Microalbuminuria Screening for Detecting Chronic Kidney Disease in the General Population: A Systematic Review

Abstract

Background: Microalbuminuria screening is widely used in high-risk populations but seldom used in the general population for detecting chronic kidney disease (CKD). Systematic reviews focused on screening for CKD are rare, and the issues about microalbuminuria screening in the general population have never been reviewed. We systematically reviewed studies regarding microalbuminuria screening and evaluated the benefits and harms of this screening method in the general population. Methods: We systematically searched MEDLINE, PubMed, and the Cochrane Library for English articles published from January 1970 to 13 December 2011. Quality assessments were performed using the QUADAS tool or the Drummond’s 10-point checklist. Due to the high heterogeneity of the study designs, meta-analysis for the study results was not possible. Therefore, we performed a narrative synthesis. Results: Six articles from four studies made up our final study population, with four articles evaluating different screening methodologies and two reporting cost-effectiveness analyses. The qualities of the included articles ranged from fair to high. Spot urine albumin concentration and spot urine albumin:creatinine ratio had a similar diagnostic performance for microalbuminuria screening in the general population. Screening for microalbuminuria in high-risk populations, such as patients with diabetes, hypertension, or old age, was cost-effective. However, there was no consensus regarding the cost-effectiveness for microalbuminuria screening in the general population. Conclusions: Microalbuminuria screening in high-risk populations is cost-effective. However, the cost-effectiveness of screening for microalbuminuria in the general population deserves further study. To keep costs low, spot urine albumin concentration may be preferable than the albumin:creatinine ratio.

• albuminuria
• mass screening
• proteinuria
• kidney failure
• chronic
• renal insufficiency
• chronic

INTRODUCTION

Though dialysis therapies have been introduced to patients with end-stage renal disease for decades, improvement in dialysis techniques and the quality of medical care for dialysis patients have recently led to increasing prevalence rates of dialysis patients, which is placing a greater burden on the health insurance system of many countries.^1,2 In 2008, Wen et al. reported that the national prevalence of chronic kidney disease (CKD) in Taiwan was 11.93%, with only one-third of CKD patients being aware of their disease.³ Moreover, most CKD patients have already entered the middle or late stages when they are initially diagnosed.

With earlier detection and intervention for CKD patients, it is more likely that disease progression will be prevented from entering into the later stages.⁴ Because of the long asymptomatic early phase, as well as the high costs of taking care of the dialysis patients, early screening and regular follow-up for CKD patients are being more closely attended to in recent years.

Current guidelines recommend screening for CKD in high-risk populations. Kidney Disease: Improving Global Outcomes (KDIGO) recommends screening patients with hypertension, diabetes, or cardiovascular disease.⁵ The National Kidney Foundation suggests screening individuals at increased risk of developing CKD.^6,7 The American Diabetes Association advocates annual screening of all diabetic patients.⁸ However, the benefits and harms, as well as the cost-effectiveness, of CKD screening in the general population remain unclear. Furthermore, no consensus has been achieved regarding the best method of CKD screening.⁹ Recently, Crews and colleagues reported that the topics regarding CKD screening were rated as high priority for systematic review by stakeholders in the CKD community, which reflects that there is significant uncertainty regarding CKD screening and implies that high-quality systematic reviews could improve clinical practice and policy.¹⁰

Urine dipstick for proteinuria is widely applied as the screening method to detect CKD, and while it is cheaper, it is also less sensitive. Methods to detect microalbuminuria, on the other hand, are widely used in screening the high-risk population but seldom used in the general population for detecting CKD. Moreover, systematic reviews focused on screening for CKD are rare, and the issues about microalbuminuria screening in the general population have never been reviewed. Herein, we systematically reviewed clinical studies regarding microalbuminuria screening for CKD and evaluated the benefits and harms of this screening method in the general population.

MATERIALS AND METHODS

Data Sources

We systematically searched the databases of MEDLINE, PubMed, and the Cochrane Library for articles published from January 1970 to 13 December 2011, using a combination of free keywords and MeSH terms related to microalbuminuria, chronic kidney disease, and mass screening (detailed searching strategies are provided in the supplement http://informahealthcare.com/doi/suppl/[doinumber]). We searched additional relevant studies in the reference lists of all identified publications.

Study Selection

We followed three predefined inclusion criteria: (1) full-text English publication of randomized controlled trials (RCTs), cohort studies, case–control studies, or cross-sectional studies; (2) studies conducted among individuals from the general population; and (3) studies that collected information on microalbuminuria screening in the general population.

Data Extraction and Quality Assessment

The following information was extracted and entered into databases by two investigators independently (H-Y Wu and J-W Huang): interventions, study design, inclusion and exclusion criteria, methodological quality criteria, study population, patient baseline characteristics, use of concurrent screening methods, benefits, and harms. If relevant information regarding design or outcomes was unclear, or doubt existed for duplicate publications, the authors were contacted to obtain the necessary information. The methodological quality of eligible trials was evaluated independently by two investigators (H-Y Wu and J-W Huang). The Quality Assessment of Diagnostic Accuracy Studies (QUADAS) tool was used to critically appraise the studies reporting on the methodological quality of test accuracy.¹¹ Drummond’s 10-point checklist was used to critically appraise the studies reporting the elements of an economic evaluation.¹² Disagreements between the two authors were resolved by consensus.

Data Synthesis and Analysis

We first summarized the trials qualitatively. Due to the high heterogeneity of study designs, as well as lack of adequate detailed data, meta-analysis for the study results was not possible. Therefore, we performed a narrative synthesis.

RESULTS

Search Results

Figure 1 details the search and selection process. Briefly, of the 60 unique citations retrieved, 16 were determined to be relevant to our study questions and were identified for full-text article review. Another two articles were identified from the reference list and reviewed. Finally, six published articles from four studies made up our final study population, including one RCT following screening,¹³ one non-randomized controlled trial following screening,¹⁴ one cohort study,¹⁵ two cross-sectional studies,^16,17 and one cost-effectiveness model simulation.¹⁸

Figure 1.  Summary of trial identification and selection.

Qualitative Summary

The characteristics of the six included articles are summarized in Tables 1 and 2. Three articles reported results from the PREVEND study (the Prevention of Renal and Vascular Endstage Disease study), which was conducted in the Netherlands, with 95% of the study population being Caucasian.^13,15,16 Another two studies were conducted among Australian Aborigines and Pakistani Indo-Asians,^14,17 and the final study used a US population for the model simulation.¹⁸ The mean age of the participants ranged from 43 to 52 years, the percentage of male participants ranged from 45% to 65%, and the follow-up period ranged from 0 to 84 months. Four articles evaluated different screening methodologies,^14–17 and two articles reported cost-effectiveness estimates.^13,18 None of the included studies reported potential harms associated with screening.

Table 1. Characteristics of articles included in the systematic review.

Study	Year	Design	Evaluation type	Country and race	Population	Participant number	Mean age (years)	Sex (female)	Follow-up (months)	Note
Hoy et al.^14	2003	Controlled study	Screening program	Australian Aborigines, Australia	Screen general population. Treat HTN, DM, albuminuric subjects	267	43.4	53%	40.7
Gansevoort et al.^16	2005	Cross-sectional	Screening program	Caucasian (95.4%), The Netherlands	General population (28–75 years)	2527	48.8	52.9%	—	PREVENDstudy
Atthobari et al.^13	2006	RCT, double-blind, placebo-controlled	CEA, use of Euros/LYG	Caucasian (95%), The Netherlands	Screen general population (28–75 years). Treat microalbuminuric subjects	864	51	35%	46	PREVEND study
Jafar et al.^17	2007	Cross-sectional	Screening program	Indo-Asian, Pakistan	General population (>40 years)	577	51.8	54.4%	—
Hoerger et al.^18	2010	Cost-effectiveness model simulation	CEA, use ofUS$/QALY	USA	General population (50–90 years)	—	—	—	—
van der Velde et al.^15	2010	Cohort study	Screening program	Caucasian (95%), The Netherlands	General population (28–75 years)	3398	49	55%	84	PREVEND study

Note: RCT, randomized controlled trial; CEA, cost-effectiveness analysis; DM, diabetes mellitus; HTN, hypertension; LYG, life-year gained; QALY, quality-adjusted life-years; PREVEND study, Prevention of Renal and Vascular Endstage Disease study.

Table 2.  Intervention and outcomes of articles included in the systematic review.

Study	Year	Intervention	Main outcomes	Sensitivity and specificity	Comments	Note
Hoy et al.^14	2003	ACR. Perindopril (n = 267) vs. historical matched controls (n = 327)	Treatment group showed 50%reduction in deaths and 57% reduction in renal failure.		Screening ACR in the high riskpopulation may be cost-effective.
Gansevoort et al.^16	2005	UAC, ACR.	UAC was comparable to ACR, diagnostic performance was satisfactory in both exams.	UAC: AUC 0.92, DV 11.2 mg/L,Sen 85%, Spe 85%.	To keep the cost lower,screening UAC in the generalpopulation is suggested.	PREVEND study
				ACR: AUC 0.93, DV 9.9 mg/g,Sen 88%, Spe 88%.
Atthobari et al.^13	2006	UAC. Fosinopril 20 mg + pravastatin 40 mg (n = 431)vs. placebo (n = 433), 2 × 2 factorial	Lower CV events in fosinoprilgroup.The cost-effectiveness of screening was 16,700 Euros/LYG.		Screening UAC in the general population and treat high-risk subjects with fosinopril maybe cost-effective.	PREVEND study
Jafar et al.^17	2007	UAC, ACR.	Both UAC and ACR were acceptable for population screening, but Sen was suboptimal (especially female).	Female	Lowering the existingthresholds of UAC or ACR is suggested.
				UAC: AUC 0.86, DV 0.5 mg/dL,Sen 87%, Spe 75%.
				ACR: AUC 0.86, DV 9.3 mg/g,Sen 74%, Spe 81%.
				Male
				UAC: AUC 0.88, DV 1.7 mg/dL,Sen 74%, Spe 94%.
				ACR: AUC 0.90, DV 6.7 mg/g,Sen 90%, Spe 77%.
Hoerger et al.^18	2010	ACR.	Universal annual screening had a cost-effectiveness ratio of 73,000 US$/QALY relative to no screening.	Model parameterACR: Sen 73%, Spe 96%.	Screening ACR is cost-effective for DM or HTN population,but not for the generalpopulation.
van der Velde et al.^15	2010	UAC (>10 or >20 mg/dL) vs. high CV risk vs. high CV risk + age >55 years	Only UAC prescreening approaches detected CKD patients with accelerated renal function loss and increased CV risk.	UAC >10 mg/dL: Sen 40%, Spe 96%.	UAC is the most efficientscreening method comparedto risk factor screening inthe general population.	PREVEND study
				UAC >20 mg/dL: Sen 58%, Spe 81%.
				High CV risk: Sen 28%, Spe 90%.
				High CV risk + age >55 years: Sen 65%, Spe 71%.

Note: ACR, urine albumin creatinine ratio; AUC, area under the curve; DV, discriminator value; DM, diabetes mellitus; CKD, chronic kidney disease; CV, cardiovascular; HTN, hypertension; Sen, sensitivity; Spe, specificity; UAC, urine albumin concentration; PREVEND study, Prevention of Renal and Vascular Endstage Disease study.

Table 3.  Quality assessment of articles reporting methodological quality of test accuracy.

The QUADAS tool		Hoy et al.(2003)^14	Gansevoort et al. (2005)^16	Jafar et al. (2007)^17	van der Velde et al. (2010)^15
1.	Was the spectrum of patients representative of the patients who will receive the test in practice?	Yes	Yes	Yes	Yes
2.	Were selection criteria clearly described?	Yes	Yes	Yes	Yes
3.	Is the reference standard likely to correctly classify the target condition?	Yes	Yes	Yes	Yes
4.	Is the time period between reference standard and index test short enough to be reasonably sure that the target condition did not change between the two tests?	Yes	Unclear	Yes	Unclear
5.	Did the whole sample or a random selection of the sample receive verification using a reference standard of diagnosis?	Yes	Yes	Yes	Yes
6.	Did patients receive the same reference standard regardless of the index test result?	Yes	Yes	Yes	Yes
7.	Was the reference standard independent of the index test (i.e., the index test did not form part of the reference standard)?	Yes	Yes	Yes	Yes
8.	Was the execution of the index test described in sufficient detail to permit replication of the test?	Yes	Yes	Yes	Yes
9.	Was the execution of the reference standard described in sufficient detail to permit its replication?	Yes	Yes	Yes	Yes
10.	Were the index test results interpreted without knowledge of the results of the reference standard?	Unclear	Yes	Unclear	Yes
11.	Were the reference standard results interpreted without knowledge of the results of the index test?	Unclear	Unclear	Unclear	Unclear
12.	Were the same clinical data available when test results were interpreted as would be available when the test is used in practice?	Unclear	Unclear	Unclear	Unclear
13.	Were uninterpretable/intermediate test results reported?	No	No	No	No
14.	Were withdrawals from the study explained?	No	No	No	No

Note: QUADAS, Quality Assessment of Diagnostic Accuracy Studies.

Table 4.  Quality assessment of articles reporting the elements of economic evaluation.

Drummond’s checklist	Atthobari et al. (2006)^13	Hoerger et al. (2010)^18
1. Was a well-defined question posed in answerable form?	Yes	Yes
2. Was a comprehensive description of the competing alternatives given?	Yes	Yes
3. Was the effectiveness of the program or services established?	Yes—double-blind RCT	Unsure—model simulation
4. Were all the important and relevant costs and consequences for each alternative identified?	Yes	Yes
5. Were costs and consequences measured accurately in appropriate physical units?	Yes	Yes
6. Were the cost and consequences valued credibly?	Yes	Yes
7. Were costs and consequences adjusted for differential timing?	Yes	Yes
8. Was an incremental analysis of costs and consequences of alternatives performed?	Yes—use of Euros/LYG	Yes—use of US$/QALY
9. Was allowance made for uncertainty in the estimates of costs and consequences?	Yes—1-way SA	Yes—1-way SA
10. Did the presentation and discussion of study results include all issues of concern to users?	Yes	Yes

Note: RCT, randomized controlled trial; LYG, life-year gained; SA, sensitivity analyses; QALY, quality-adjusted life-years.

Among the screening methodologies, two articles used urine albumin concentration (UAC),^13,15 two articles used the ratio of spot urine albumin to creatinine (ACR),^14,18 and two articles compared UAC with ACR.^16,17 Both Gansevoort and Jafar compared the performances of UAC and ACR in spot urine versus the reference test of urine albumin excretion in 24-h urine collection.^16,17 The study by van der Velde comprised a cross-sectional part that compared four different approaches in their ability to detect CKD, with a longitudinal part that compared risks for progression of CKD.¹⁵ The study by Hoy reported the follow up from a multifaceted interventional study that targeted high-risk Australian Aborigines in remote areas, including those with diabetes, hypertension, macroalbuminuria, or microalbuminuria.¹⁴

Both Hoy and Atthobari conducted controlled trials among albuminuric subjects after screening the general population.^13,14 Hoy treated the albuminuric subjects with perindopril and compared the treatment group with a historical matched control group.¹⁴ Atthobari conducted a randomized double-blind, placebo-controlled trial among albuminuric subjects with fosinopril and pravastatin in a 2 × 2 factorial design.¹³ Among the two studies reporting a cost-effectiveness analysis (CEA), Atthobari presented the findings in Euros per life-year gained (LYG), and Hoerger presented the findings in US$ per quality-adjusted life-years (QALY) gained.^13,18

Among the four articles reporting screening methodologies and test accuracy, we used the QUADAS tool for quality assessment (Table 3).^14–17 Two of the four articles were based on one prospective Dutch study (the PREVEND study).^15,16 Overall the qualities of the four articles were fair. All four articles adequately selected a representative cohort of consecutive patients suspected for CKD. All four articles had an adequate description of the patient-selection procedure, characteristics of the study participants, the reference standard, and the used cutoff value of the marker. The time between index test and reference test was insufficiently reported in two articles.^15,16 Blinding for the results of the marker when interpreting the reference test was insufficiently reported in all four articles. Uninterruptable test results or withdrawals were not reported in all of the studies.

Among the two articles reporting elements of economic evaluation, we used Drummond’s 10-point checklist for quality assessment (Table 4).^13,18 Overall the qualities of the two articles were high. Both studies provided comprehensive descriptions of the interventions and comparators and mostly provided clear details of the resource types and how these were measured and valued. Both studies analyzed the incremental cost-effectiveness ratio and adjusted costs and effects for discounting. Sensitivity analyses were performed in both studies.

Among the two articles related to the PREVEND study using UAC as the screening method, Atthobari concluded that screening the general population and treating albuminuric subjects with fosinopril might be cost-effective, whereas van der Velde concluded that UAC was the most efficient screening method compared to risk factor screening in the general population.^13,15 Of the two studies using ACR as the screening method, Hoy concluded that screening the general population and treating the albuminuric patients might be cost-effective; nevertheless, Hoerger concluded that screening was only cost-effective in diabetic or hypertensive patients but not in the general population.^14,18 Regarding the two articles comparing ACR with UAC, Jafar concluded that both screening methods were acceptable tests for population screening but the sensitivity was suboptimal, and Gansevoort concluded that diagnostic performance was satisfactory in both exams and that screening UAC in the general population was favored in order to keep the costs down.^16,17

The cost-effective analysis by Atthobari included a cross-sectional part on screening UAC in the general population and a double-blind RCT part to treat the microalbuminuria subjects,¹³ which demonstrated that screening the general population by UAC was cost-effective (€16,700/LYG for the study population, which was below the suggested Dutch threshold of €20,000/LYG).¹³ Hoerger reported the CEA of screening ACR in a simulated CKD cohort.¹⁸ The model cohort was based on the US Renal Data System data, and screened microalbuminuria at 1-, 2-, 5-, or 10-year intervals, followed by treatment with angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers.¹⁸ Relative to nonscreening, targeted annual screening had cost-effectiveness ratios of $21,000/QALY, $55,000/QALY, and $155,000/QALY for persons with diabetes, persons with hypertension, and persons with neither diabetes nor hypertension, respectively.¹⁸ Hoerger, therefore, concluded that microalbuminuria screening was cost-effective for patients with diabetes or hypertension but not for the general population.¹⁸

DISCUSSION

To the best of our knowledge, this study is the first systematic review evaluating microalbuminuria screening in the general population. Our findings show a similar diagnostic performance for microalbuminuria screening using either UAC or ACR in the general population. Yet, in terms of maintaining low costs, UAC may be preferable than ACR as the screening method. While screening for microalbuminuria in the high-risk population, such as patients with diabetes, hypertension, or advanced age, is cost-effective, screening for microalbuminuria in the general population might still remain cost-effective.

Few systematic reviews have evaluated screening for CKD in the general population.^19,20 A recent systematic review conducted by Fink and colleagues has tried to identify RCT of CKD screening but no study has been retrieved; therefore, the benefits and harms of screening were not analyzed.^19,21 Craig has performed a systematic review followed by model simulation and concluded that screening for dipstick proteinuria in population over 50 years of age was cost-effective; however, model simulation for the general population could not be performed due to limited data.²⁰ Nevertheless, microalbuminuria screening, a method more sensitive and specific than screening dipstick proteinuria, has never been systematically reviewed for its benefits, harms, and cost-effectiveness in the general population.

Although screening microalbuminuria in the high-risk population, especially the diabetic patients, has already been widely used and studied,²² studies involving microalbuminuria screening in the general population are rare, probably due to the relatively high cost of microalbuminuria examinations. In this systematic review, only six articles from four studies have been identified to be related to microalbuminuria screening in the general population. The qualities of those included articles ranged from fair to high, with heterogeneous study designs. Among the six articles, three articles reported the results from the PREVEND study, and this might introduce bias into the result of this systematic review.

The cost-effectiveness of screening for microalbuminuria in the general population remains unclear as only two studies reported a CEA, and their conclusions were different. Atthobari conducted a double-blind RCT on microalbuminuria patients following the mass screening in the PREVEND study, which demonstrated that screening the general population was cost-effective.¹³ On the other hand, Hoerger performed a cost-effectiveness model simulation for the US population and concluded that only screening the high-risk population was cost-effective.¹⁸ Cost-effectiveness studies with a larger population and longer follow-up time are necessary to disambiguate this issue.

We identified no evidence of harms associated with microalbuminuria screening because none of the included studies reported that. Potential harms of screening include follow-up of false-positive results, psychological effects from labeling asymptomatic individuals as diseased, medication adverse effects, and increased difficulty in keeping health insurance coverage.¹⁹ Whether microalbuminuria screening in the general population results in substantial harms relative to their benefits deserves further studies to clarify.

Our findings have some limitations. First, due to the limited number and heterogeneous nature of the included studies, meta-analysis could not be performed. However, we have performed a comprehensive systematic review on three electronic databases with a predefined protocol and tried to provide the best evidence from available studies through our narrative synthesis. Second, the best evidence should be a randomized screening trial performed in a large number of participants throughout the country with lifetime follow-up. Nevertheless, none of the included studies was designed to perform randomization at the screening phase. Future studies with randomized screening design are preferred, and model simulation for mass screening studies could be an alternative method if RCTs are difficult and expensive to perform.

In conclusion, our systematic review has shown that UAC and ACR yielded a similar diagnostic performance for microalbuminuria screening in the general population. According to the available data, in order to keep the cost lower, UAC may be preferable. However, while screening for microalbuminuria in the high-risk population, such as patients with diabetes, hypertension, or old age, is certainly cost-effective, further research is required to verify the benefits, harms, and cost-effectiveness of microalbuminuria screening in the general population in order to facilitate a consensus on screening strategies.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Notes

Prof. Kuo-Liong Chien and Prof. Mei-Shu Lai contributed equally to this work.