Purpose of review

Diabetes mellitus is a leading cause of chronic kidney disease (CKD) that confers faster kidney disease progression, higher mortality, and various metabolic derangements including hypoglycemia.

Recent findings

Even in the absence of diabetes mellitus, growing research demonstrates that CKD patients are at heightened risk for hypoglycemia via multiple pathways. In CKD patients transitioning to end-stage renal disease (ESRD), spontaneous resolution of hyperglycemia and frequent hypoglycemia resulting in reduction and/or cessation of glucose-lowering medications are frequently observed in a phenomenon described as 'burnt-out diabetes'. In non-CKD patients, it is well established that hypoglycemia is causally associated with mortality, with pathways including arrhythmias, sudden cardiac death, stroke, and seizures. Increasing evidence shows that, in CKD and ESRD patients with and without diabetes mellitus, hypoglycemia is associated with cardiovascular complications and mortality risk.

Summary

Given the high prevalence of hypoglycemia in CKD patients and the morbidity and mortality associated with this metabolic complication, a multimodal strategy is needed to prevent dysglycemia, including individualization of glycemic targets, selection of glucose-lowering medications less likely to induce hypoglycemia, medical nutrition therapy administered by trained dietitians, and accurate and precise hypoglycemia detection methods, such as self-monitored blood glucose or continuous glucose monitoring including during dialysis treatment.

Abstract The renin-angiotensin system (RAS), bone morphogenetic protein (BMP), and WNT pathways are involved in pathogenesis of diabetic kidney disease (DKD). This study characterized assays for urinary angiotensinogen (AGT), gremlin-1, and matrix metalloproteinase 7 (MMP-7), components of the RAS, BMP, and WNT pathways and examined their excretion in DKD. We measured urine AGT, gremlin-1, and MMP-7 in individuals with type 1 diabetes and prevalent DKD (n = 20) or longstanding (n = 61) or new-onset (n = 10) type 1 diabetes without DKD. These urine proteins were also quantified in type 2 DKD (n = 11) before and after treatment with candesartan. The utilized immunoassays had comparable inter- and intra-assay and intraindividual variation to assays used for urine albumin. Median (IQR) urine AGT concentrations were 226.0 (82.1, 550.3) and 13.0 (7.8, 20.0) μg/g creatinine in type 1 diabetes with and without DKD, respectively (P < 0.001). Median (IQR) urine gremlin-1 concentrations were 48.6 (14.2, 254.1) and 3.6 (1.7, 5.5) μg/g, respectively (P < 0.001). Median (IQR) urine MMP-7 concentrations were 6.0 (3.8, 10.5) and 1.0 (0.4, 2.9) μg/g creatinine, respectively (P < 0.001). Treatment with candesartan was associated with a reduction in median (IQR) urine AGT/creatinine from 23.5 (1.6, 105.1) to 2.0 (1.4, 13.7) μg/g, which did not reach statistical significance. Urine gremlin-1 and MMP-7 excretion did not decrease with candesartan. In conclusion, DKD is characterized by markedly elevated urine AGT, MMP-7, and gremlin-1. AGT decreased in response to RAS inhibition, suggesting that this marker reflects therapeutic response. Urinary components of the RAS, BMP, and WNT pathways may identify risk of DKD and aid development of novel therapeutics.

Importance

Diabetic kidney disease is the leading cause of chronic and end-stage kidney disease in the United States and worldwide. Changes in demographics and treatments may affect the prevalence and clinical manifestations of diabetic kidney disease.

Objective

To characterize the clinical manifestations of kidney disease among US adults with diabetes over time.

Design, setting, and participants

Serial cross-sectional studies of adults aged 20 years or older with diabetes mellitus participating in National Health and Nutrition Examination Surveys from 1988 through 2014.

Exposures

Diabetes was defined as hemoglobin A1c greater than 6.5% or use of glucose-lowering medications.

Main outcomes and measures

Albuminuria (urine albumin-to-creatinine ratio ≥30 mg/g), macroalbuminuria (urine albumin-to-creatinine ratio ≥300 mg/g), reduced estimated glomerular filtration rate (eGFR <60 mL/min/1.73 m2), and severely reduced eGFR (<30 mL/min/1.73 m2), incorporating data on biological variability to estimate the prevalence of persistent abnormalities.

Results

There were 6251 adults with diabetes included (1431 from 1988-1994, 1443 from 1999-2004, 1280 from 2005-2008, and 2097 from 2009-2014). The prevalence of any diabetic kidney disease, defined as persistent albuminuria, persistent reduced eGFR, or both, did not significantly change over time from 28.4% (95% CI, 23.8%-32.9%) in 1988-1994 to 26.2% (95% CI, 22.6%-29.9%) in 2009-2014 (prevalence ratio, 0.95 [95% CI, 0.86-1.06] adjusting for age, sex, and race/ethnicity; P = .39 for trend). However, the prevalence of albuminuria decreased progressively over time from 20.8% (95% CI, 16.3%-25.3%) in 1988-1994 to 15.9% (95% CI, 12.7%-19.0%) in 2009-2014 (adjusted prevalence ratio, 0.76 [95% CI, 0.65-0.89]; P < .001 for trend). In contrast, the prevalence of reduced eGFR increased from 9.2% (95% CI, 6.2%-12.2%) in 1988-1994 to 14.1% (95% CI, 11.3%-17.0%) in 2009-2014 (adjusted prevalence ratio, 1.61 [95% CI, 1.33-1.95] comparing 2009-2014 with 1988-1994; P < .001 for trend), with a similar pattern for severely reduced eGFR (adjusted prevalence ratio, 2.86 [95% CI, 1.38-5.91]; P = .004 for trend). Significant heterogeneity in the temporal trend for albuminuria was noted by age (P = .049 for interaction) and race/ethnicity (P = .007 for interaction), with a decreasing prevalence of albuminuria observed only among adults younger than 65 years and non-Hispanic whites, whereas the prevalence of reduced GFR increased without significant differences by age or race/ethnicity. In 2009-2014, approximately 8.2 million adults with diabetes (95% CI, 6.5-9.9 million adults) had albuminuria, reduced eGFR, or both.

Conclusions and relevance

Among US adults with diabetes from 1988 to 2014, the overall prevalence of diabetic kidney disease did not change significantly, whereas the prevalence of albuminuria declined and the prevalence of reduced eGFR increased.

The adoption of electronic health records (EHRs) has made patient data increasingly accessible, precipitating the development of various clinical decision support systems and data-driven models to help physicians. However, missing data are common in EHR-derived datasets, which can introduce significant uncertainty, if not invalidating the use of a predictive model. Machine learning (ML)-based imputation methods have shown promise in various domains for the task of estimating values and reducing uncertainty to the point that a predictive model can be employed. We introduce Autopopulus, a novel framework that enables the design and evaluation of various autoencoder architectures for efficient imputation on large datasets. Autopopulus implements existing autoencoder methods as well as a new technique that outputs a range of estimated values (rather than point estimates), and demonstrates a workflow that helps users make an informed decision on an appropriate imputation method. To further illustrate Autopopulus' utility, we use it to identify not only which imputation methods can most accurately impute on a large clinical dataset, but to also identify the imputation methods that enable downstream predictive models to achieve the best performance for prediction of chronic kidney disease (CKD) progression.

Aims

To determine if serum levels of serum amyloid A (SAA) predict death and end-stage renal disease in a cohort of people with diabetic kidney disease.

Methods

In a longitudinal cohort study of 135 participants with type 2 diabetes and diabetic kidney disease, serum samples were assayed for SAA. Censored time-to-event analyses in Cox-proportional hazard models were utilized to assess SAA as a predictor of the primary outcome of death and end-stage renal disease.

Results

Participants were 73% Mexican-American (99/135) and 55% men (75/135), with a mean±SD age of 57±7.5years. At baseline, participants had hemoglobin A1c of 8.6±2.3%, systolic blood pressure of 153±27mm Hg, body mass index of 31±9kg/m², median urine-albumin-to-creatinine ratio of 1861mg/g (interquartile range 720-3912mg/g), and estimated glomerular filtration rate of 55.7±22.3ml/min/1.73m². Over a median duration of follow-up of 3.5years, 44% (60/135) of participants experienced a primary outcome event. The hazards ratio for the primary outcome was 3.03 (95% CI 1.43-6.40, p=0.003) in the highest (>1.0 μg/ml) compared to the lowest (<0.55 μg/ml) SAA tertile in a model adjusted for urine-albumin-to-creatinine ratio, estimated glomerular filtration rate, age, sex, and race/ethnicity. Addition of SAA as a covariate improved the model C-statistic (Δ c=0.017).

Conclusions

In a longitudinal cohort study of participants with type 2 diabetes and DKD, higher levels of serum SAA predicted higher risk of death and ESRD. SAA is a promising targetable biomarker for DKD.

People with diabetes and chronic kidney disease (CKD) are at high risk for kidney failure, atherosclerotic cardiovascular disease, heart failure, and premature mortality. Recent clinical trials support new approaches to treat diabetes and CKD. The 2022 American Diabetes Association (ADA) Standards of Medical Care in Diabetes and the Kidney Disease: Improving Global Outcomes (KDIGO) 2022 Clinical Practice Guideline for Diabetes Management in Chronic Kidney Disease each provide evidence-based recommendations for management. A joint group of ADA and KDIGO representatives reviewed and developed a series of consensus statements to guide clinical care from the ADA and KDIGO guidelines. The published guidelines are aligned in the areas of CKD screening and diagnosis, glycemia monitoring, lifestyle therapies, treatment goals, and pharmacologic management. Recommendations include comprehensive care in which pharmacotherapy that is proven to improve kidney and cardiovascular outcomes is layered on a foundation of healthy lifestyle. Consensus statements provide specific guidance on use of renin-angiotensin system inhibitors, metformin, sodium-glucose cotransporter 2 inhibitors, glucagon-like peptide 1 receptor agonists, and a nonsteroidal mineralocorticoid receptor antagonist. These areas of consensus provide clear direction for implementation of care to improve clinical outcomes of people with diabetes and CKD.

Importance

Chronic kidney disease (CKD) is serious and common, yet recognition and public health responses are limited.

Objective

To describe clinical features of, prevalence of, major risk factors for, and care for CKD among patients treated in 2 large US health care systems.

Design, setting, and participants

This cohort study collected data from the Center for Kidney Disease Research, Education, and Hope (CURE-CKD) registry, an electronic health record-based registry jointly curated and sponsored by Providence St Joseph Health and the University of California, Los Angeles. Patients were adults and children with CKD (excluding end-stage kidney disease) and adults at risk of CKD (ie, with diabetes, hypertension, or prediabetes) identified by laboratory values, vital signs, prescriptions, and administrative codes. Data were collected from January 2006 through December 2017, with analyses performed from March 2019 through November 2019.

Exposures

Diabetes, hypertension, and prediabetes.

Main outcomes and measures

Clinical and demographic characteristics, prevalence, and prescribed medications.

Results

Of 2 625 963 adults and children in the sample, 606 064 adults (23.1%) with CKD had a median (interquartile range [IQR]) age of 70 (59-81) years, with 338 785 women (55.9%) and 434 474 non-Latino white individuals (71.7%). A total of 12 591 children (0.4%) with CKD had a median (IQR) age of 6 (1-13) years, with 7079 girls (56.2%) and 6653 non-Latino white children (52.8%). Median (IQR) estimated glomerular filtration rate was 53 (41-61) mL/min/1.73 m2 among adults and 70 (50-95) mL/min/1.73 m2 in children. Prevalence rates for CKD in adults were 4.8% overall (606 064 of 12 669 700) with 1.6% (93 644 of 6 011 129) during 2006 to 2009, 5.7% (393 455 of 6 903 084) during 2010 to 2013, and 8.4% (683 574 of 8 179 860) during 2014 to 2017 (P < .001). A total of 226 693 patients (37.4%) had category 3a CKD; 100 239 (16.5%), category 3b CKD; 39 125 (6.5%), category 4 CKD; and 20 328 (3.4%), category 5 CKD. Among adults with CKD, albuminuria and proteinuria assessments were available in 52 551 (8.7%) and 25 035 (4.1%) patients, respectively. A renin-angiotensin system inhibitor was prescribed to 124 575 patients (20.6%), and 204 307 (33.7%) received nonsteroidal anti-inflammatory drugs or proton pump inhibitors. Of 1 973 258 adults (75.1%) at risk, one-quarter had diabetes or prediabetes (512 299 [26.0%]), nearly half had hypertension (955 812 [48.4%]), and one-quarter had both hypertension and diabetes or prediabetes (505 147 [25.6%]).

Conclusions and relevance

This registry-based cohort study revealed a burgeoning number of patients with CKD and its major risk factors. Rates of identification and use of kidney protective agents were low, while potential nephrotoxin use was widespread, underscoring the pressing need for practice-based improvements in CKD prevention, recognition, and treatment.

Aim

Guideline-directed medical therapy (GDMT) is designed to improve clinical outcomes. The study aim was to assess GDMT prescribing rates and prescribing-persistence predictors in patients with diabetes and chronic kidney disease (CKD) from the Center for Kidney Disease Research, Education, and Hope Registry.

Materials and methods

Data were obtained from adults ≥18 years old with diabetes and CKD between 1 January 2019 and 31 December 2020 (N = 39 158). Baseline and persistent (≥90 days) prescriptions for GDMT, including angiotensin converting enzyme (ACE) inhibitor/angiotensin receptor blocker (ARB), sodium-glucose cotransporter-2 (SGLT2) inhibitor and glucagon-like peptide 1 (GLP-1) receptor agonist were assessed.

Results

The population age (mean ± SD) was 70 ± 14 years, and 49.6% (n = 19 415) were women. Baseline estimated glomerular filtration rate (2021 CKD-Epidemiology Collaboration creatinine equation) was 57.5 ± 23.0 ml/min/1.73 m² and urine albumin/creatinine 57.5 mg/g (31.7-158.2; median, interquartile range). Baseline and ≥90-day persistent prescribing rates, respectively, were 70.7% and 40.4% for ACE inhibitor/ARB, 6.0% and 5.0% for SGLT2 inhibitors, and 6.8% and 6.3% for GLP-1 receptor agonist (all p < .001). Patients lacking primary commercial health insurance coverage were less likely to be prescribed an ACE inhibitor/ARB [odds ratio (OR) = 0.89; 95% confidence interval (CI) 0.84-0.95; p < .001], SGLT2 inhibitor (OR 0.72; 95% CI 0.64-0.81; p < .001) or GLP-1 receptor agonist (OR 0.89; 95% CI 0.80-0.98; p = .02). GDMT prescribing rates were lower at Providence than UCLA Health.

Conclusions

Prescribing for GDMT was suboptimal and waned quickly in patients with diabetes and CKD. Type of primary health insurance coverage and health system were associated with GDMT prescribing.

Background

Uromodulin is a kidney-derived glycoprotein and putative tubular function index. Lower serum uromodulin was recently associated with increased risk for kidney allograft failure in a preliminary, longitudinal single-center -European study involving 91 kidney transplant recipients (KTRs).

Methods

The Folic Acid for Vascular Outcome Reduction in Transplantation (FAVORIT) trial is a completed, large, multiethnic controlled clinical trial cohort, which studied chronic, stable KTRs. We conducted a case cohort analysis using a randomly selected subset of patients (random subcohort, n = 433), and all individuals who developed kidney allograft failure (cases, n = 226) during follow-up. Serum uromodulin was determined in this total of n = 613 FAVORIT trial participants at randomization. Death-censored kidney allograft failure was the study outcome.

Results

The 226 kidney allograft failures occurred during a median surveillance of 3.2 years. Unadjusted, weighted Cox proportional hazards modeling revealed that lower serum uromodulin, tertile 1 vs. tertile 3, was associated with a threefold greater risk for kidney allograft failure (hazards ratio [HR], 95% CI 3.20 [2.05-5.01]). This association was attenuated but persisted at twofold greater risk for allograft failure, after adjustment for age, sex, smoking, allograft type and vintage, prevalent diabetes mellitus and cardiovascular disease (CVD), total/high-density lipoprotein cholesterol ratio, systolic blood pressure, estimated glomerular filtration rate, and natural log urinary albumin/creatinine: HR 2.00, 95% CI (1.06-3.77).

Conclusions

Lower serum uromodulin, a possible indicator of less well-preserved renal tubular function, remained associated with greater risk for kidney allograft failure, after adjustment for major, established clinical kidney allograft failure and CVD risk factors, in a large, multiethnic cohort of long-term, stable KTRs.