#NephJC Chat
Tuesday July, 9th, 2024 at 9 pm Eastern (AEST = July 10th, 11am)
Wednesday July 10th, 2024, at 9 pm Indian Standard by Time and 3:30 pm GMT (AEST = July 11th, 2am)
Lancet 2024 Mar 30;403(10433):1279-1289. doi: 10.1016/S0140-6736(23)02843-X. Epub 2024 Mar 13.
Effects of rare kidney diseases on kidney failure: a longitudinal analysis of the UK National Registry of Rare Kidney Diseases (RaDaR) cohort
Wong K, Pitcher D, Braddon F, Downward L, Steenkamp R, Annear N, Barratt J, Bingham C, Chrysochou C, Coward RJ, Game D, Griffin S, Hall M, Johnson S, Kanigicherla D, Karet Frankl F, Kavanagh D, Kerecuk L, Maher ER, Moochhala S, Pinney J, Sayer JA, Simms R, Sinha S, Srivastava S, Tam FWK, Turner AN, Walsh SB, Waters A, Wilson P, Wong E, Taylor CM, Nitsch D, Saleem M, Bockenhauer D, Bramham K, Gale DP; RaDaR consortium.
PMID: 38492578
Introduction
Part of the attraction of many physicians towards nephrology is the dichotomy of care: inpatient versus outpatient, acute versus chronic, and some of the most common versus some of the rarest diseases. Perhaps because of this diversity of disease management, we are also prone to lumping patients into one basket, for example, “ESKD on dialysis”. We all know from experience, however, that many elements interact (including, but not limited to, comorbidities, genetics, etiology of kidney decline, and overall frailty) to determine the course of disease in patients with CKD and ESKD. The current study used the national UK registry of rare diseases to gather longitudinal data and study disease progression, kidney failure, and death in patients with rare diseases versus patients with more common causes of kidney disease.
How rare does a disease have to be to be called rare? In the renal world, this includes over a hundred kidney diseases with an incidence of <5 per 10,000 people in Europe (Moliner AM, Waligora J, Adv Exp Med Biol. 2017) or <200,000 diagnosed patients in the USA (Schieppati A et. al. Lancet, 2008) - which at the current US population (~ 330 million) would be about 6 per 10,000 people. These may be cases that we diagnose and treat many times a year to once (or possibly never) in a career. Rare kidney diseases include both sporadic and inherited diseases, listed in the image below. The graphic is not exhaustive, and the list at this ERKNet link (European Reference Network for Rare Kidney Diseases) has more than 350 diseases. Using the definition provided above, even the ‘most common glomerular disease’ makes an appearance in the list of rare diseases as seen below.
Figure from Devuyst O et. al. Lancet 2014.
However, does ‘rare’ mean these individuals rarely end up with kidney failure? Not really; we already knew the importance of rare kidney diseases leading to kidney failure, from the 2014 ERA-EDTA registry study. About 12% of adults and 58% of children on RRT had an etiology classified as a rare disease (Wühl E et al. Nephrol Dial Transplant, 2014). There is much more to know about rare kidney disease beyond kidney failure incidence. Clinicians and researchers alike are interested in the age of diagnosis, rate of eGFR decline, time from diagnosis to RRT, and life expectancy on RRT.
A KDIGO conference report on rare kidney diseases noted many challenges and proposed cohort studies to monitor subsequent health outcomes (Aymé S et. al. Kidney International 2017). Complete data is often unavailable due to the small numbers of affected patients, unidentified causes of disease, and a lack of biomarkers for monitoring disease progression. In addition, improved diagnostics, management of kidney functional decline, translation of advances in research to clinical care, and integrated patient support are needed to address patients’ needs. Let’s examine how the UK based RaDaR (Rare kidney Diseases Registry) helps target some of these questions.
The Study
Methods
The RaDaR database was established in 2010 by the UK Kidney Association, and has linkages with hospital labs and with the UK renal registry for RRT initiation and death outcomes. It is “the largest rare kidney disease registry in the world” (check it out here: https://rarerenal.org). The present study was thus a retrospective cohort design linked to metadata including clinical and laboratory information from over 30,000 patients. The registry began in Jan 2010 and recruited participants until July 2022, with a follow up time defined as from the date of diagnosis until July 25, 2022. They examined all patients with an age range of “a lifetime” (really, from 0 to 96 years)
RaDaR eligibility criteria varied among participants based on the disease (details on page 27 onwards in supplement). Depending on the disease, the authors took into account the gold standard for the diagnosis of each condition. The exclusion criteria were defined by secondary or idiopathic causes of kidney damage or a spectrum of the disease without kidney involvement. Thus diagnosis was based on clinical diagnosis alone for some, while for IgA nephropathy one needed a biopsy + proteinuria > 0.5 g/day or GFR < 60. Ten other rare diseases (included in the registry) were excluded from this study because of either a small sample size, or not being a primary rare renal disease diagnosis. Among the excluded diagnoses were calciphylaxis/ calcific uremic arteriopathy, Fabry disease, and mitochondrial renal disease.
Outcomes
The primary outcomes were cumulative incidences of death and kidney failure. Kidney failure was defined as the need for kidney replacement therapy (dialysis or transplantation) or an eGFRcr CKD-EPI without adjustment for race <15 ml/min for >4 weeks (or by Schwartz equation for those ≤16 years).
Secondary outcomes
Median age at kidney failure
Median age at death
Time from start of dialysis to death
Time from diagnosis to eGFR thresholds of the therapeutic trial window from last eGFR >75 ml/min to <30 ml/min (which is considered therapeutic trial window)
The authors compared the survival on kidney replacement therapy of this population versus those with a primary renal diagnosis of diabetes, hypertension or renovascular disease. The estimations of CKD prevalence and incidence of KRT were taken from the National CKD Audit and the UK Renal Registry annual report.
Funding
The Medical Research Council, Kidney Research UK, Kidney Care UK, and the Polycystic Kidney Disease Charity funded this study. The funders did not have a role in study design, data collection, analysis, interpretation or writing.
Statistical analysis
Cumulative incidence analysis of time to kidney failure and median age at kidney failure were performed, and the censoring of deceased patients and sensitivity analysis were calculated, with death being a competing risk. Another censored result was transplantation rates because its rates differ between groups and survival rates also vary depending on the KRT received. Analysis was performed by rare disease group and sex.
‘Time in therapeutic trial window’ was defined as time from the last eGFR ≥75 to first eGFR <30 ml/min without subsequent higher eGFR values.
The 2 year kidney failure risk (KFRE; Tangri et al JAMA 2011) at RaDaR recruitment was calculated using a recalibrated four-variable kidney failure risk equation and was compared with the already observed kidney failure to assess equation accuracy in rare kidney diseases. Mortality data were obtained from the Office of National Statistics. The RaDAR population was indirectly age and sex-standardized to match the general population of England and Wales. Adjusted hazard ratios (aHR) for all-cause mortality were calculated based on a large population-based study, with adjustment for eGFR at recruitment. These ratios were compared to those of the general CKD population.
Results
The study population consisted of 27,285 participants with 27,294 diagnoses of rare kidney diseases: ADPKD and IgAN being the most frequent, with Bartter syndrome being the least frequent. A total of 28 types of rare kidney diseases were included. As can be seen in the figure below, the N for each analytic outcome varied somewhat as appropriate.
Supp. Figure 1. Flow diagram of included patients, from Wong et al, Lancet, 2024.
The median age at diagnosis for the entire RaDar population was approximately 40 years (IQR 23.7–57.1 years), with variations across different groups of rare diseases. The median follow-up was 9.6 years (IQR 5.9 to 16.7 years), and a variation of median time from age of recruitment, diagnosis to kidney failure, or death between the diseases can be seen below. Very high cumulative kidney failure events were seen in certain diseases (e.g. cystinosis and anti-GBM both at > 80%), and were not unexpected. On the other hand, the not-insignificant event risk in other diseases (e.g. 30% in X-linked Alport, 16% in thin basement membrane, and a whopping 60% in IgA) is quite eye-opening. There is a lot more to unpack in the extremely informative table 1, reproduced below.
Table 1. Characteristics of participants and primary outcome. Wong et al, Lancet, 2024.
Primary outcomes
There is significant heterogeneity in age of kidney failure between disease groups. In patients with cystinosis, kidney failure was reached at a median age of 15.4 years (IQR 11.6 to 19.9 years). In contrast, a median age of 65 years was seen for kidney failure in diseases like vasculitis, HNF1B mutations, thin basement membrane nephropathy (TBMN), monoclonal gammopathy of renal significance (MGRS), and membranous nephropathy.
Supplementary Figure 2. Cumulative incidence of kidney failure. a) Glomerular, b) Cystic, c) Metabolic, d) Tubular, e) Other kidney conditions, f) Alport Syndrome. From Wong et al, Lancet, 2024.
Similar results were seen in the sensitivity analyses for age at kidney failure and time from diagnosis with death as a competing risk.
The median age at death was either not evaluable (due to few events), or over 75 years for most groups. The one exception was for cystinosis (median age 56.4 years, 95% CI 40.9–not evaluable).
Secondary outcomes
These are outlined in table 2 in great detail (below).
Table 2. Secondary outcomes. Wong et al, Lancet, 2024.
Kidney function at diagnosis
To no one’s surprise, severity of kidney function does vary depending on kidney disease etiology. eGFR at time of diagnosis, was >90 ml/min in patients with thin basement membrane disease and tuberous sclerosis complex. Contrastingly, patients with atypical HUS and anti-GBM disease had a significantly decreased kidney function at diagnosis, with a median eGFR of only 29 ml/min and 11 ml/min respectively. Similar observations were made with respect to the highly variable median time from diagnosis to kidney failure. For example, a median time of 61 years was observed for tuberous sclerosis complex versus a paltry 0.1 years for anti-GBM disease.
Therapeutic trial window
This concept (time in therapeutic trial window) tries to encapsulate how long we have to do an intervention - or complete RCTs for a particular disease. It ranged from 20 years for retroperitoneal fibrosis to 1 year in monoclonal gammopathy of renal significance. Shorter times in the therapeutic trial window could reflect either a steeper eGFR slope or a lower eGFR at diagnosis. For example, the median time in the therapeutic window was ~ 11 years for ADPKD, but only a surprisingly short 4 years for IgA Nephropathy.
Mortality and kidney failure incidence
The reported mortality of people with rare kidney diseases was more than twice the English and Welsh population with a standardized mortality ratio of 2.29 (95% CI, 2.09 to 2.49). In those with kidney failure the standardized mortality ratio was understandably even higher at 3.39 (95% CI, 3.55 to 4.42). On the other hand, comparing the RaDaR population versus people with all-cause CKD (e.g. including diabetic nephropathy), the standardized mortality ratio was only 0.42 (95% CI, 0.32 to 0.52). This points to the fact that unlike common causes of CKD (like hypertension and diabetes that significantly add to cardiovascular mortality), rare diseases may have a lower mortality risk.
Table 3 Observed and Expected mortality, Wong et al, Lancet, 2024.
Among RaDaR participants who were not already on RRT at recruitment, the cumulative incidence of kidney failure (up to 5 years) was higher compared to UK patients with all-cause CKD (28% vs 1%). The higher risk of KRT might be attributed to rare diseases being linked to direct renal pathophysiologic processes, versus more indirect (systemic) damage of other common diseases that may occur over a longer timeline.
Supp. Table 2. Cumulative incidence of kidney failure. Wong et al, Lancet, 2024.
Predictive accuracy with KFRE
Observed rates of kidney failure in the RaDaR population exceeded predictions by the four-variable kidney failure risk equation overall (p<0.0001).
Figure 3. Observed vs predicted kidney failure within 2 years in RaDaR patients. Wong et al, Lancet, 2024.
This was true in all age groups, except in patients older than 65 years.
Supplementary Figure 8. Calibration plot for 2-year 4 variable Kidney Failure Risk Equation (KFRE) for RaDaR patients stratified by age group, from Wong et al, Lancet, 2024
Gender differences
Men had a higher median age at diagnosis pointedly in ADPKD (40 vs 36 years; p<0.0001), and IgA nephropathy (42 vs 38 years; p<0.0001), though it came with a lower eGFR such that age at kidney failure was similar between the sexes. A case worth mentioning is X-linked Alport syndrome. Male patients were younger at diagnosis (19 vs 26 years; p= 0.015), they were younger at time of KRT (29 vs 64 years;) and with a shorter time from diagnosis to kidney failure (15 vs 40 years). Apart from such notable exceptions, there were no significant differences by sex in age at kidney failure.
Supplementary Figure 12. Sex stratified Kaplan Meier Survival analyses of years from dialysis start to death (not censored for transplant)
a) RaDaR patients (Recruited) b) Patients eligible for RaDaR in the UKRR but not yet recruited (Eligible) c) Patients with a Primary Renal Disease of Diabetes, Renovascular Disease or Hypertension (Other PRD). From Wong et al, Lancet, 2024.
Discussion
Nephrologists often contemplate the uniqueness of finding a rare disease, a similar sentiment captured by the old Latin phrase “Rara avis in terris, nigroque simillima cygno” (a rare bird in the world, very similar to the black swan). Unlike the rarity of a black swan, the diagnosis of a rare disease comes with a sense of isolation for both the nephrologist and the patient. The comprehensive RaDaR study helps nephrologists navigate into the unknown territory of rare kidney diseases, by including a very comprehensive database evaluation of 28 rare kidney diseases. The large numbers, the long follow-up (median of 9.6 years), and the diversity of rare diseases included, provide a reliable basis for understanding the natural history, prognosis, and outcomes of many less rare as well as obscure diagnoses. The study provides precise estimates of kidney function decline, kidney failure, and risk of death. By comparing these estimates with those from the general population with CKD, the study highlights the distinct trajectory of rare kidney diseases, emphasizing the need for tailored mitigation strategies.
Limitations
However, one of the most important limitations of the RaDaR study's findings is that unfortunately they may not be generalizable beyond the UK due to differences in clinical practices, ethnic composition, and genetic backgrounds. Additionally, recruitment criteria for certain disease groups, such as IgA nephropathy, may favor the inclusion of patients with more severe disease, potentially skewing findings. Indirect comparison with the UK’s national data sources, such as the Office of National Statistics, might have increased the potential bias, affecting some estimates.
Due to the length of follow-up and potential survivor bias, with limited data on age at death, could also possibly influence the accuracy of mortality estimates. Next, the study combines data from various subcategories of diseases, which might obscure meaningful nuances. While the RaDaR cohort provides an excellent overview, more detailed separate analyses of individual diseases are needed. Finally, the study focused on kidney failure and mortality outcomes, but it didn’t account for extra-renal manifestations that can cause significant morbidity.
The study’s findings of higher survival and faster disease progression in rare kidney disease compared to the general CKD population are new. The study confirms that rare kidney diseases tend to result in kidney failure rather than premature death, unlike common causes of CKD, where cardiovascular mortality is more common.
RaDaR is the biggest cohort so far on rare kidney diseases, with over twenty-five thousand patients being followed for almost 10 years. It creates a new reservoir of information to differentiate between CKD etiologies, and assist in individualizing the management of our patients. When scanning for a cause for CKD, and ultimately discussing prognosis, we are obliged to see the tree, and not just the forest.
Conclusion
We now have evidence that rare kidney diseases have a higher survival rate on KRT than other common causes of CKD. Unfortunately, rare diseases are also linked with higher probabilities of rapid eGFR decline and kidney failure. Although these findings are not generalizable to all CKD cohorts throughout the world, the more we can learn about rare diseases, the better we can serve our unique patients with kidney disease.
Summary prepared by
Samantha Verdugo
NSMC 2024 intern
Guadalajara, Jalisco, Mexico
Reviewed by Brian Rifkin,Cristina Popa, Husam Alzayer, and Swapnil Hiremath