Abstract
BACKGROUND:
Albuminuria, indicated by an elevated urine albumin-to-creatinine ratio (UACR) at baseline, is consistently associated with poor clinical outcomes and increased economic burden. The effect of a change in albuminuria over time on health care resource utilization is not well understood.
OBJECTIVE:
To assess the association between changes in UACR and economic outcomes in patients with chronic kidney disease (CKD) associated with type 2 diabetes (T2D).
METHODS:
The Optum electronic health records database (January 2007 to September 2021) was used to identify adult patients with albuminuria, measured by UACR of 30 mg/g or more (initial test) after diagnosis of T2D and CKD. UACR change was categorized as increased (>30% change), stable (30% increase to 30% decrease), or decreased (>30% change) based on the percentage of change between the initial test and the follow-up test (the last test within 0.5 to 2 years after the initial test). All-cause inpatient (IP) admissions, emergency department (ED) visits, outpatient (OP) visits, and total medical costs were evaluated during the year after the follow-up test. The association of UACR change with health care resource utilization (HRU) was evaluated using Poisson regression, adjusting for key baseline characteristics. Medical costs (2022 US dollars) were estimated using a unit costing approach based on HRU frequencies.
RESULTS:
Among 144,814 eligible patients included in the study, 81,084 (56%) had decreased, 31,766 (22%) had stable, and 31,964 (22%) had increased UACR. Patients with increased UACR had higher HRU (IP admissions: 0.24 per-person per-year [PPPY]; ED visits: 0.35 PPPY; OP visits: 21.20 PPPY) and annual medical costs ($15,013 PPPY) than patients with stable UACR (IP: 0.18 PPPY; ED: 0.31 PPPY; OP: 19.13 PPPY; costs: $12,521 PPPY) and decreased UACR (IP: 0.17 PPPY, ED: 0.31 PPPY, OP: 19.90 PPPY; costs: $12,329 PPPY). Compared with patients with increased UACR, those with decreased UACR had adjusted incidence rate ratios of 0.79 (95% CI = 0.76-0.82) for IP, 0.88 (0.85-0.92) for ED, and 0.96 (0.95-0.97) for OP, and patients with stable UACR had adjusted incidence rate ratios of 0.82 (0.78-0.86) for IP, 0.91 (0.87-0.95) for ED, and 0.94 (0.92-0.95) for OP (all P values of <0.001).
CONCLUSIONS:
Among patients with CKD and T2D who had albuminuria, an increase in UACR over time was associated with significantly higher HRU and costs compared with patients with stable or decreased UACR. Managed care organizations and other health care decision-makers should consider strategies that enhance monitoring and management of UACR in patients with CKD and T2D to potentially reduce HRU and associated costs.
Plain language summary
Patients with high levels of urine protein (ie, albuminuria), measured through urine albumin-to-creatinine ratio (UACR) test, have a higher risk of kidney and cardiovascular events and higher medical costs. This study assessed the impact of albuminuria management on health care resource use among patients with chronic kidney disease and type 2 diabetes and found that patients with stable or decreased UACR levels had fewer hospital visits and lower medical costs compared with patients with increased UACR.
Implications for managed care pharmacy
In patients with chronic kidney disease and type 2 diabetes, achieving stable or decreasing UACR is associated with lower health care resource use and medical costs. Considering the underutilization of UACR testing in practice, it is important to highlight the value of UACR monitoring and management as actionable strategies for clinical practice and managed care pharmacy to improve outcomes and reduce economic burden in patients with chronic kidney disease and type 2 diabetes.
Chronic kidney disease (CKD) imposes a significant health care burden, impacting approximately 36 million adults in the United States.1 Diabetes is one of the leading causes of CKD, and approximately 25% of adults with diabetes develop CKD.2,3 Furthermore, patients with diabetes have faster progression of CKD, and the disease progression is associated with higher health resource utilization (HRU) and increased annual medical costs.4,5
Albuminuria is an early marker of kidney damage and is diagnosed by an elevated level of urine albumin-to-creatinine ratio (UACR) over 30 mg/g. According to recent clinical guidelines, albuminuria control is one of the key treatment targets for effective kidney-heart risk management.6,7 The American Diabetes Association recommends targeting a decrease of at least 30% for CKD patients with a UACR of 300 mg/g or more.7 This guidance aligns with the introduction of novel treatments like sodium-glucose cotransporter 2 inhibitors,8 glucagon-like peptide-1 receptor agonists,9 and finerenone,10 which show promise in controlling albuminuria, offering new avenues for mitigating both kidney and cardiovascular risks.
Prior studies have demonstrated strong associations between albuminuria, kidney events, and cardiovascular events.11–13 A meta-analysis of randomized trials found support for the use of change in albuminuria as a surrogate endpoint for progression of CKD.14 Similarly, post hoc analyses of large clinical trials also linked reductions in UACR levels with lower risks for major cardiovascular events.15–17 Furthermore, results from a post hoc mediation analysis using clinical trial data in patients with CKD associated with type 2 diabetes (T2D) showed that reduction in UACR accounted for 84% and 37% of finerenone’s treatment effect on long-term kidney and cardiovascular outcomes, respectively.18
Albuminuria, partially because of its association with elevated risks of various clinical outcomes, is linked to increased HRU and incurs higher annual medical costs from $3,500 to $12,800 per patient among patients with T2D.19,20
With new treatments being available for controlling albuminuria, it would be valuable to quantify the impact of UACR reduction on HRU and medical costs in patients with CKD and T2D. These findings can be used to inform payer decisions on health plan benefits and formularies, as well as managed care recommendations and guidelines for this population. Specifically, quantifying the relationship between UACR improvement and reduction in economic burden can guide managed care professionals, including pharmacists, to advocate for integrating routine UACR testing and setting UACR as a treatment target within standard CKD management protocols, thereby supporting timely, targeted therapeutic interventions, informing formulary decisions, and optimizing health care and resource use. Therefore, the study aimed to assess the patterns of UACR change in current clinical practice and their association with economic outcomes in patients with CKD, albuminuria, and prior T2D.
Methods
DATA SOURCE
This retrospective cohort study used the Optum electronic health record (EHR) database from January 2007 to September 2021. The database represents more than 150,000 medical providers from over 2,000 hospitals and 7,000 clinics. The database contains information on demographic characteristics, type of health care provider, medical history, and diagnoses for all types of encounters within the network, detailed area of care during hospitalization, in-hospital procedures, inpatient medications, physician prescriptions, and laboratory data.
STUDY POPULATION AND DESIGN
Adult patients with CKD and T2D who had albuminuria (ie, UACR of ≥30 mg/g) were included in the study. T2D was identified using a modified EMERGE algorithm that is applicable to EHRs and incorporates International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) and ICD-10-CM diagnosis codes, T2D medications, and laboratory measurements.21,22 CKD on or after the date of T2D diagnosis was identified based on the Kidney Disease: Improving Global Outcomes clinical guidelines using diagnosis codes, serum creatinine, and UACR.23 Patients were required to have at least 1 elevated UACR test (≥30 mg/g, initial test) on or after CKD diagnosis following T2D diagnosis, as well as at least 1 UACR test within 0.5 to 2 years after the initial UACR test. The follow-up UACR is defined as the last UACR measurement within this time range, and the date of the follow-up UACR test was defined as the index date. In addition, patients were required to be at least 18 years old as of the initial elevated UACR test and to have continuous eligibility from 1 year before to 2 years after the initial UACR test and 1 year after the follow-up UACR test. The continuous eligibility was defined as an uninterrupted period of time during which records indicating any clinical activity (ie, disease diagnoses, procedures, laboratory visits, medication prescriptions and administrations, and other clinical visits and encounters) had gaps of less than 6 months. Patients were excluded from the study if they had diagnoses or procedures indicating preexisting albuminuria/proteinuria or end-stage kidney disease (ESKD), including chronic dialysis and kidney transplant) or preexisting conditions associated with kidney function decline (including systemic lupus erythematosus, polycystic kidney disease, and kidney cancer), during the 1-year period before the initial elevated UACR test. Baseline characteristics were assessed in the 6-month pre-index period and economic outcomes were assessed during the 12-month post-index period (ie, following the follow-up UACR test) (Supplementary Figure 1 (251.1KB, pdf) , available in online article).
UACR CHANGE PATTERNS
UACR measurements were either directly reported values from laboratory test records of UACR or calculated as a ratio of the reported urine albumin concentration and the urine creatinine concentration. UACR change was assessed as the percentage of change in UACR from the initial UACR test to the follow-up UACR test and was categorized into 3 categories using clinical meaningful cutoffs: more than 30% increase, stable (30% decrease to 30% increase), or more than 30% decrease, in accordance with the existing literature and clinical guidelines.6,7,24
HRU AND COSTS
The outcomes of interest, including all-cause and CKD-related HRU and all-cause and CKD-related medical costs, were evaluated during the 12-month post-index period. All-cause HRU was defined as annual counts of inpatient (IP) stays, emergency department (ED) visits, and outpatient (OP) visits, as well as the number of days hospitalized for each patient per year. CKD-related HRU was identified as HRU of which the medical records were associated with a diagnosis for CKD or related diseases and complications. Medical costs, including IP, ED, and OP costs, were estimated using a unit costing approach by multiplying the respective annual frequencies by the average cost for each HRU type. The unit costs were derived from a previous study among patients with CKD and T2D using the Optum Clinformatics claims database.5 Specifically, among adult patients with CKD and T2D, unit costs were derived by averaging the standardized costs associated with each type of encounter across patients, assessed during the 1-year period after having both CKD and T2D diagnoses. The unit costs for all-cause HRU were $4,430 per diem for IP stays, $2,042 per ED visit, and $361 per OP visit. For CKD-related HRU, the unit costs were $4,757 per diem for IP stays, $2,011 per ED visit, and $359 per OP visit. All costs were converted to 2022 US dollars. This study focused on medical costs and HRU. Medication use and pharmacy costs were not assessed, as the EHR data do not contain detailed prescription information, such as medication fill records or associated charges.
STATISTICAL ANALYSIS
Patient demographics and clinical characteristics were summarized for the overall patient population and by UACR change category during the 6-month pre-index period. For all-cause and CKD-related HRU, the count and proportion of patients having each encounter visit type during the 12-month post-index period were summarized. The frequencies of annual visits for each visit type and annualized length of stay for hospitalization were summarized using means and SDs. Adjusted HRU frequencies and incidence rate ratios were estimated using Poisson regression models, with UACR change category as the primary exposure variable, controlling for age, sex, race, region, insurance type, estimated glomerular filtration rate (eGFR) level, body mass index, time from initial UACR test to follow-up test, hemoglobin A1c level, hypertension, ischemic heart disease, heart failure, stroke, diabetes-related microvascular complications diseases, hyperlipidemia, anemia, and acidosis. To account for violations of the equidispersion assumption, SEs were calculated using the Huber-White heteroskedastic robust estimator.
The annual costs associated with ED and OP were calculated for each patient by multiplying the frequency of annual visits for ED and OP and associated unit costs, respectively. The annual costs associated with IP stays were calculated as the annual number of days in the hospital multiplied by the IP per diem unit cost to account for the variability of hospital durations. Total annual costs for each patient were calculated as the sum of the IP, ED, and OP costs. All costs were described using means and SDs, both overall and stratified by UACR change category. Adjusted mean medical costs were calculated by multiplying the values for adjusted HRU by the unit cost for that type of HRU. The total adjusted medical cost was calculated by summing the adjusted medical costs for each type of HRU. SEs and 95% CIs of the adjusted mean HRU and costs were calculated via bootstrapping by resampling the original patients with replacement.
All statistical analyses were conducted using SAS (SAS Studio Release: 3.8 [Enterprise Edition], SAS Institute Inc.) and R version 4.2.2 (R Foundation for Statistical Computing). P < 0.05 was considered statistically significant for all analyses, and tests were 2-tailed.
Results
CHARACTERISTICS OF THE STUDY POPULATION
Among 1,826,223 patients identified with CKD and T2D, 1,124,649 (61.6%) patients had UACR tests, and 773,534 (42.4%) patients had an elevated UACR test on or after CKD diagnosis; among them, 408,152 (22.3%) patients had a follow-up UACR test within 0.5 to 2 years after initial elevated UACR. The study further excluded patients aged 18 years and older, patients with ESKD or any of the relevant preexisting conditions, and patients without continuous eligibility from 1 year before to 2 years after the initial UACR test or without 12-month continuous eligibility after the follow-up UACR test, leaving 144,814 patients in the final sample (Figure 1).
FIGURE 1.
Study Population Flowchart
CKD = chronic kidney disease; eGFR = estimated glomerular filtration rate; ESKD = end-stage kidney disease; T2D = type 2 diabetes; UACR = urine albumin-to-creatinine ratio.
The demographic and clinical characteristics of the study population by UACR change category are shown in Table 1. The median time from the initial UACR test to the last follow-up test (within 0.5 to 2 years after the initial UACR test) was 16.2 months. The overall population had a mean age of 65.9 (SD = 11.9) years at the time of follow-up UACR test, and 48.2% were male. Most patients were White (82.3%), resided in the Midwest (67.1%), and had commercial insurance (57.6%). The most common comorbid conditions were hypertension (74.5%) and hyperlipidemia (71.4%).
TABLE 1.
Patient Baseline Demographic and Clinical Characteristics by UACR Change Categorya
| Total | UACR change category | P value | |||
|---|---|---|---|---|---|
| >30% decrease | Stable | >30% increase | |||
| N=144,814 | n=81,084 | n=31,766 | n=31,964 | ||
| Demographics | |||||
| Age, years (mean±SD) | 65.9 ± 11.9 | 65.0 ± 11.9 | 66.7 ± 11.7 | 67.3 ± 11.6 | <0.001 |
| Sex, n (%) | |||||
| Female | 74,969 (51.8) | 45,030 (55.5) | 15,198 (47.8) | 14,741 (46.1) | <0.001 |
| Male | 69,804 (48.2) | 36,024 (44.4) | 16,563 (52.1) | 17,217 (53.9) | |
| Unknown | 41 (0.0) | 30 (0.0) | 5 (0.0) | 6 (0.0) | |
| Race and ethnicity, n (%) | |||||
| African American | 16,408 (11.3) | 8,918 (11.0) | 3,595 (11.3) | 3,895 (12.2) | <0.001 |
| Asian | 2,907 (2.0) | 1,614 (2.0) | 643 (2.0) | 650 (2.0) | |
| White | 119,134 (82.3) | 66,968 (82.6) | 26,126 (82.2) | 26,040 (81.5) | |
| Other/Unknown | 6,365 (4.4) | 3,584 (4.4) | 1,402 (4.4) | 1,379 (4.3) | |
| Region, n (%) | |||||
| Midwest | 97,183 (67.1) | 54,558 (67.3) | 20,904 (65.8) | 21,721 (68.0) | <0.001 |
| South | 20,243 (14.0) | 10,947 (13.5) | 5,092 (16.0) | 4,204 (13.2) | |
| Northeast | 16,072 (11.1) | 9,271 (11.4) | 3,354 (10.6) | 3,447 (10.8) | |
| West | 8,143 (5.6) | 4,533 (5.6) | 1,738 (5.5) | 1,872 (5.9) | |
| Other/Unknown | 3,173 (2.2) | 1,775 (2.2) | 678 (2.1) | 720 (2.3) | |
| Insurance type, n (%) | |||||
| Commercial | 75,104 (57.6) | 43,391 (59.1) | 16,175 (56.7) | 15,538 (54.4) | <0.001 |
| Medicaid | 7,146 (5.5) | 4,152 (5.7) | 1,445 (5.1) | 1,549 (5.4) | |
| Medicare | 44,234 (33.9) | 23,611 (32.2) | 10,048 (35.2) | 10,575 (37.0) | |
| Other payer type | 833 (0.6) | 496 (0.7) | 177 (0.6) | 160 (0.6) | |
| Uninsured | 1,843 (1.4) | 1,029 (1.4) | 404 (1.4) | 410 (1.4) | |
| Unknown | 1,314 (1.0) | 724 (1.0) | 270 (0.9) | 320 (1.1) | |
| Index year, n (%) | |||||
| 2008-2013 | 51,717 (35.7) | 28,121 (34.7) | 11,658 (36.7) | 11,938 (37.3) | <0.001 |
| 2014-2019 | 93,097 (64.3) | 52,963 (65.3) | 20,108 (63.3) | 20,026 (62.7) | |
| Clinical characteristics | |||||
| eGFR category, n (%) | |||||
| G1 - Normal or high: eGFR ≥90 mL/min/1.73 m2 | 45,537 (31.4) | 26,851 (33.1) | 9,894 (31.1) | 8,792 (27.5) | <0.001 |
| G2 - Mildly decreased: eGFR between 60 and <90 mL/min/1.73 m2 | 54,508 (37.6) | 31,055 (38.3) | 11,837 (37.3) | 11,616 (36.3) | |
| G3a - Mildly to moderately decreased: eGFR between 45 and <60 mL/min/1.73 m2 | 20,924 (14.4) | 10,918 (13.5) | 4,586 (14.4) | 5,420 (17.0) | |
| G3b - Moderately to severely decreased: eGFR between 30 and <45 mL/min/1.73 m2 | 11,290 (7.8) | 5,534 (6.8) | 2,597 (8.2) | 3,159 (9.9) | |
| G4 - Severely decreased: eGFR between 15 and <30 mL/min/1.73 m2 | 2,946 (2.0) | 1,316 (1.6) | 666 (2.1) | 964 (3.0) | |
| G5 - Kidney failure: eGFR <15 mL/min/1.73 m2 | 330 (0.2) | 119 (0.1) | 82 (0.3) | 129 (0.4) | |
| Unknown | 9,279 (6.4) | 5,291 (6.5) | 2,104 (6.6) | 1,884 (5.9) | |
| eGFR level (mL/min/1.73 m2) | |||||
| Mean±SD | 76.4 ± 23.4 | 77.9 ± 22.8 | 76.0 ± 23.5 | 73.0 ± 24.2 | <0.001 |
| Missing, n (%) | 9,279 (6.4) | 5,291 (6.5) | 2,104 (6.6) | 1,884 (5.9) | |
| Initial UACR (mg/g) | |||||
| Mean±SD | 144.4 ± 434.6 | 146.3 ± 468.9 | 151.2 ± 449.2 | 132.6 ± 312.1 | <0.001 |
| Median (Q1, Q3) | 52.6 (37.2, 98.0) | 55.0 (38.3, 101.0) | 47.6 (35.5, 85.6) | 52.0 (37.0, 98.3) | |
| Follow-up UACR (mg/g) | |||||
| Mean±SD | 158.1 ± 665.9 | 53.1 ± 403.9 | 164.1 ± 624.2 | 418.6 ± 1,054.5 | <0.001 |
| Median (Q1, Q3) | 35.5 (16.4, 93.0) | 18.3 (10.6, 31.9) | 48 (33.7, 90.7) | 145.2 (80.3, 339.1) | |
| Body mass index | |||||
| Mean±SD | 34.0 ± 7.8 | 34.2 ± 7.8 | 33.8 ± 7.7 | 34.0 ± 7.9 | <0.001 |
| Missing, n (%) | 14,484 (10.0) | 7,909 (9.8) | 3,243 (10.2) | 3,332 (10.4) | |
| A1c, % | |||||
| Mean±SD | 7.3 ± 1.5 | 7.2 ± 1.4 | 7.3 ± 1.5 | 7.5 ± 1.6 | < 0.001 |
| Missing, n (%) | 7,486 (5.2) | 4,096 (5.1) | 1,648 (5.2) | 1,742 (5.4) | |
| Comorbidities, n (%) | |||||
| Hypertension | 107,937 (74.5) | 59 604 (73.5) | 23,752 (74.8) | 24,581 (76.9) | <0.001 |
| Ischemic heart diseases | 24,880 (17.2) | 13,093 (16.1) | 5,592 (17.6) | 6,195 (19.4) | <0.001 |
| Heart failure | 9,852 (6.8) | 5,234 (6.5) | 2,033 (6.4) | 2,585 (8.1) | <0.001 |
| Stroke | 2,547 (1.8) | 1,331 (1.6) | 539 (1.7) | 677 (2.1) | <0.001 |
| Diabetic ketoacidosis | 98 (0.1) | 42 (0.1) | 22 (0.1) | 34 (0.1) | <0.01 |
| Diabetes-related microvascular complications | 22,827 (15.8) | 12 607 (15.5) | 4,664 (14.7) | 5,556 (17.4) | <0.001 |
| Hyperlipidemia | 103,336 (71.4) | 58,212 (71.8) | 22,622 (71.2) | 22,502 (70.4) | <0.001 |
| Hyperkalemia | 2,175 (1.5) | 1,073 (1.3) | 480 (1.5) | 622 (1.9) | <0.001 |
| Hypoglycemia | 679 (0.5) | 379 (0.5) | 137 (0.4) | 163 (0.5) | 0.346 |
| Hyponatremia | 1,950 (1.3) | 982 (1.2) | 419 (1.3) | 549 (1.7) | <0.001 |
| Anemia (nonhereditary) | 16,301 (11.3) | 8,786 (10.8) | 3,484 (11.0) | 4,031 (12.6) | <0.001 |
| Acidosis | 701 (0.5) | 362 (0.4) | 142 (0.4) | 197 (0.6) | <0.001 |
| Acute kidney injury | 3,074 (2.1) | 1,524 (1.9) | 618 (1.9) | 932 (2.9) | <0.001 |
| Volume depletion | 1,862 (1.3) | 1,023 (1.3) | 369 (1.2) | 470 (1.5) | <0.01 |
| Edema | 8,263 (5.7) | 4,444 (5.5) | 1,708 (5.4) | 2,111 (6.6) | <0.001 |
| Urinary tract infections | 9,045 (6.2) | 5,004 (6.2) | 1,855 (5.8) | 2,186 (6.8) | <0.001 |
Body mass index was calculated as weight in kilograms divided by height in meters squared.
Age was assessed on the date of the follow-up UACR. Insurance type, eGFR category, body mass index, and all the comorbidities were assessed within 6 months prior to or on the date of the follow-up UACR.
A1c = hemoglobin A1c; eGFR = estimated glomerular filtration rate; UACR = urine albumin-to-creatinine ratio.
At index date (ie, follow-up UACR test), 22.2% of patients had CKD stage G3 (eGFR between 30 and <60 mL/min/1.73 m2) and 2.2% of patients were in stage G4/5 (eGFR <30 ml/min/1.73 m2). The median value of the initial UACR was 52.6 (first quartile [Q1], third quartile [Q3]: 37.2, 98.0) mg/g; the median of the follow-up UACR was 35.5 (Q1, Q3: 16.4, 93.0) mg/g.
UACR CHANGE PATTERNS
There were 81,084 patients (56.0%) with elevated UACR who experienced a decrease greater than 30% in UACR level in 2 years after the initial UACR, whereas 31,964 (22.1%) patients had an increase greater than 30% in UACR level in 2 years. Compared with patients with stable or a decrease greater than 30% in UACR, those with an increase greater than 30% increase in UACR were older, more likely to be male and to have baseline comorbidities, and had higher baseline eGFR.
The mean eGFR at index was 73.0 (SD = 24.2) ml/min/1.73 m2 for patients with an increase greater than 30% in UACR, 76.0 (SD = 23.5) ml/min/1.73 m2 for those with a stable UACR, and 77.9 (SD = 22.8) ml/min/1.73 m2 for patients with a decrease greater than 30% in UACR. The median of initial UACR measurement was 55.0 mg/g for those with a decrease greater than 30% in UACR, 47.6 mg/g for those with stable UACR, and 52.0 mg/g for those with an increase greater than 30% in UACR. The median follow-up UACR measurement was 18.3 mg/g for those with a decrease greater than 30% in UACR, 48.0 mg/g for those with stable UACR, and 145.2 mg/g for those with an increase greater than 30% in UACR.
ASSOCIATION BETWEEN UACR CHANGE AND HRU
There were 18,270 patients (12.6%) who had at least 1 IP admission during the 1-year follow-up period, with 0.18 mean IP stays per-person per-year (PPPY), and on average 1.15 days per year in the hospital. Patients with increasing UACR levels had a higher number of IP admissions (mean 0.24 stays PPPY) and longer length of stay in-hospital (1.50 days PPPY) than patients with decreased and stable UACR, with 0.17 and 0.18 IP admissions PPPY and 1.02 and 1.12 days of IP stay PPPY, respectively. Similar to IP admissions, patients with increasing UACR had a higher number of ED (0.35 visits PPPY) and OP (21.20 visits PPPY) visits compared with patients with decreased UACR (ED: 0.31 visits PPPY, OP: 19.90 visits PPPY) and stable UACR (ED: 0.31 visits PPPY, OP: 19.13 visits PPPY) (Table 2).
TABLE 2.
All-Cause HRU by UACR Change Category During the 1-Year Study Period After UACR Change
| Total N = 144,814 | Index UACR categories | P value | |||
|---|---|---|---|---|---|
| >30% Decrease n = 81,084 | Stable n = 31,766 | >30% Increase n = 31,964 | |||
| IP admission | |||||
| Patients with ≥1 IP admission | 18,270 (12.6%) | 9,322 (11.5%) | 3,963 (12.5%) | 4,985 (15.6%) | <0.001 |
| Number of IP admissions (PPPY), mean ± SD | 0.18 ± 0.60 | 0.17 ± 0.57 | 0.18 ± 0.58 | 0.24 ± 0.68 | <0.001 |
| Length of stay (PPPY), mean ± SD | 1.15 ± 6.13 | 1.02 ± 5.59 | 1.12 ± 6.36 | 1.50 ± 7.11 | <0.001 |
| ED visit | |||||
| Patients with ≥1 ED visit | 28,184 (19.5%) | 15,401 (19.0%) | 6,054 (19.1%) | 6,729 (21.1%) | <0.001 |
| Number of ED visits (PPPY), mean ± SD | 0.32 ± 0.91 | 0.31 ± 0.91 | 0.31 ± 0.85 | 0.35 ± 0.95 | <0.001 |
| OP visit | |||||
| Patients with ≥1 OP visit | 144,093 (99.5%) | 80,715 (99.5%) | 31,584 (99.4%) | 31,794 (99.5%) | <0.05 |
| Number of OP visits (PPPY), mean ± SD | 20.02 ± 17.08 | 19.90 ± 16.93 | 19.13 ± 16.54 | 21.20 ± 17.89 | <0.001 |
ED = emergency department; HRU = health care resource utilization; IP = inpatient; OP = outpatient; PPPY = per-person per-year; UACR = urine albumin-creatinine ratio.
The results were consistent after adjusting for demographic and key clinical characteristics. Adjusted HRU was substantially lower for patients with decreasing UACR and those with stable UACR levels relative to patients with increasing UACR, with adjusted incidence rate ratios (using increasing UACR as the reference category) of 0.79 (95% CI = 0.76-0.82) for average number of IP admissions, 0.78 (95% CI = 0.73-0.83) for length of IP stay per year, 0.88 (95% CI = 0.85-0.92) for average number of ED visits, and 0.96 (95% CI = 0.95-0.97) for average number of OP visits, all statistically significant with P < -.001. There were no differences in HRU between those with stable and decreasing UACR (Table 3). Similar trends were observed for the association between UACR and CKD-related HRU (Supplementary Tables 1 and 2 (251.1KB, pdf) ).
TABLE 3.
| Total IP stays IRR (95% CI) | IP length of stay IRR (95% CI) | Total ED visits IRR (95% CI) | Total OP visits IRR (95% CI) | |
|---|---|---|---|---|
| UACR change (ref: >30% increase) | ||||
| UACR change: >30% decrease | 0.79 (0.76-0.82)c | 0.78 (0.73-0.83)c | 0.88 (0.85-0.92)c | 0.96 (0.95-0.97)c |
| Stable | 0.82 (0.78-0.86)c | 0.82 (0.75-0.89)c | 0.91 (0.87-0.95)c | 0.94 (0.92-0.95)c |
Poisson models with a log-link were used to estimate the IRR. Coefficients presented have been exponentiated.
All the models adjust for age at follow-up UACR test, sex (reference = female), race and ethnicity (reference = White), region (reference = Midwest), insurance type (reference = Medicare/Medicaid), eGFR level (reference = eGFR level: G1), body mass index (reference = 18.5-25), time from the initial UACR test to the follow-up UACR test, hemoglobin A1c (reference = 6.5%-7%), hypertension, ischemic heart disease, heart failure, stroke, diabetes-related microvascular complications diseases, hyperlipidemia, anemia, and acidosis.
P < 0.001.
ED = emergency department; eGFR = estimated glomerular filtration rate; IP = inpatient; IRR = incidence rate ratio; OP = outpatient; ref = reference; UACR = urine albumin-creatinine ratio.
ASSOCIATION BETWEEN UACR CHANGE AND COSTS
The mean annual all-cause medical costs were $15,013 for patients with increasing UACR, which were significantly higher than those for patients with stable UACR (mean: $12,521; difference vs increasing UACR: $2,492) and for patients with decreasing UACR (mean: $12,329; difference: $2,684). There was no significant difference in annual costs between those with stable and decreasing UACR (Figure 2). The annual cost differences between those with either decreasing or stable UACR and those with increasing UACR were primarily driven by lower costs associated with IP and ED, which accounted for $1,748 (70.1%) of the difference among those with stable UACR and $2,216 (82.6%) of the difference among those with decreasing UACR. There were similar trends for CKD-related costs, with $3,795 in annual CKD-related costs for those with increasing UACR, $2,674 for those with stable UACR, and $2,313 for those with decreasing UACR.
FIGURE 2.
Observed Costs by UACR Change Category
ED = emergency department; HRU = health care resource utilization; UACR = urine albumin-creatinine ratio.
aP < 0.001.
Discussion
In this large, real-world retrospective cohort study of patients with CKD associated with T2D and albuminuria, approximately 22% of patients had an increase in UACR levels within 2 years. Our study found that when compared with those with an increased UACR, patients who maintained or decreased their urine ACR over time had lower costs across IP, ED, and OP settings. These findings were independent of baseline clinical and demographic characteristics, including baseline eGFR. In addition, the cost savings associated with stable or decreasing UACR resulted from reductions in costs associated with IP and ED visits, rather than outpatient drug use. These results demonstrate the potential economic benefits of UACR monitoring and effective management of UACR in patients with CKD associated with T2D.
A previous study found that achieving a 30% UACR reduction was associated with clinical benefits, including reducing risk of cardiovascular events and mortality and slowing down CKD progression.24 It is notable that the American Diabetes Association clinical guidelines recommend targeting at least 30% reduction in UACR to slow disease progression for patients with CKD and UACR levels of 300mg/g or more.7 Our study found that compared with patients having a stable or decreased UACR, those with an increased UACR were strongly associated with higher frequency of IP visits, which is the driver of cost difference between these two groups. These clinical benefits can translate to savings in medical expenditures with costs attributable to cardiovascular events on the order of $73,300 for myocardial infarction, $36,000 for angina, and $71,600 for stroke.25 Cardiovascular-related costs are particularly high for patients in high-risk groups, including those with diabetes and CKD.26,27
Of note, our study focused on the short-term economic impact associated with UACR change patterns. The long-term benefit could be amplified considering the long-term benefits in reducing clinical outcomes. Previous studies provided projections of the 10-year risk of developing ESKD, finding that those with decreasing UACR had a lower probability of developing ESKD than those with no worsening in UACR, and this difference in probabilities grew over time.24 That the clinical benefits of reducing UACR are not only persistent but seem to grow larger over time supports the notion that the long-term economic benefits could be even greater, proportionally, than suggested in the present short-term study.
There are important clinical, research and policy implications of our results. From a clinical perspective, these findings emphasize the importance of remeasuring UACR in practice, especially among adults who have an initial positive test result. Current clinical guidelines recommend at least annual monitoring of albuminuria among patients with CKD and more frequent testing among T2D patients or other high-risk groups.23 However, undertesting of UACR has been commonly observed in clinical practice with a study showing an average annual testing frequency of 0.6 tests among patients with CKD and T2D.28,29 In our sample selection, approximately 40% patients with CKD and T2D did not have any UACR testing, and among those with an elevated UACR test, about half did not have a follow-up UACR measurement within 1.5 years. Of note, the testing rate may be overestimated, as our sample selection criteria were enriched for patients with UACR measurement. Our study demonstrates that effective management of albuminuria, which would require sufficient UACR monitoring, may help reduce health care resource use and medical costs. This is in line with a previous study that found that lower UACR testing rates were associated with higher health care costs at the state level.29 Therefore, sufficient UACR monitoring and early interventions are crucial in effective disease management of CKD. From a research and policy perspective, the incorporation of UACR testing and monitoring as a Healthcare Effectiveness Data and Information Set quality measure has the potential to improve testing rates in health systems in the future, and studies should be performed to evaluate the impact of increased UACR testing on appropriate treatment of CKD.
A strength of this study is the use of a large-scale EHR database, which permitted the selection of a large sample of patients with CKD and prior T2D across all US geographic regions and age groups. However, the final sample size was limited by the availability of laboratory data. Another strength was that eGFR/UACR laboratory measures were used to identify patients with CKD and characterize disease stages in addition to diagnosis codes. This approach is advantageous as the use of CKD diagnosis codes alone may result in underdetection of CKD, particularly patients with an early-stage impairment, and errors in classifying disease stages. Additionally, change in UACR level was carefully defined to minimize potential selection bias and reflect clinical meaningful categorization to inform clinical practice. Moreover, a large proportion of our study population were early-stage CKD patients, which represents an important group of T2D patients who have albuminuria but with normal or mildly impaired kidney function in clinical practice. Finally, specific types of HRU were identified, allowing for a more granular examination of patterns of HRU reduction.
LIMITATIONS
The study also had some limitations that need to be considered. UACR measurements were less frequently tested and recorded in the data, which led to the use of a single UACR measurement to define albuminuria and a relatively long interval to identify the follow-up UACR (ie, 0.5 to 2 years after the initial UACR). Patients with longer intervals between the two selected UACR tests may differ from those with shorter intervals in terms of disease monitoring and access to health services. Second, as measures of UACR are highly variable, there may have been misclassification of UACR categories and change status. Indeed, we observed a large proportion of patients experiencing UACR reduction, which may partially be due to the regression-to-the mean phenomenon. However, misclassification caused by the variability of laboratory results is not expected to differentially affect subgroups of the study population and is not expected to be associated with the economic outcomes of interest. Third, the study sample was restricted to patients with at least 2 UACR measurements, which may not represent patients with less frequent UACR monitoring. Patients with more frequent UACR testing in the database may be different from those without in terms of disease severity and health resource accessibility. Therefore, the generalizability of our study results to the broader patient population with CKD and T2D in the United States warrants further assessment. Fourth, continuous clinical activity with gaps of less than 6 months was used as a proxy for continuous enrollment in the EHR data. However, some patient records, for example, those from outside the health care network, may not have been captured. Fifth, the requirement for a relatively long period of continuous clinical activity (ie, from 1 year before to 2 years after the initial UACR test and 1 year after the follow-up test) may have selected patients with less severe disease and thus with longer continuous clinical activity, who could have different cost profiles compared with those excluded. Sixth, because EHR data do not contain cost information, costs in the current study were estimated using a unit costing approach, which may be less precise than using actual cost data. Additionally, although medication use and pharmacy costs are important components in the economic burden evaluation, they were not assessed in this study because of the lack of detailed prescription information in the EHR data, such as medication fill records and associated charges. As a result, our analysis focused on medical costs and medical service–related HRU. Future studies assessing the impact of UACR change on economic outcomes in a more comprehensive manner, including medication use and pharmacy costs, are warranted. The more accurate cost information from a claims database was traded off for more precise identification of CKD and defining UACR levels offered by the laboratory results present in the EHR database. Seventh, the data period partially overlaps with the COVID-19 pandemic. Changes in clinical practice during the pandemic may have influenced the economic outcomes assessed near the end of the study period. However, given that the overall data period is much longer than the portion affected by the pandemic, any impact is likely to be small. Finally, as with most observational studies, although the analysis adjusted for observed baseline differences, unmeasured and residual confounding may still exist.
Conclusions
In patients with CKD and T2D who had albuminuria, more than 1 in 5 experienced an increase greater than 30% in UACR, which was associated with significantly higher HRU and costs compared with stable or decreased UACR. These findings highlight that early and appropriate UACR monitoring and management, with the goal of reducing UACR by more than 30% is important for improving outcomes and reducing costs for patients with CKD and T2D.
Disclosures
This study was funded by Bayer US, LLC. The sponsor was involved in the study design, analysis and interpretation of data, writing the manuscript, and the decision to submit the report for publication.
Dr Pantalone provided consulting services to AstraZeneca, Corcept, Diasome, Eli Lilly, Merck, Novo Nordisk, Sanofi Aventis, Twinhealth, and Bayer US, LLC, which funded the development and conduct of this study and development of the manuscript; and conducted teaching and speaking for AstraZeneca, Corcept, Merck, and Novo Nordisk and research for Eli Lilly, Merck, Novo Nordisk, Twinhealth, and Bayer US, LLC.
Dr Tangri received grants from the Canadian Institutes of Health Research, National Institutes of Health, Kidney Foundation of Canada, AstraZeneca, Boehringer Ingelheim, Janssen Pharmaceuticals, Research Manitoba, Otsuka Pharmaceutical, Tricida, Eli Lilly, and Bayer US, LLC, which funded the development and conduct of this study and development of the manuscript; provided consulting services to AstraZeneca, Boehringer Ingelheim, GSK, Janssen Pharmaceuticals, Otsuka Pharmaceutical, Prokidney, Roche, Tricida, and Bayer US, LLC; received honoraria for lectures, presentations, speakers, bureaus, manuscript writing, or educational events from AstraZeneca, Boehringer Ingelheim, Janssen Pharmaceuticals, Eli Lilly, Otsuka Pharmaceutical, and Tricida; has a pending patent for a microfluidic device for point-of-care detection of urine albumin; served on the advisory board of AstraZeneca, Janssen Pharmaceuticals, BI-Lilly, Otsuka Pharmaceutical, and National Kidney Foundation; reported ownership of ClinPredict, Klinrisk, Quanta, and Marizyme; and holds stock options of Mesentech, Renibus Therapeutics, PulseData, and Tricida.
Drs Singh, Farag, Beeman, Du, Kong, Williamson, and Li were employees of Bayer US, LLC during the development of this study, and may own stock options. Dr Farag is now an employee of Alexion AstraZeneca Rare Disease Unit.
Drs Chen, Betts, and Wu are employees of Analysis Group, Inc., a company that has provided paid consulting services to Bayer US, LLC, which funded the development and conduct of this study and development of the manuscript. Dr Rabideau was an employee of Analysis Group, Inc. during the development of this study.
References
- 1.Kidney Disease Statistics for the United States (2023). National Institute of Diabetes and Digestive and Kidney Diseases. Accessed September 12, 2024. https://www.niddk.nih.gov/health-information/health-statistics/kidney-disease.
- 2.Kovesdy CP. Epidemiology of chronic kidney disease: An update 2022. Kidney Int Suppl (2011). 2022;12(1):7-11. doi: 10.1016/j.kisu.2021.11.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Zelnick LR, Weiss NS, Kestenbaum BR, et al. Diabetes and CKD in the United States population, 2009-2014. Clin J Am Soc Nephrol. 2017;12(12):1984-90. doi: 10.2215/CJN.03700417 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Zhang X, Fang Y, Zou Z, et al. Risk factors for progression of CKD with and without diabetes. J Diabetes Res. 2022;9613062. doi: 10.1155/2022/9613062 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Mullins CD, Pantalone KM, Betts KA, et al. CKD Progression and economic burden in individuals with CKD associated with type 2 diabetes. Kidney Med. 2022;4(11):100532. doi: 10.1016/j.xkme.2022.100532 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Group KDW; Kidney Disease: Improving Global Outcomes (KDIGO) Diabetes Work Group. KDIGO 2022 Clinical Practice Guideline for Diabetes Management in Chronic Kidney Disease. Kidney Int. 2022;102(5S):S1-127. doi: 10.1016/j.kint.2022.06.008 [DOI] [PubMed] [Google Scholar]
- 7.ElSayed NA, Aleppo G, Aroda VR, et al. ; on behalf of the American Diabetes Association. 11. Chronic kidney disease and risk management: Standards of care in diabetes-2023. Diabetes Care. 2023;46(suppl 1):S191-202. doi: 10.2337/dc23-S011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Mavrakanas TA, Tsoukas MA, Brophy JM, Sharma A, Gariani K. SGLT-2 inhibitors improve cardiovascular and renal outcomes in patients with CKD: A systematic review and meta-analysis. Scientific Reports. 2023;13(1):15922. doi: 10.1038/s41598-023-42989-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Perkovic V, Tuttle KR, Rossing P, et al. ; FLOW Trial Committees and Investigators. Effects of semaglutide on chronic kidney disease in patients with type 2 diabetes. N Engl J Med. 2024;391(2):109-21. doi: 10.1056/NEJMoa2403347 [DOI] [PubMed] [Google Scholar]
- 10.Ruilope LM, Pitt B, Anker SD, et al. Kidney outcomes with finerenone: An analysis from the FIGARO-DKD study. Nephrol Dial Transplant. 2023;38(2):372-83. doi: 10.1093/ndt/gfac157 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Matsushita K, Coresh J, Sang Y, et al. ; CKD Prognosis Consortium. Estimated glomerular filtration rate and albuminuria for prediction of cardiovascular outcomes: A collaborative meta-analysis of individual participant data. Lancet Diabetes Endocrinol. 2015;3(7):514-25. doi: 10.1016/S2213-8587(15)00040-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Pasternak M, Liu P, Quinn R, et al. Association of albuminuria and regression of chronic kidney disease in adults with newly diagnosed moderate to severe chronic kidney disease. JAMA Netw Open. 2022;5(8):e2225821. doi: 10.1001/jamanetworkopen.2022.25821 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ärnlöv J, Nowak C. Association between albuminuria, incident cardiovascular events, and mortality in persons without hypertension, diabetes, and cardiovascular disease. Eur J Prev Cardiol. 2022;29(1):e4-6. doi: 10.1093/eurjpc/zwaa095 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Heerspink HJL, Greene T, Tighiouart H, et al. ; Chronic Kidney Disease Epidemiology Collaboration. Change in albuminuria as a surrogate endpoint for progression of kidney disease: a meta-analysis of treatment effects in randomised clinical trials. Lancet Diabetes Endocrinol. 2019;7(2):128-39. doi: 10.1016/S2213-8587(18)30314-0 [DOI] [PubMed] [Google Scholar]
- 15.Jun M, Ohkuma T, Zoungas S, et al. ; ADVANCE Collaborative Group. Changes in albuminuria and the risk of major clinical outcomes in diabetes: Results from ADVANCE-ON. Diabetes Care. 2018;41(1):163-70. doi: 10.2337/dc17-1467 [DOI] [PubMed] [Google Scholar]
- 16.Persson F, Bain SC, Mosenzon O, et al. ; LEADER Trial Investigators. Changes in albuminuria predict cardiovascular and renal outcomes in type 2 diabetes: A post hoc analysis of the LEADER Trial. Diabetes Care. 2021;44(4):1020-6. doi: 10.2337/dc20-1622 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Waijer SW, Xie D, Inzucchi SE, et al. Short-term changes in albuminuria and risk of cardiovascular and renal outcomes in type 2 diabetes mellitus: A post hoc analysis of the EMPA-REG OUTCOME Trial. J Am Heart Assoc. 2020;9(18):e016976. doi: 10.1161/JAHA.120.016976 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Agarwal R, Tu W, Farjat AE, et al. ; FIDELIO-DKD and FIGARO-DKD Investigators. Impact of finerenone-induced albuminuria reduction on chronic kidney disease outcomes in type 2 diabetes: A mediation analysis. Ann Intern Med. 2023;176(12):1606-16. doi: 10.7326/M23-1023 [DOI] [PubMed] [Google Scholar]
- 19.Zhou Z, Chaudhari P, Yang H, et al. Healthcare resource use, costs, and disease progression associated with diabetic nephropathy in adults with type 2 diabetes: A retrospective observational study. Diabetes Ther. 2017;8(3):555-71. doi: 10.1007/s13300-017-0256-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Nichols GA, Vupputuri S, Lau H. Medical care costs associated with progression of diabetic nephropathy. Diabetes Care. 2011;34(11):2374-8. doi: 10.2337/dc11-0475 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Kho AN, Hayes MG, Rasmussen-Torvik L, et al. Use of diverse electronic medical record systems to identify genetic risk for type 2 diabetes within a genome-wide association study. J Am Med Inform Assoc. 2012;19(2):212-8. doi: 10.1136/amiajnl-2011-000439 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Pantalone KM, Hobbs TM, Wells BJ, et al. Changes in characteristics and treatment patterns of patients with newly diagnosed type 2 diabetes in a large United States integrated health system between 2008 and 2013. Clin Med Insights Endocrinol Diabetes. 2016;9:CMED.s39761. doi: 10.4137/CMED.s39761 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Kidney Disease: Improving Global Outcomes (KDIGO) Diabetes Work Group . KDIGO 2020 Clinical Practice Guideline for Diabetes Management in Chronic Kidney Disease. Kidney Int. 2020;98(4s):S1-115. 2020; [DOI] [PubMed] [Google Scholar]
- 24.Coresh J, Heerspink HJL, Sang Y, et al. ; Chronic Kidney Disease Prognosis Consortium and Chronic Kidney Disease Epidemiology Collaboration. Change in albuminuria and subsequent risk of end-stage kidney disease: An individual participant-level consortium meta-analysis of observational studies. Lancet Diabetes Endocrinol. 2019;7(2):115-27. doi: 10.1016/S2213-8587(18)30313-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.O’Sullivan AK, Rubin J, Nyambose J, Kuznik A, Cohen DJ, Thompson D. Cost estimation of cardiovascular disease events in the US. PharmacoEconomics. 2011;29(8):693-704. doi: 10.2165/11584620-000000000-00000 [DOI] [PubMed] [Google Scholar]
- 26.Duh MS, Fulcher NM, White LA, et al. Costs associated with cardiovascular events in patients with hypertension in US managed care settings. J Am Soc Hypertens. 2009;3(6):403-15. doi: 10.1016/j.jash.2009.09.001 [DOI] [PubMed] [Google Scholar]
- 27.Kent S, Schlackow I, Lozano-Kühne J, et al. What is the impact of chronic kidney disease stage and cardiovascular disease on the annual cost of hospital care in moderate-to-severe kidney disease? BMC Nephrology. 2015;16(1):65. doi: 10.1186/s12882-015-0054-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Chu CD, Xia F, Du Y, et al. Estimated prevalence and testing for albuminuria in US adults at risk for chronic kidney disease. JAMA Netw Open. 2023;6(7):e2326230. doi: 10.1001/jamanetworkopen.2023.26230 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Betts KA, Song J, Elliott J, et al. Geographical variation in kidney function testing and associations with health care costs among patients with chronic kidney disease and type 2 diabetes. Am J Manag Care. 2022;28(6 suppl):S112-9. doi: 10.37765/ajmc.2022.89211 [DOI] [PubMed] [Google Scholar]