Ann Intern Med. 2023 Jun;176(6):788-797. doi: 10.7326/M22-3228. Epub 2023 May 23.
Population-Wide Screening for Chronic Kidney Disease: A Cost-Effectiveness Analysis
Marika M Cusick, Rebecca L Tisdale, Glenn M Chertow, Douglas K Owens, Jeremy D Goldhaber-Fiebert
PMID: 37216661
JAMA Health Forum. 2024 Nov 1;5(11):e243892. doi: 10.1001/jamahealthforum.2024.3892.
Population wide Screening for Chronic Kidney Disease: A Cost-Effectiveness Analysis
Marika M Cusick, Rebecca L Tisdale, Glenn M Chertow, Douglas K Owens, Jeremy D Goldhaber-Fiebert, Joshua A Salomon
PMID: 39514193
Why was this study done?
Should we screen for kidney disease? As fellow nephron readers, preserving every glomerulus (and tubule) is worthwhile, so of course, one should detect kidney disease as early as possible and prevent it, right? But the simple answer may not always be right. We discussed an aspirational study of screening as our first #NephJC (Kurella Tamura et al, Kidney Int 2014 | Summary), but the evidence has been pretty weak sauce. We need to have effective therapies to slow down the progression, and until recently, we only had RASi, that also in proteinuric settings. Screening the whole population is expensive, especially when the marginal benefit in capturing a small number of individuals, for whom all one can offer is lisinopril or losartan, is quite underwhelming. Hence, past studies have not demonstrated the benefit of CKD screening (Manns et al, BMJ 2010; USPSTF statement Moyer et al, Annals IM 2012).
But since 2012, the renal therapeutic space has broadened considerably. We have flozins. We have finerenone. The GLP1RAs are flowing into the nephrology world. We have new treatments for the most common GN, IgAN. Will this change the screening calculus in favor of screening?
How was the study done, and what did it show?
These are actually two studies - the first one was the base model looking at a fixed screening at ages 35 - 45 years, and the second one looked at multiple different decades (35, 45, 55, 65, or 75 years). This was otherwise a standard cost-effectiveness analysis. A Markov cohort model was created (see schematic below) with certain assumptions.
Screening was for albuminuria (UACR > 29 mg/g; $49), then creatinine for calculating GFR ($23). Those with GFR < 60 would get an ultrasound ($420) as well. The interventions would be either RASi alone ($14 a month; slowing down CKD progression) or RASi + flozination ($14 + $365 a month; which would slow down CKD progression even more and lengthen survival). Adverse effects, including genital infections ($153), euglycemic DKA ($31,304), and angioedema ($3879) were also built in. Developing CKD also lowers quality of life (mathematically computed at ~ 0.74 to 0.85, with 1.0 being perfect health; dialysis at 0.60). The monthly cost of having CKD was thus estimated at ~ $136 for CKD 3A, up to $1069 for CKD 4, and ballooning to $7018 once on RRT. One can appreciate how preventing kidney failure requiring dialysis can save money, counterbalanced on the amount spent on screening and purchasing RASi/flozins per month. The NHANES dataset was used to build the study population, with estimates from DAPA CKD (NephJC summary) for the flozination effects.
Short story of results: flozination adds on ~ 0.11 QALY, at a ~ $9000 extra cost, thus providing a $86,300 spent per QALY gained calculus. Screening more often does give you more benefit, at a higher cost - see how it breaks down in the figure below.
Figure 2 from Cusick et al, Annals IM 2023
To see how the benefits alone accrue, figure 1 shows it quite nicely below, for RRT on the left, and survival on the right. The blue bar is the status quo (assuming sporadic treatment with RASi), and the other bars show how this changes with active case detection and treatment.
Figure 1 from Cusick et al, Annals IM 2023
Next, in the JAMA Health Forum - everything else remaining the same, the authors report the effect of screening every 5 or 10 years, and starting at different ages. The earlier one starts and the more frequently you screen, the more benefit one accrues - but with a higher cost, best exemplified in the table/figure below.
Figure 2 from Cusick et al, JAMA Health Forum 2024.
What are the implications?
Compared to the cynical nihilism of 2 decades ago, we now do have extremely effective agents making screening a viable proposition. Clearly, from a nephron-preserving standpoint (and not a palliative dialysis-ologist attitude), detecting CKD, especially proteinuric CKD, early is better. But the cost at a population level is something to consider, especially if we consider that every dollar spent on nephron preservation is one less dollar spent on cancer detection, or HIV treatment, vaccination, or even hunger prevention. This study is useful for healthcare stewards and policy people to use numbers to compare across programs. Other studies from non-US perspective also agree with this overall results (eg Pouwels et al EClin Med 2024 from Netherlands; Wen et al Lancet RHWP 2025 from China). Even KDIGO has weighed in as far back as 2019 (Shlipak et al Kidney Int 2019).
Some people might still balk at the costs. Could there be ways to make these lower? Surely there are tweaks that might be more ‘cost-effective’? As an aside, cost effectiveness was traditionally established as $50,000 based on the cost of dialysis (i.e., spending up to $50,000 more for one QALY is justified from a societal perspective, see Manns et al, Health Econ 2003). People have argued this threshold should be updated (~ $128,000 per QALY according to Lee et al, Value in Health 2009). Many of the strategies would be deemed ‘cost-effective’ from that threshold. But one could tweak this to something along the lines of: screen everyone once at age 35, and screen those who have diabetes every 5 years? Variations like this are bound to be less expensive than screening everyone every 5 years - but are less likely to be diligently followed as well.
The other wrinkle will be the cost of drugs. The cost of flozination was modelled at $365 a month. In Canada this is already C$20 a month. More useful is if we could make GFR estimation and UACR cheaper than ~ $25 and $50 a pop, respectively. On the other hand, we have other medications coming online. GLP1RAs are a different factor of cost - and may not come down in cost as rapidly as flozins have, though we hope the incoming small oral molecules (e.g., orfoglipron) will be cheaper than the existing options. Many of the IgAN therapies require you to sell your house for a few months of therapy, and all the savings in dialysis will not make such strategies cost-effective. Clearly, there is work to be done. But these are exciting times in nephrology - time to move on from talking about dialysis planning to dialysis prevention.
Summary by Swapnil Hiremath
Reviewed by Cristina Popa and Brian Rifkin