Background
Chronic Kidney Disease (CKD) is a significant public health problem worldwide because of the poor outcomes experienced by patients and high associated costs [
1‐
3]. It is the most common chronic disease in Australia with approximately 10% of adults (1.7 million) affected according to estimates by the Australian Institute of Health and Welfare [
4] and a prevalence of 14% as per the AusDiab study [
5]. Control of CKD is confounded by its heterogeneity; of patient demographics, etiology of disease, and number and type of complications and co-morbidities. Patients in whom CKD is not controlled experience excess morbidity and mortality, especially cardiovascular mortality [
6‐
9]. Furthermore the cost of managing these poor health outcomes is high with renal replacement therapy (RRT) for end stage kidney disease (ESKD) alone costing a huge amount of money to health systems [
2]. All of the opportunity of controlling CKD, once present, lies with slowing disease progression, preventing complications, especially cardiovascular complications, and preventing ESKD. Early detection, better understanding of disease mechanisms and personalized treatment of the CKD complex are all needed. However, there are paucity of data of the interactions of demographics, etiologies, complications and co-morbidities, including their impact on the postulated pathway of a self-perpetuating vicious cycle of fibrosis independent of the initial renal injury [
10].
Furthermore, the weight of different components of the complex causal interactions is changing. Elderly people are the fastest growing population group [
11] and elderly patients are characterized by different causes of kidney disease and higher co-morbidities than younger age groups [
12]. Therapies based on evidence generated in younger age groups may therefore not be applicable to those of an older demographic. Furthermore personalized treatment of CKD in the elderly must be cognizant of likely interacting and coexisting health needs in a cohort who have accumulated deficits in multiple organs requiring further customizing of their bundle of care. Modification of drug selection and dosing is one example [
13].
We aim to describe patient, kidney and co-morbid characteristics across age groups in a large and current population of the Chronic Kidney Disease in Queensland (CKD.QLD) cohort. CKD.QLD [
14]. CKD.QLD, established in late 2009, is a statewide multidisciplinary research and practice program, conducted through a collaboration of academics and specialists in kidney care in the public health sector in Queensland. CKD.QLD encompasses a CKD registry involving all consenting patients referred to public renal practices (prior to renal replacement therapy - RRT) in Queensland, which is the third most populous Australian state/territory. CKD.QLD has identified some important differences between CKD.QLD registry patients (a pre-dialysis population) and those patients who have survived a renal death and are on RRT in Australia. Notably, the CKD.QLD cohort included older people with no gender difference, unlike the male-dominant RRT population, and has a higher percentage of patients with renal vascular disease and a lower percentage with diabetic nephropathy [
15].
Methods
This is a prospective observational cohort study of 1265 participants who were enrolled, with informed consent, in the CKD.QLD registry [
15] at the Kidney Health Service, Royal Brisbane and Women’s Hospital (Metro North Hospital and Health Service), Queensland Health. Their characteristics were documented at baseline; they were followed longitudinally and, for the purpose of this description of outcomes, censored at commencement of renal replacement therapy (RRT), death, discharge or September 2015.
“Chronic Kidney Disease in Queensland (CKD.QLD) Registry Study” has Royal Brisbane and Women’s Hospital Human Research Ethics Committee approval - HREC Reference Number (MNHHS): HREC/15/QRBW/294 as of June 2015 (Previous HREC Reference Number (OHMR): HREC/10/QHC/41 from Nov 2010 to June 2015 University of Queensland Medical Research Ethics Number: 2011000029).
Participant population
The study population is a group of 1265 patients managed by a single service provider, the Kidney Health Service (KHS), Metro North Hospital and Health Service (MNHHS), a specialty service in the publicly funded health system. The first participant gave informed consent to join the CKD.QLD cohort in May 2011. Participants were progressively enrolled at clinical consultations at different clinic locations, all under the umbrella of KHS, MNHHS. Patients were grouped according to age at consent into <35, 35–44, 45–54, 55–64, 65–74, 75–84 and 85+ years.
Inclusion and exclusion criteria
All adult patients referred to the MNHHS KHS, not on RRT, and able to give informed consent, were offered enrolment in the CKD.QLD Registry. Patients were excluded if pediatric, already on RRT, or unable to give informed consent. Rates of patients declining Registry participation were less than 1%.
Data collection and patient characteristics
Data were extracted from the clinical records of participants, both electronic and hardcopy, at the time point of consent and throughout the study period. Outcomes were sourced from a combination of these records and organisational datasets, including hospital records for each admission and ANZDATA (
http://www.anzdata.org.au) for a comparative population of RRT patients. Data extractions included 1) Demographics – age, gender, weight, height and derived body mass index (BMI) 2) cause/s of chronic kidney disease (clinical diagnosis) as determined by the treating Nephrologist; these causes were grouped into seven categories in both genders i.e. glomerulonephritis (GN), genetic renal disease (GRD), loss of single kidney, diabetic nephropathy (DN), renovascular disease including hypertension, and categories of other and uncertain. We also recorded the number of renal diagnoses in all patients, both males and females. Some participants were accorded more than one cause for CKD by their treating nephrologist. 3) Stage of CKD (1, 2, 3A, 3B, 4 and 5) at the time of consent to the registry 4) The presence of specific clinically diagnosed co-morbidities were recorded: these included coronary artery disease, cerebrovascular disease, hypertension, peripheral vascular disease, diabetes, gout, obstructive sleep apnoea, chronic lung disease and cognitive impairment/dementia 5) Outcomes measured included death, start of RRT (cause specific), hospitalisation including length of stay, causes of specific hospitalisation, and changes in eGFR over the follow-up period.
Stage of CKD was calculated by using CKD-EPI formula for estimated glomerular filtration rate (eGFR) [
16]. The outcome of death was recorded as the raw number of patients who died and the incidence of death per 100 person years of follow up was calculated. The outcome of initiation of RRT was recorded as the raw number of patients who started RRT, the incidence of RRT per 100 person years and the primary renal diagnosis in that patient. Rates of hospitalisation per 100 person years and the top 10 causes of hospitalisation were recorded by using diagnosis related group (DRGs) code set. Changes in eGFR were transformed to rate of change of eGFR over 12 months and categorised as progressive if declining by >5 ml/min/1.73m
2, improved if there was gain of >5 ml/min/1.73m
2 and stable if there was less than +/− 5 ml/min/1.73m
2 of eGFR change over at least 12 months follow up.
Discussion
Australia has excellent available data regarding End Stage Kidney Disease (ESKD) patients on RRT in the form of the ANZDATA registry [
19] but there is paucity of data regarding earlier stages of CKD before the need for RRT, or those who do not go on to RRT. Our study is one of the few providing opportunities to look at pre-terminal CKD, through examining patients referred by other health care providers (mostly in primary care) to renal specialty services, across all age groups from young adults through to very elderly. This study population consists of 1265 patients with CKD referred for evaluation and/or management to a single renal specialty service, although with multiple clinics, in a metropolitan setting (Kidney Health Service (KHS), Metro North Hospital and Health Service (MNHHS)), within the publicly funded Queensland Health system.
There were no exclusions in our study based on age or range of eGFR. This contrasts with other published studies of renal specialty-based CKD patients. Other CKD cohort studies like German Chronic Kidney Disease (GCKD) [
20], the Chronic Renal Insufficiency Cohort (CRIC) [
21] and the Chronic Kidney Disease Japan Cohort (CKD-JAC) [
22] excluded patients over 75 years of age or excluded patients with milder forms of CKD, represented by eGFR over 60 ml/min/1.73m
2 (GCKD, CKD-JAC and CKD-REIN – The French Chronic Kidney Disease Renal Epidemiology and Information Network cohort study [
23]) or over 45 ml/min/1.73m
2 (CanPREDDICT [
24] Cohort profile: Canadian study of prediction of death, dialysis and interim cardiovascular events). The International Network of CKD cohort studies (iNET-CKD) has recently been established to promote research and conduct observational studies of appropriate sample size, follow-up and defined objectives including patient level outcomes. This will facilitate comparative research and identify the differences (genetic, behavioural and health services factors) associated with the course of CKD [
25].
Additional strengths of this study include the diversity of subjects, representing the very multicultural mix of the Australian population. In contrast, the GCKD [
20] cohort had 100% Caucasian patients, CanPREDDICT [
24] had a low prevalence of non-Caucasian individuals and the CKD-JAC cohort [
22] had only Japanese patients who have lower cardiovascular disease prevalence. Our cohort also had uniform access to health resources and shared practice protocols within the entire Queensland public health service. Moreover, our study links participant demographics, clinical characteristics and laboratory data to clinically significant outcomes, while some studies e.g. Alberta Kidney Disease Network (AKDN) [
26] rely on only laboratory data to describe a study population. The comprehensive nature of our data provides a basis for international comparison and cross-validation of findings.
In this study we found significant heterogeneity by age in terms of CKD stage, etiology and number of renal disease diagnosis, comorbidities as well as outcomes of renal replacement therapy, death, rate of progression of CKD and of hospitalizations. Being sourced from specialty nephrology practices, the majority of patients (66.8%) were in more advanced stages of CKD (stage 3B, 4 and 5) at entry into the CKD.QLD cohort. However, the level of kidney dysfunction at the time of entry was dependent on age. Younger patients entered at earlier stages of CKD, with three quarters (75.2%) younger than 65 years in either Stage 1 or 2. This contrasts with only 24.7% of patients over 65 years entering at stage 1 or 2. This could represent variability in referrers practice. 63% of our patients were ≥65 years old, themselves a heterogeneous population, and 34.7% were ≥75 years old at consent. Better understanding of these individuals is needed, so efforts can be made to slow their disease progression and reduce the burden of ESKD and non-renal complications, because specific therapies to target CKD are very limited as of today, and the elderly are the fastest growing population demographic.
Not surprisingly the prevalence of vascular disease, including renovascular disease, increased from younger to elderly age groups. It was the most common attributable diagnosis in the majority of our patients (38.4%). Interestingly, diabetic nephropathy contributed to only 17.4% of CKD in our cohort in comparison to much higher prevalence in other studies [
27] and it varied with age. Older participants tended to have more than one primary renal diagnosis but a significant minority of younger participants also had multiple renal disease diagnoses. The elderly cohort also had a greater burden of co-morbid diseases, which they had presumably accumulated over the course of their lives or might have been facilitated or accelerated by CKD.
The renal function of the majority (55.9%) of participants did not deteriorate over the study interval. Whilst this may represent successful models of care, other explanations include a more benign natural history of CKD than expected: it is noteworthy that these data align with the results of CanPREDICCT study [
24]. However CKD did progress in a subgroup. This has implications for CKD prognostication to target those who are more likely to progress.
Not unexpectedly, participants in the older age groups have higher incidence of death and lower incidence of RRT, either dialysis or transplantation. The lower rates of RRT may indicate increasing provision of renal supportive/conservative care as an alternative choice to RRT in those elderly frail patients with multiple medical co-morbidities. Older patients also tended to have more hospitalizations with longer length of stay and utilized more health resources as opposed to younger patients. Tonelli et al. [
28] recently demonstrated that comorbidity in CKD patients is an important driver of adverse clinical and economic consequences.
Our study has some limitations. It is a single center study although patients are drawn from multiple hospital and community clinics. As only referred and consented patients with CKD are included in the CKD.QLD registry, therefore, we are unable to generalize our findings to the CKD patients being managed by primary care physicians with no specialist nephrology input. There may be a referral bias as elderly patients with renal impairment are less likely to be referred to specialist renal services than their younger counterparts. Furthermore there are some uncertainties in the etiology of CKD as many physicians made the diagnoses without biopsy confirmation.
Acknowledgements
The NHMRC CKD Centre of Research Excellence and Chronic Kidney Disease in Queensland Research Collaboratives’ (CKD.QLD) (
http://www.ckdqld.org)
Kidney Health Service, Metro North Hospital and Health Service (KHS MNHHS), Queensland Health
We thank all of the patients who are participating in the CKD.QLD Registry Study.