min/1.73 m(2)) when compared with a similar group that had no access to nephrology services until dialysis was required (eGFR decline 12.0 ml/min/1.73 m(2)). The second study, among a group of veterans with two creatinine levels of at least 140 mg/dl, reported that a composite end point of death or progression was lower in the group receiving nephrology follow-up than in those receiving only primary care follow-up. The greatest effect was observed in those with stage 3 or worse disease after adjustment for comorbidities, age, race, smoking and proteinuria {stage 3: hazard ratio (HR) 0.8 [95% confidence interval (CI) 0.61 to 0.9)]; stage 4: HR 0.75 (95% CI 0.45 to 0.89)}. In the base-case analysis, all early referral strategies produced more quality-adjusted life-years (QALYs) than referral upon transit to stage 5 CKD (eGFR 15 ml/min/1.73 m(2)). Referral for everyone with an eGFR below 60 ml/min/1.73 m(2) (stage 3a CKD) generated the most QALYs and, compared with referral for stage 4 CKD (eGFR < 30 ml/min/1.73 m(2)), had an incremental cost-effectiveness ratio of approximately 3806 pounds per QALY. Limitations Because of a lack of data on the natural history of CKD in individuals without diabetes, and a lack of evidence on the costs and effects of early referral, the Markov model relied on many assumptions. The findings were particularly sensitive to changes in eGFR decline rates and the relative effect of early referral on CKD progression and cardiovascular events; the latter parameter being derived from a single non-randomised study. Conclusions Despite substantial focus on the early identification and proactive management of CKD in the last few years, we have identified significant evidence gaps about how best to manage people with CKD. There was some evidence to suggest that the care of people with CKD could be improved and, because these people are at risk from both renal and cardiovascular outcomes, strategies to improve the management of people with CKD have the potential to offer an efficient use of health service resources. Given the number of people now being recognised as having markers of kidney impairment, there is an urgent need for further research to support service change