Risk prediction and outcomes in chronic kidney disease stage 3: a prospective cohort study in primary care

Shardlow, Adam (2019) Risk prediction and outcomes in chronic kidney disease stage 3: a prospective cohort study in primary care. PhD thesis, University of Nottingham.

[img] PDF (Thesis - as examined) - Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (2MB)



Chronic Kidney Disease (CKD) is common, and affects many people who are predominantly managed in primary care. Guidelines for management of CKD are often based around evidence from referred secondary care populations. Many risk factors for progressive CKD have been described, but not examined in a primary care population. Outcomes in CKD are variable, and concern long-term changes. Therefore, risk stratification is key to guide clinical management. This thesis will examine outcomes over 5 years in a cohort of people with CKD stage 3 recruited from primary care. It will evaluate cystatin C, Fibroblast Growth Hormone (FGF23), vitamin D and parathyroid hormone (PTH) as biochemical markers of risk in this population, and additionally will assess Skin autofluorescence (SAF) as a non-invasive measure of tissue advanced glycation endproducts (AGEs) and another potential strategy for risk stratification.


The Renal Risk in Derby (RRID) study is a prospective cohort of 1,741 people with CKD stage 3, defined by two GFR measurements more than 90 days apart, prior to study entry. Participants were individually assessed at baseline, 1 year, and 5 year follow-up visits. Blood and urine were sampled for biochemistry, and anthropometric data were additionally collected. SAF was measured at each study visit using an AGE reader (Diagnoptics).


Five-year follow-up demonstrated low rates of CKD progression. The largest group of participants (593 of 1,741, 34.1%) evidenced stable CKD over the 5 year study period. End Stage Kidney Disease (ESKD) was rare, and was seen in 0.2% (4 of 1,741) of participants. Additionally, a significant minority of participants (19.3%, 336 of 1,741) demonstrated no evidence of CKD at their year 5 visit. The risk of CKD progression can be predicted with moderate accuracy using common clinical variables including age, gender, baseline eGFR, urine albumin to creatinine ratio (uACR), haemoglobin, bicarbonate, and diabetes.

Cystatin C was evaluated in the context of national and international guidance suggesting it be used to confirm diagnosis of CKD in those with CKD 3a A1 on creatinine based estimates of GFR. In this cohort, mean cystatin C based eGFR values were significantly lower than those derived from creatinine (45.1 ml/min/1.73m2, 95% Confidence Interval (CI) 44.4 to 45.9 versus 53.6 ml/min/1.73m2, 95% CI 53.0 to 54.1). Use of cystatin C based eGFR in this population would therefore reclassify many participants as having more severe disease.

Markers of mineral bone disease, including FGF23, vitamin D, and parathyroid hormone (PTH) have been of interest in CKD risk prediction. In the RRID study, vitamin D deficiency (Hazard Ratio 1.62, 95% CI 1.01 to 2.58) and raised PTH (HR 1.42, 95% CI 1.09 to 1.84) were independently associated with increased risk of all-cause mortality, but FGF23 was not.

Skin Autofluorescence (SAF) was examined as a risk marker of all-cause mortality and CKD progression. In fully adjusted multivariable analysis, both baseline SAF (HR 1.16, 95% CI 1.02 to 1.32), and the change in SAF over the first year of the study (HR 1.16, 95% CI 1.00 to 1.34) were associated with all-cause mortality but not CKD progression.


The results presented in this thesis show that, in a primary care cohort of predominantly older adults, CKD progression is uncommon and can be predicted using common clinical variables. Cystatin C did not improve diagnosis or risk prediction in CKD in the way that it was intended in recent national and international guidance. Some novel risk markers evaluated in this study did predict risk in this population, including SAF, vitamin D, and PTH; however, these do not necessarily contribute additionally to common clinical variables. Risk prediction models in primary care should therefore focus on these. Thus, these results suggest that people with CKD stage 3 can be adequately managed and risk stratified in primary care without the need for additional, potentially expensive, investigations.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Taal, Maarten
Keywords: Chronic kidney disease; Outcomes; Risk predicition
Subjects: W Medicine and related subjects (NLM Classification) > WJ Urogenital system
Faculties/Schools: UK Campuses > Faculty of Medicine and Health Sciences > School of Medicine
Item ID: 56430
Depositing User: Shardlow, Adam
Date Deposited: 19 Jul 2019 04:40
Last Modified: 07 May 2020 11:45
URI: https://eprints.nottingham.ac.uk/id/eprint/56430

Actions (Archive Staff Only)

Edit View Edit View