Chronic kidney disease: determining chronicity, prevalence, variation and survival in a community chronic kidney disease (CKD) cohort

So, Beng Hock (2018) Chronic kidney disease: determining chronicity, prevalence, variation and survival in a community chronic kidney disease (CKD) cohort. MD thesis, University of Glasgow.

Full text available as:
Download (2MB) | Preview
Printed Thesis Information:


Chronic kidney disease (CKD) is insidious and most cases are diagnosed through opportunistic serum creatinine (SCr) testing before symptoms develop. However, efforts to accurately assess prevalence have been hampered by the lack of a universally agreed definition of SCr thresholds for the diagnosis of CKD. At the turn of the millennium, two crucial developments occurred. The first was the description of the Modification of Diet in Renal Disease estimated Glomerular Filtration Rate (eGFR) which closely correlated to cumbersome measured GFR and could be used instead in daily clinical practice. The second was the publication of the Kidney Disease Outcomes Quality Initiative (KDOQI) clinical practice guideline for the evaluation, classification and stratification of CKD detailing a new definition of CKD based on GFR thresholds. Together, these two developments formed the basis of CKD as we know it today.
Prevalence of CKD varies, and accurate prevalence estimates are difficult to obtain especially with respect to fulfilling the chronicity criterion (reduced eGFR ≥ 90 days). Traditional risk factors for CKD are well described and non-traditional risk factors such as socio-economic status (SES), health literacy and rurality are gaining interest. SCr sampling patterns in the community mean that most individuals with CKD are tested routinely every year. This information may not be considered in its entirety by primary care providers (PCP) which may explain inaccuracies in PCP CKD registers. Accurate identification is important to direct evidence based clinical interventions to this patient group.
In chapter 2, a novel algorithm for detecting CKD and confirming chronicity from a laboratory database was developed to identify a CKD cohort of the population served by NHS Ayrshire & Arran. Data linkage of additional laboratory data, Scottish Morbidity Records for co-morbidity, statin dispensing information from Prescribing Information Scotland, area SES, rurality and deaths from Information Services Division Scotland enriched the cohort. Patients on renal replacement therapy were identified and excluded through the Scottish Renal Registry. Multiple imputations were applied where appropriate to address missing values. There were 21,037 individuals from 2010 to 2012 fulfilling the definition of CKD stage 3 – 5. Prevalence of adults with CKD was 5.6% – 5.8%. Average age (± SD) of the cohort was 75 ± 11 years. 64.6% were female and average eGFR for the cohort was 47.32 ± 11.53 mL/min/1.73m2.
In chapter 3, laboratory ascertainment of CKD identified 7% more cases than PCP CKD registers. Furthermore, around 25% of patients on PCP CKD registers may be wrongly coded as having CKD. There was a 3.9-fold variation in CKD prevalence amongst PCPs, ranging from 2.8% - 11.0%. Variation fell to 3-fold with laboratory ascertainment, ranging from 3.0% - 9.1%. This fell further with age and gender stratification. Stratified laboratory CKD prevalence was positively associated with SES and rurality, a novel finding, but in multivariate linear regression, only SES, in addition to age and gender, were significant predictors for CKD prevalence.
Chapter 4 explored the association between SES, eGFR and all-cause mortality. One-way ANOVA demonstrates a linear relationship between eGFR and SES (F (4,15078) = 2.52, p = 0.039) with a mean difference in eGFR of 0.83 mL/min/1.73m2 between the lowest and the highest SES quintile. However, linear regression modelling found proteinuria, hypertension, peripheral arterial disease, age, gender and serum albumin to be significant predictors for eGFR, but not SES. After adjustment for age and gender imbalance, survival demonstrated substantial influence by SES, but weakened in effect with full adjustment with only Scottish Index of Multiple Deprivation (SIMD) quintile 3 demonstrating a 13% increased risk (HR 1.13, 95% CI 1.03 to 1.24) with no progressive increase in risk associated with lower levels of SES.
As a quality of care marker, the dispensing of statin was examined in chapter 5. Having another diagnosis where statins are indicated, male gender, higher serum albumin, CKD stage 3B and age between 65 – 80 were associated with higher odds ratio for statin dispensing. 64% of the cohort was dispensed a statin in 2010, but the proportion fell by 5% to 58% in 2012. This fall in dispensing disproportionately affected younger and less co-morbid CKD patients who were all eligible for a statin. SES and gender did not appear to be a factor in falling dispensing rates. Average LDL levels were lower in the statin group by (mean difference) 0.78 mmol/L (95% CI 0.74 to 0.81) in 2010 and 0.93 mmol/L (0.90 to 0.97) in 2012. 37.2% of all statin prescriptions was for Simvastatin 40 mg.
Statins reduce cardiovascular events and mortality in CKD. However, in older patients typical of CKD, evidence is lacking. Chapter 6 examines survival in those dispensed a statin. Those dispensed a statin were younger, more likely to be male, had higher serum albumin and more co-morbid. After full adjustment, statin dispensing was associated with a 24% lower risk of death (HR 0.76, 95% CI 0.71 to 0.83) overall, 18% benefit for primary prevention (no prior coronary heart disease or cerebrovascular disease) (0.82, 0.74 to 0.91), 32% benefit in secondary prevention (0.68, 0.60 to 0.77), 22% benefit in younger (<76 years) CKD patients (0.78, 0.67 to 0.92) and 22% benefit in the older (≥ 76 years) CKD patients (0.78, 0.71 to 0.85) over 4.5 years follow-up. To illustrate absolute risk reduction, the number needed to treat to avoid one death for all patients is 15.8 (95% CI 12.3 to 22.2) and 12.4 (9.3 to 18.5) for older CKD patients.
This thesis demonstrates that centralised ascertainment of CKD is better at case finding, than existing PCP CKD registers. The linkage of additional, routinely collated healthcare data can develop CKD registers into a powerful tool for monitoring quality of care, efficacy of therapy and hypothesis generation which can, and should be, integrated into clinical IT systems with the appropriate information governance oversight in place.

Item Type: Thesis (MD)
Qualification Level: Doctoral
Additional Information: Chapter 3 was published in the journal NDT. Accessible here:
Keywords: Chronic kidney disease, CKD, epidemiology, prevalence, survival, Ayrshire and Arran, variation, statins.
Subjects: R Medicine > R Medicine (General)
R Medicine > RA Public aspects of medicine > RA0421 Public health. Hygiene. Preventive Medicine
Colleges/Schools: College of Medical Veterinary and Life Sciences > School of Medicine, Dentistry & Nursing
Supervisor's Name: Jardine, Professor Alan and Mark, Doctor MacGregor
Date of Award: 2018
Depositing User: Dr Beng Hock So
Unique ID: glathesis:2018-30671
Copyright: Copyright of this thesis is held by the author.
Date Deposited: 04 Jul 2018 09:04
Last Modified: 24 Aug 2018 16:17
Related URLs:

Actions (login required)

View Item View Item


Downloads per month over past year