Early Detection of CKD: Implications for Low-Income, Middle-Income, and High-Income Countries

Abstract

CKD is common, costly, and associated with adverse health outcomes. Because inexpensive treatments can slow the rate of kidney function loss, and because CKD is asymptomatic until its later stages, the idea of early detection of CKD to improve outcomes ignites enthusiasm, especially in low- and middle-income countries where renal replacement is often unavailable or unaffordable. Available data and prior experience suggest that the benefits of population-based screening for CKD are uncertain; that there is potential for harms; that screening is not a wise use of resources, even in high-income countries; and that screening has substantial opportunity costs in low- and middle-income countries that offset its hypothesized benefits. In contrast, some of the factors that diminish the value of population-based screening (such as markedly higher prevalence of CKD in people with diabetes, hypertension, and cardiovascular disease, as well as high preexisting use of kidney testing in such patients) substantially increase the appeal of searching for CKD in people with known kidney risk factors (case finding) in high-income countries as well as in low- and middle-income countries. For both screening and case finding, detection of new cases is the easiest component; the real challenge is ensuring appropriate management for a chronic disease, usually for years or even decades. This review compares and contrasts the benefits, harms, and opportunity costs associated with these two approaches to early detection of CKD. We also suggest criteria (discussed separately for high-income countries and for low- and middle-income countries) to use in assessing when countries should consider case finding versus when they should consider foregoing systematic attempts at early detection and focus on management of known cases.

“All screening programs do harm. Some do good as well and of these, some do more good than harm at reasonable cost.”¹

CKD is a common and costly condition that is associated with substantial morbidity and mortality.² As for most noncommunicable chronic diseases, the prevalence and burden of CKD are increasing most rapidly in low- and middle-income countries (LMICs).^3,4 The corresponding increase in the burden of kidney failure is a major challenge for health systems worldwide because of growing demand for expensive RRTs (dialysis and kidney transplantation). Unfortunately, renal replacement is unavailable or unaffordable for the large majority of people who live in LMICs,⁵ and thus CKD that progresses to kidney failure kills millions of people each year.⁶ Because inexpensive treatments can slow the rate of kidney function loss, and because CKD is asymptomatic until its later stages, there is some enthusiasm for population-based screening to enable early intervention in both LMICs and high-income countries (HICs).

History teaches that stakeholders often attribute unrealistic benefits to screening for “their” disease, as exemplified by screening mammography (the American College of Radiology and the American College of Surgeons), screening for prostate-specific antigen levels (the American Urological Association and Prostate Cancer Canada), and screening for mental illness.^7–9 Unfortunately, many screening programs are markedly less beneficial than expected, and the situation for CKD is no different. Available data suggest that the benefits of population-based screening for CKD are uncertain¹⁰ and that screening is not a wise use of resources even in HICs.^11,12 In addition, experience with other diseases suggests that there is some potential for harms associated with CKD screening, and that screening has substantial opportunity costs in LMICs that offset its hypothesized benefits.^10,13

In contrast, some of the factors that diminish the value of population-based screening (such as markedly higher prevalence of CKD in people with diabetes, hypertension, and cardiovascular disease, as well as high preexisting use of kidney testing in such patients) substantially increase the appeal of searching for CKD in people with known kidney risk factors, or case finding, in HICs as well as in LMICs.

This review compares and contrasts the benefits, harms, and opportunity costs associated with these two approaches to early detection of CKD. We also suggest criteria for assessing when countries should consider population-based screening or case finding compared with foregoing systematic attempts at early detection and focusing on management of known cases. The suggested criteria are discussed separately for HICs and LMICs.

Early Detection Strategies: Screening and Case Finding

Screening is a sorting process that is designed to find a few people who have higher risk of disease, among a group who currently believe themselves to be well.¹ Positive screening tests increase the likelihood that an individual has the disease, but further testing is needed to confirm the diagnosis. However, simply identifying disease is futile unless it is linked to actions that improve clinical outcomes, ideally as part of a systematic program. Thus, early detection per se is simply one part of a program that must be based on sound evidence and performed with high quality; otherwise, there is risk of causing more harm than good.

In assessing the potential benefits of early detection, three specific forms of bias must be considered: length bias, lead-time bias, and—often—volunteer bias.¹⁴ These biases, which tend to exaggerate the apparent benefits of early detection, are typically used to describe issues in cancer screening, but are also applicable to the detection of CKD. Only randomized trials can fully eliminate the effects of these biases.

Alternatively, rather than searching through the whole population (population-based screening), the sorting process may focus on defined groups of people with a particular condition or set of diseases (case finding).¹⁵ In case finding, the target group’s risk of developing an unwanted outcome is much higher than it is among the general population, and it can facilitate prompt recognition of clinical problems in patients who do not recognize a disease’s early signs and symptoms. A program of case finding must also work to high standards, because it can also lead to harm.

Because early detection programs search for a relatively uncommon outcome that will occur in the future, a relatively small proportion of participants are likely to benefit from the process. Consequently, even a small risk of harm from testing and its consequent follow-up (which occur in the present) may potentially outweigh the benefits that may accrue at a future time. Testing may identify many patients who have less severe forms of disease and who would experience a good outcome if the condition had never been detected. Perhaps the best example of inadvisable screening was a program of urine testing for catecholamines in infants, intended to detect infantile neuroblastoma.¹ The program found many infants with high levels of catecholamines, some of whom had tumors that were treated with surgery and chemotherapy. However, although the program increased the total number of cases detected, it detected many tumors that would have regressed if left alone and it did not reduce the risk of death from the clinically relevant tumors. This experience shows that, even when screening appears appropriate, real-world programs may reveal unanticipated problems identified by a better understanding of the target condition’s natural history.

Cervical screening by cytology or testing for human papillomavirus is an effective program among women >30 years but, among younger women, it identifies many abnormalities that would resolve spontaneously, with no benefit from the treatment offered.¹⁶ Harm from treatment-related cervical damage is uncommon but potentially serious. In addition, although the optimal interval for cervical testing has been subject to debate, annual cytology testing adds very little marginal benefit over testing every third year and increases the risk of false positive test results. These examples show that even effective screening must be managed carefully.

An example of useful case finding is routine ophthalmoscopy for people with diabetes, to detect retinopathy and offer early treatment. However, just as for screening, case finding may cause harm, consume resources, or both, without helping patients. For example, diabetic retinopathy is routinely identified and treated although not all cases will progress to cause serious loss of vision and, although the interventions used to treat identified cases are relatively safe, adverse events can still occur. Therefore, the costs and consequences of case finding should also be evaluated, especially when the natural history of disease changes over time, as for type 1 diabetes.¹⁷

Both screening and case finding lead to overdiagnosis. This occurs when asymptomatic people are correctly diagnosed with disease and enter the management pathway, but do not benefit from it. This includes people whose CKD does not progress or progresses too slowly to cause adverse consequences before death from an unrelated cause.¹⁸ Thus for some older people, early detection of CKD will not improve their prognosis, because their likelihood of living long enough to develop kidney failure is low whether CKD is detected or not.

The Principles for Evaluating the Merits of Screening

The venerable principles for screening published by Wilson and Jungner¹⁹ in 1968 are often the starting place for assessing the merits of early detection programs. These principles are often used simplistically as a checklist or score sheet, in which having a majority of positive answers is sufficient. However, each principle requires complex decision making rather than a binary answer, and the benefit of early detection relies on a chain that is only as strong as the weakest link. Since Wilson and Jungner’s publication, additional factors have emerged as critical for assessing the merits of screening: the importance of measuring all outcomes (including harms), the importance of integrating early detection into the overall health system to ensure continuity of care, and the importance of quality assurance. Consequently, many authors have developed newer versions of principles for screening, summarized by Dobrow et al,²⁰ who created a list of 12 principles using a Delphi process among experts. Their list includes most of the Wilson and Jungner principles and adds several more (Table 1). Two additional criteria that could also be considered include whether an alternative strategy to reduce the burden of disease exists and could be used instead, and whether the benefits of screening are supported by high-quality evidence from randomized controlled trials.²¹

Table 1.

Consolidated principles for population-based screening

Principle	HICs	LMICs
Disease/condition
1. Epidemiology of the disease must be understood, and it must be an important health problem	++++	++++
	The burden of CKD is unquestionably high and increasing worldwide. Much is known about the epidemiology of CKD and about risk factors for progressive kidney function loss.	The burden of CKD is high in LMICs, and the potential risks associated with kidney failure are potentially higher because RRT may not be available.
2. Natural history should be understood, and a preclinical phase must be detectable	++++	+++
		Although there may be special populations for which less is known about the natural history, this criterion is generally met.
3. Target population for screening must be defined	+++	+++
	There are some uncertainties about the precise criteria that should be used. For example, should the target population be defined by age >18 yr or >40 yr? Should there be an upper limit or exclusions of certain groups?
Test/intervention principles
4. Screening test performance characteristics: test(s) should be accurate, safe, acceptable, and affordable	++++	++++
	Test performance and costs are generally favorable.	There may be opportunities to reduce cost by using dipstick urinalysis rather than ACRs to detect cases, but this might increase false positive rates.
5. Interpretation of screening test results: tests should have clear thresholds	++ ++
	There are some uncertainties about which criteria should define a positive screening test. For example, should eGFR 45–59.9 ml/min per 1.73 m2 but without albuminuria be considered a positive or a negative test? Still, this criterion is generally met.
6. Postscreening test options: there should be an agreed-upon course of action for follow-up, treatment, and improved outcomes	++	+
	For people with diabetes, hypertension, and perhaps vascular disease, post-test management for newly identified cases of CKD will differ little from those without CKD.	Same as for HICs. Also, some treatments that might be prescribed to CKD cases in HICs (e.g., sodium-glucose cotransporter 2 inhibitors for people with concomitant diabetes) are unaffordable in LMICs.
Program/system principles
7. Screening program infrastructure: there should be adequate existing resources or a plan to develop sufficient resources for all	+++	+
	Population-based screening would require substantial new resources. Depending on the target population, there may be opportunities to integrate CKD screening with screening for other noncommunicable diseases.	Infrastructure needed for testing alone is not available in most LMICs. Also, management of detected cases over years or decades would be difficult or impossible in most LMICs.
8. Screening program coordination and integration: screening should be coordinated and integrated into broader health system	+++	+
	This would be generally possible in HICs. Avoiding overtesting is a potential concern, as well as overinvestigation (unnecessary follow-up of positive screening tests).	Few LMICs would be able to integrate CKD cases identified by screening into the broader health system. Many LMICs struggle to manage known CKD cases.
9. Screening program acceptability and ethics: all components should be ethically acceptable to participants and professionals, and methods to ensure informed choice should be in place	++++	++
	No concerns.	Labeling people as having CKD without providing effective management causes harm.
10. Screening program benefits and harms: benefits such as increased function and quality of life or decreased mortality should be greater than harms (such as from overdiagnosis and overtreatment)	++	+
	In the absence of high-quality evidence demonstrating benefit, it is unknown whether this criterion is met. Given the comments for criteria 6 and 9, there is more uncertainty about a favorable benefits/harms ratio in LMICs than in HICs.
11. Economic evaluation of screening program: economic evaluation should assess full costs of operating screening program, compared with opportunity costs of allocating resources to alternatives	+	+
	Even making very favorable assumptions about the benefits/harms ratio, population-based screening is not cost effective in HICs and may be more problematic LMICs.
12. Screening program quality and performance management: screening program should have clear goals, objectives, and monitoring for quality control and performance targets	+++	+
	Electronic medical records and associated algorithms to guide management could enhance the quality of care for detected cases, while simplifying quality control. However, such integrated systems are not yet widely available.	Most LMICs do not have the capacity to ensure or measure the quality of care in existing CKD cases; existing capacity is likely to be adversely affected by an influx of new cases.

+, extent to which each principle is met in high-income and low-middle-income countries.

Such trials of screening approaches have proven necessary because new problems or questions arise when tests are used widely for screening, such as determining a threshold value for a positive test to gain the maximum benefit while minimizing the disturbance to those whose outcome will be unchanged by being screened. Thus, any plan to initiate a screening program should be accompanied by an embedded evaluation. Ideally, such an evaluation would be a randomized trial, or at least a parallel group cohort study comparing outcomes between centers that screen and those that do not. Once screening tests are widely used, they must be applied and monitored closely, because loss of quality can produce excess harm that will outweigh the net benefit. Australia has a national Standing Committee on Screening to monitor cancer screening, whereas the United Kingdom National Screening Committee covers screening for cancers, illness in newborns, diabetic retinopathy, and aortic aneurysms.^22,23 These committees audit every stage of screening and the relevant actions that follow screening, from determining the population that receives the initial tests, follow-up processes, diagnosis, and management. When necessary, they may hold inquiries to remedy problems or undertake research to improve the program.

What Tests Can Be Used for Early Detection of CKD?

CKD is defined chiefly by the presence of eGFR <60 ml/min per 1.73 m² or abnormal albuminuria, and thus both blood- and urine-based tests could be used for early detection.²⁴ The diagnosis of CKD requires at least two abnormal values >3 months apart, and therefore a confirmatory test is always needed, at least in theory.²⁵ Serum creatinine is usually used to estimate eGFR because of its wide availability and long history of clinical use. Compared with equations based solely on serum creatinine, equations using both serum cystatin and creatinine have a lower risk of false positives and slightly higher diagnostic accuracy for CKD.²⁶ If these advantages translated into a lower risk of harms or reduced the need for follow-up testing, serum cystatin might offer advantages for early detection programs in HICs and LMICs alike. However, serum cystatin is more costly and much less widely available than serum creatinine, so its use for early detection would require further study. The ideal frequency for estimating eGFR as part of early detection is unknown, but given the chronic nature of CKD, more frequent measurements seem less likely to yield benefit compared with broader population coverage or other proven methods of noncommunicable disease control.

Either dipstick urinalysis or urine albumin-creatinine ratio (ACR) could also be used for the early detection of CKD. The presence of severe albuminuria (2+, 3+, or 4+) on dipstick has a high positive predictive value for clinically meaningful albuminuria,²⁷ but a normal finding on dipstick urinalysis does not rule out significant albuminuria.²⁸ When dipstick urinalysis suggests the presence of less severe albuminuria (especially trace or 1+), there is substantial potential for misclassification, which is undesirable for early detection programs that seek to minimize harms and wasted resources from overdiagnosis.²⁹ This suggests that if dipstick urinalysis is used for early detection, a more stringent threshold (e.g., ≥2+) might be used to define an abnormal test, although this remains speculative.

Assessment of the ACR is not operator dependent and has better diagnostic properties compared with dipstick urinalysis.²⁹ However, because creatinine excretion is proportional to muscle mass, the ACR may be falsely low in muscular people (e.g., younger males) and lead to underestimation of albuminuria, or falsely high in smaller or malnourished people (e.g., elderly females) and lead to overestimation of albuminuria. These considerations are potentially relevant in LMICs, where sensitivity of standard ACR cutoffs for clinically relevant albuminuria appears lower than in HICs,^30–33 perhaps because of lower average muscle mass or racial variation in baseline creatinine excretion. In addition, ACR assays to assess albuminuria are more expensive than dipstick urinalyses,²⁴ although the magnitude of the excess cost varies between countries.

Similar to eGFR assessments, the ideal frequency of urine testing to detect CKD is unknown. Because albuminuria does not always accompany reduced eGFR, it is also unknown whether assessment of albuminuria is the optimal method for early detection of CKD, or whether it should be replaced or combined with eGFR assessment.^34,35

Theoretic Basis for Early Detection of CKD

In theory, early detection of CKD should be beneficial because it enables clinicians to initiate effective treatment of mild disease, preventing loss of kidney function and delaying or avoiding progression to kidney failure. Strategies that are known to reduce mortality or prevent progressive loss of kidney function in CKD include control of BP, use of an angiotensin-converting enzyme inhibitor or angiotensin receptor blocker, and statin treatment, as well as control of blood glucose and use of sodium-glucose cotransporter 2 inhibitors in people with diabetes.^36–43 None of these treatments are curative, but together they can substantially reduce rates of kidney function loss in people with progressive disease and they may also improve cardiovascular outcomes for those at risk, such as the subset of patients with diabetes or known cardiovascular disease. For a small proportion (<10%) of people with CKD, early detection leads to the identification of an uncommon underlying condition such as GN, which enables specific treatment in addition to these general measures.

In theory, the incremental benefit of early detection occurs during the period beginning when CKD is identified by screening or case finding, and ending when CKD would have been detected by testing ordered during clinical practice. However, this assumes that early detection of CKD will change management. This assumption is uncertain, because many of the beneficial treatments (such as BP control in people with hypertension and control of blood glucose in those with diabetes) should be provided whether CKD is present.

There are other critical assumptions for realizing these theoretic benefits. These include the accessibility of the early detection strategy for the target population; the availability, affordability, and acceptability of the treatments that follow for people with newly identified CKD; the effectiveness of treatment in preventing progression outweighing any harms it causes; the likelihood these patients will continue to receive the treatments for years or decades; and the ability of the health system to cope with the influx of new cases and manage appropriate follow-up.

Besides preventing or delaying kidney failure, there could be other benefits associated with early detection of CKD, such as increased eligibility for living donor transplantation, better modality choice if renal replacement is required, more cautious use of intravenous contrast, drug dosing that has been adjusted for eGFR, or perhaps the opportunity to receive medications that would not necessarily be indicated in the absence of CKD (e.g., sodium-glucose cotransporter 2 inhibitors or statins). However, these benefits are extremely speculative, are unlikely to have large health benefits, and are generally not part of the stated rationale for early detection. Although we acknowledge they could exist, we do not discuss them further.

To recapitulate, the large majority of people with CKD but without kidney failure are asymptomatic and the chief goal of early detection of CKD is prevention of kidney failure; however, detection of CKD may not lead to management changes. Therefore, when evaluating the merits of early detection, milder forms of CKD should arguably be conceptualized as high-risk intermediate states (analogous to cervical dysplasia in screening to prevent cervical cancer) rather than the target condition per se.

What Factors Influence the Likelihood That Early Detection of CKD Will Be Beneficial?

Certain factors will diminish any benefits of implementing an early detection strategy (Box 1). These include a relatively high preexisting rate of kidney testing in at-risk groups, a relatively low prevalence of disease in people with and without such risk factors, a relatively low baseline likelihood of developing kidney failure in the reservoir of cases (e.g., because of relatively slow disease progression or short life expectancy in the absence of early detection), limited capacity to treat newly identified cases of CKD, overtesting in the target population once early detection begins (e.g., testing more frequently than is necessary or duplicate testing arranged by multiple providers),⁴³ and relatively high baseline uptake of effective treatments in people without known CKD.

Factors That Will Offset Any Benefits of Early Detection

• Potential harms (e.g., false positives, labeling effects)

• Costs of testing, treatment, and follow-up (including confirmation of abnormal tests and false positives)

• Opportunity costs, meaning potential benefits resulting from other programs that must be foregone because of the investment in early detection

Factors That Reduce the Likelihood of Benefits Associated with Early Detection

• Higher preexisting rate of kidney testing in at-risk groups

• A lower prevalence of disease in people with and without such risk factors

• A higher prevalence of other diseases that require similar treatment among people with CKD, and which could be detected instead

• A lower baseline likelihood of developing kidney failure in the reservoir of cases (e.g., slower progression or shorter life expectancy in the absence of early detection)

• Lower capacity to treat newly identified cases of CKD

• Overtesting in the target population upon initiation of early detection (e.g., testing more frequently than necessary, duplicate testing from multiple providers)

• Higher baseline uptake of effective treatments in people without known CKD.

One key challenge is that, unlike tests that are ordered predominantly for screening (e.g., a test for prostate-specific antigen level) or those ordered in response to clinical suspicion (e.g., blood cultures), serum creatinine is ordered frequently as part of clinical practice—nearly 300 million times annually in the United States alone,⁴⁴ often without a strong indication. This high preexisting use will tend to reduce the benefits of early detection strategies. A second challenge that diminishes or eliminates any potential for benefit—especially in LMICs but also among disadvantaged groups in HICs—is that early detection strategies usually have greater uptake among those with reasonable access to care who may already be receiving effective treatment (e.g., an angiotensin receptor blocker for treatment of hypertension), even if they are not recognized as having CKD.

On the other hand, in some circumstances, the benefits of early detection are more pronounced. These situations include having a higher baseline rate of underlying cases that may benefit from specific treatment (e.g., GN in Japan)⁴⁵ or scenarios in which the consequences of developing kidney failure are more severe (e.g., in most LMICs).⁴⁶

Any benefits of early detection are offset by its potential harms, including the costs and discomforts of investigation (including follow-up of abnormal screening tests, whether false positive or true positive), as well as harms attributable to treatment and follow-up. In addition, benefits of early detection can also be offset by opportunity costs—i.e., potential benefits resulting from other activities that must be foregone because of the investment of resources in early detection efforts (Table 1).

Harms of early detection and its consequences are often minimized by providers. Because most people with CKD do not require kidney biopsy, nephrologists might believe that early detection of kidney disease poses little risk of harm. But this is not necessarily the case, because harms are not limited to complications of invasive procedures. For example, any imaging done as a consequence of early detection may detect abnormalities of questionable clinical significance (e.g., small calculi, solitary cysts, renal artery stenosis, functional hydronephrosis, and incidentalomas⁴⁷), which in turn have potential to trigger further investigations, treatment, or both. To prevent such problems, Welch⁴⁸ argues that thyroid findings should not be mentioned in reports of computed tomography performed to detect lung lesions.

Other harms occur simply from the fact of diagnosis, which may trigger nephrology consultation, follow-up visits, and laboratory testing, leading to cost and inconvenience for patients and families. A few patients with newly detected CKD will have side effects from medications, and a small minority of these adverse events will be potentially serious. New diagnoses may cause considerable anxiety to patients and families,⁴⁹ especially when effective treatment is not available or it causes economic hardship.⁵⁰ False positive diagnoses are relatively common when a single eGFR value is used to define CKD⁵¹; on a population level, any detrimental effects of such false positive results may outweigh the economic benefits in HICs.⁵² Positive screening results may increase anxiety even when not confirmed on follow-up testing, as demonstrated by an Australian study of fecal occult blood testing.⁵³ Finally, data from HICs suggest that labeling with a chronic disease may have lifelong implications for insurance, work performance,⁵⁴ and potentially for occupation.⁵⁵

Although it is tempting to dismiss these harms as rare or clinically inconsequential, the ethical principal of nonmaleficence suggests they should be carefully considered, because substantially more people may experience harms compared with those who experience benefit. The fact that such harms may occur does not mean that early detection strategies should be abandoned. However, making a well informed decision about whether to proceed will require careful assessment of net benefit, which by definition requires consideration of all harms.

For LMICs, it is important to recognize that the opportunity costs of investing resources in early detection could include foregone benefits from interventions outside the health system that might lead to substantial health benefits, such as clean water, safer roads, or better education.⁵⁶ Unfortunately, stakeholders focused on advocating for a particular disease typically do not give much consideration to these interventions. A special opportunity cost that is particularly relevant to LMICs is that funds spent on early detection of CKD are not available to manage known CKD in people who cannot afford or access effective treatment. Therefore, it does not seem appropriate to spend time and money detecting additional cases until most known cases of CKD have been appropriately managed.

Critical Assessment of Population-Based Screening for CKD

At a superficial level, screening for CKD has theoretic appeal because many of the Wilson–Jungner criteria seem to be met. As for many other diseases, closer scrutiny indicates that population-based screening faces challenges. Key issues include a high rate of baseline testing for CKD (meaning a screening program would provide only small added value), barriers to uptake of screening among previously untested people, and relatively little change in recommended management for people with key risk factors such as diabetes and hypertension, whether CKD is detected or not.

Unlike many of the conditions for which population-based screening was initially contemplated, CKD is not usually curable but rather requires lifelong treatment aimed at slowing progression and reducing the risk of complications. This, in turn, highlights the importance of the program- and system-level issues among the principles summarized by Dobrow et al.²⁰ when assessing the merits of population-based screening for CKD (Table 1). Lack of available infrastructure for conducting screening and ensuring follow-up treatment over years to decades, limited opportunities to integrate screening programs with the underlying health system, and substantial barriers to quality control and performance management among detected cases would all be critical issues for the large majority of LMICs. Searching for cases without resolving these issues would increase the likelihood of harms and costs without compensatory benefits. In HICs, there is potential to address all of the systemic and programmatic principles, although some work would be needed to identify the optimal approach.

However, for both HICs and LMICs, there is a lack of evidence that screening will improve clinical outcomes. Moreover, existing economic analyses demonstrate that population-based screening is unlikely to be cost effective.^10,11

Critical Assessment of Case Finding for CKD

For both HICs and LMICs, populations with a higher risk of CKD include people with diabetes, hypertension, and cardiovascular disease. These conditions are highly prevalent, readily diagnosed, and are clearly associated with an increased risk of CKD, and so existing health care systems have prioritized their management in recent years. Researchers have studied the epidemiology of CKD in these conditions extensively and, although there is potential to improve clinical outcomes, this potential has not been confirmed in high-quality clinical trials. Finally, because the population being tested for CKD already has clinically apparent disease, harms caused by labeling may be of less concern than for apparently healthy people. All of these factors increase enthusiasm for case finding, especially in HICs.

Although some economic analyses appear to show that case finding is cost effective, especially among patients with diabetes or hypertension,⁵⁷ an unresolved issue is that management of the latter conditions is not much changed by the presence of CKD when dialysis is unavailable or unaffordable. Therefore, although case finding may be a reasonable use of resources in wealthy countries, further evidence of clinical and economic benefit in LMICs would be essential before commencing such a program.

Even if the efficacy of case finding in LMICs is supported by future studies, deploying a case-finding strategy in such countries would be highly dependent on the availability of program- and system-level characteristics (principles for population screening are summarized in Table 1), because it is not rational to detect cases that cannot be appropriately managed. LMICs that are considering a case-finding strategy should critically assess their state of readiness to address these principles, because it may be more beneficial to improve the management of known cases in their particular context. LMICs will also need to carefully consider their other health-related priorities, because opportunity costs for early detection of CKD may well be greater than those associated with other new programs, especially those that focus on broader issues, such as implementing or strengthening primary health care.

Attention to factors that will make early detection of CKD less beneficial than expected (outlined in Box 1) will help increase the potential benefit of case finding for both HICs and LMICs. Integration of CKD management with management of its major risk factors (e.g., diabetes, hypertension, and vascular disease) would also be expected to maximize benefit, although this is speculative. For HICs, availability of diagnostic tests, the affordability of indicated medications, and access to appropriate expertise are much less significant issues than for LMICs, although harms and the need to minimize overtesting are arguably more significant. Although opportunity costs are not as significant for HICs as in LMICs, minimizing waste is still important. Achieving benefit from case finding in HICs could be facilitated by strong linkages between efforts to detect and manage CKD,⁵⁸ and by careful attention to the quality of care delivered to detected cases, potentially through leveraging electronic medical records.

In summary, for HICs, the key uncertainties relate to how best to implement case-finding strategies to maximize benefits while reducing costs and harms. In contrast, although case finding may make sense for some LMICs, this will not be true in all settings. Thus, careful consideration of principles for population screening and of factors that will make early detection of CKD less beneficial than expected (summarized in Box 1 and Table 1) is required.

Research Priorities for Early Detection of CKD

Much remains to be learned about how best to implement early detection programs in either LMICs or HICs, especially for case finding. The optimal criteria used to trigger case finding are unknown. Whether or not other conditions such as a family history of CKD (or a higher versus lower multivariate CKD risk score)⁵⁹ increase the diagnostic yield sufficiently to warrant case finding would require further investigation in light of the factors outlined in Table 1. In certain countries, additional target conditions (such as schistosomiasis in endemic areas⁶⁰) might merit consideration in case finding for CKD. This would also require further study, recognizing the potential for overlap between case finding and clinical management of the target condition.

ACR and estimating equations based on serum creatinine are the standard tests for diagnosis of CKD in wealthy countries. It is possible that cystatin-based equations would reduce any harms associated with early detection because of their higher positive predictive value compared with creatinine. It is also possible that the higher cost of cystatin would be partially offset by lesser costs attributable to false positive results. However, the appeal of this strategy would depend on the eGFR value used to define a positive test, because the added value of cystatin is reduced when true GFR is <45 ml/min per 1.73 m².²⁶ On the other hand, although using the ACR for CKD detection is more accurate and sensitive than dipstick urinalysis, the former is more costly. There may be less benefit associated with earlier detection of CKD among people with positive ACR but no albuminuria on dipstick, and future studies could examine whether urinalysis is actually preferable to ACR for screening or case finding, given the current lack of data.⁶¹

Regardless of the testing strategy or target population, it would be important to rigorously study the clinical and economic benefits of different approaches to case finding in a variety of settings. This should include studies in both HICs and LMICs (again, considering the factors outlined in Box 1) and ensuring that the comparator is correctly specified. Given the high preexisting frequency of testing in HICs, one should consider evaluating the merits of efforts to “increase awareness” of the importance of CKD,⁶² which might lead to better management of known cases without the economic and opportunity costs associated with early detection.⁶³

Perhaps the most interesting research priorities concern how best to integrate early CKD detection with clinical management or early detection of other chronic diseases, or both. If urine testing, eGFR, or both are used for case finding, it seems reasonable to integrate such testing with existing management programs for noncommunicable diseases that frequently coexist with CKD, such as those aimed at hypertension, diabetes, and cardiovascular disease. However, the ideal frequency of testing, strategies to mitigate overtesting, and the optimal model for management of detected cases are all unknown. Similarly, if case finding for CKD is considered (whether in HICs or LMICs), it seems important to consider simultaneously testing for diabetes and hypertension, and perhaps taking the opportunity to assess and manage overall vascular risk. A Danish trial of screening for cardiovascular risk found a substantial population that warranted treatment according to guidelines, but vigorous treatment with drugs and lifestyle advice did not change outcomes.⁶⁴ Although some programs are already doing this in HICs (e.g., the United Kingdom’s National Health Service Health Check),⁶⁵ the approach’s benefits, harms, and costs have not been thoroughly assessed, and further studies of additional models for such simultaneous testing are needed for both HICs and LMICs.

Whether or not screening or case finding for CKD could be efficiently combined with population-based screening for common malignancies such as colorectal cancer or cervical cancer is an interesting idea that warrants further study.⁶⁶ Given that many LMICs do not have well developed cancer screening programs, this model may be most applicable to HICs. Even in wealthy countries, uptake of cancer screening is lower among people of lower socioeconomic status,⁶⁷ who are at increased risk for CKD,⁶⁸ so the diagnostic yield of this strategy and its potentially adverse effects on equity would require careful consideration.

Summary

Early detection of CKD has considerable theoretic appeal. However, little is known about how best to ensure clinical benefit while minimizing harms, opportunity costs, and wasting of resources from overtesting and cases that are identified but inappropriately managed, especially in LMICs. Available data suggest population-based screening is unlikely to be effective or cost effective in either HICs or LMICs. In contrast, case-finding strategies might avoid many of the limitations associated with population-based screening, and appear to be reasonable investments for HICs and probably some LMICs. However, important uncertainties remain about the circumstances under which case finding might be beneficial and about how best to tailor case-finding strategies to local epidemiology and existing practice patterns. For many LMICs, resources would be better devoted to improving management of known cases rather than early detection of new cases, whereas HICs that choose to emphasize case finding should focus on minimizing overtesting and improving the quality of management for detected cases. Future studies in HICs and LMICs alike should evaluate new models for integrating early detection of CKD with clinical management and early detection of other chronic diseases, as well as assessing the optimal approach for identifying new cases.

Disclosures

M. Tonelli is the past chair and J. Dickinson a past member of the Canadian Task Force on Preventive Health Care. The views expressed here are their own. M. Tonelli is the Director of the World Health Organization Collaborating Centre for the Prevention and Control of CKD. M. Tonelli also reports Daichi Sankyo awarded a grant to M. Tonelli’s institution in lieu of a personal honorarium for a lecture at a scientific meeting in 2017. The lecture topic was not related to the topic of this article. M. Tonelli also received a lecture fee from B. Braun in 2019, the fee was donated to charity. The lecture topic was not related to the topic of this article.

Funding

This research was supported by Canadian Institutes of Health Research Foundation grant FRN 143211 (to M. Tonelli). M. Tonelli was supported by the University of Calgary David Freeze Chair in Health Services Research.