Of the nearly 600,000 people in the U.S. who receive dialysis for chronic kidney failure, >60% have diabetes. People receiving dialysis who have diabetes have worse overall and cardiovascular survival rates than those without diabetes. Diabetes care in the dialysis setting is complicated by kidney failure–related factors that render extrapolation of glycated hemoglobin (HbA1c) targets to the dialysis population unreliable and may change the risk-benefit profiles of glucose-lowering and disease-modifying therapies. No prospective studies have established the optimal glycemic targets in the dialysis population, and few randomized clinical trials of glucose-lowering medications included individuals receiving dialysis. Observational data suggest that both lower and higher HbA1c are associated with mortality in the dialysis population. Existing data suggest the potential for safety and effectiveness of some glucose-lowering medications in the dialysis population, but firm conclusions are hindered by limitations in study design and sample size. While population-specific knowledge gaps about optimal glycemic targets and diabetes medication safety and effectiveness preclude the extension of all general population diabetes guidelines to the dialysis-dependent diabetes population, these uncertainties should not detract from the importance of providing person-centered diabetes care to people receiving dialysis. Diabetes care for individuals with and without dialysis-dependent kidney failure should be holistic, based on individual preferences and prognoses, and tailored to integrate established treatment approaches with proven benefits for glycemic control and cardiovascular risk reduction. Additional research is needed to inform how recent pharmacologic and technological advances can be applied to support such individualized care for people receiving maintenance dialysis.
Introduction
Diabetes is the primary cause of kidney failure for nearly half of the ∼600,000 people in the U.S. who have dialysis-dependent kidney failure (1). Despite enormous Medicare investment in the dialysis population (>$50 billion in 2019), people receiving maintenance dialysis have exceedingly high mortality rates (160 deaths/1,000 person-years, median survival of 48 months), primarily driven by cardiovascular causes (2). Compared with individuals with dialysis-dependent kidney failure without diabetes, people with dialysis-dependent kidney failure with diabetes have a 63% higher risk of cardiovascular mortality and a 40% higher risk of all-cause mortality (3).
Although observational data suggest that glycemic control is associated with lower risk of mortality in the dialysis population, population-specific evidence to guide diabetes management in dialysis-dependent kidney failure is limited. Kidney failure–related factors, such as chronic uremia and anemia, contribute to the inaccuracy of glycated hemoglobin (HbA1c) and render extrapolation of HbA1c targets from the general to dialysis populations unreliable. Additionally, given differences in drug metabolism, cardiovascular and hypoglycemia risk profiles, and dialysis-related symptoms, diabetes medications that are safe, effective, and well tolerated in the nondialysis population may not have the same risk-benefit profiles in people receiving dialysis (4,5). However, people with diabetes receiving dialysis have been underrepresented in diabetes clinical trials investigating glycemic targets, glucose-lowering medications, and cardiovascular outcomes, hindering population-specific guideline development. While the standard of diabetes care in the general population has progressed to focus on comorbid disease risk reduction in addition to glycemia (e.g., cardiorenal risk reduction and weight management), population-specific knowledge gaps about optimal glycemic targets and diabetes medication safety and effectiveness preclude the extension of all general population diabetes guidelines to the dialysis-dependent diabetes population.
In this Perspective, we summarize the available evidence regarding glycemic treatment targets and glucose-lowering medication selection among people with dialysis-dependent kidney failure, consider subgroups that may require nuanced management, and highlight the need to approach diabetes care in the dialysis population with the same holistic, person-centered approach that has become the standard of diabetes care in the nondialysis population.
Glucose Control in Dialysis-Dependent Kidney Failure: Does It Matter?
Landmark randomized controlled trials (RCTs) conclusively demonstrate that intensive glycemic control, defined as targeting an HbA1c <6–7%, reduces the risk of microvascular disease, including retinopathy, neuropathy, and nephropathy, in the nondialysis population with type 1 or type 2 diabetes (6). However, targeting lower HbA1c levels (<6%) (primarily by intensifying treatment with insulin) versus standard HbA1c levels (7–7.9%) is associated with higher mortality in the nondialysis diabetes population (7). Guideline bodies such as the American Diabetes Association (ADA), European Association for the Study of Diabetes (EASD), and the American Association of Clinical Endocrinology (AACE) thus recommend targeting HbA1c of ∼6.5–7% in most people with diabetes. Less stringent targets are recommended for people at high risk for hypoglycemia (i.e., those with frequent hypoglycemic episodes or hypoglycemia unawareness) and people with shortened life expectancies. Individuals with dialysis-dependent kidney failure have higher risks of hypoglycemia and mortality than those without, but most guideline bodies do not specify HbA1c targets for this population.
Uncertainty about the accuracy of HbA1c in the setting of dialysis-dependent kidney failure is one barrier to establishing population-specific HbA1c targets. Decreased erythrocyte turnover in the setting of iron-deficient anemia can raise HbA1c values. On the other hand, erythropoietin-stimulating agents, prescribed to 92% of people receiving hemodialysis and 75% of people receiving peritoneal dialysis in the U.S., can lower HbA1c values artificially by promoting red blood cell turnover (8,9). Red blood cell lysis induced by kidney failure–related anemia or the dialysis procedure itself can also falsely lower HbA1c values. HbA1c can therefore be biased higher or lower, making it difficult to recommend strict HbA1c targets among people with diabetes receiving dialysis.
No RCTs have tested intensive versus conventional glycemic control in people receiving dialysis, but observational studies suggest that both higher and lower HbA1c values associate with mortality in this population. In a study of 23,618 people with diabetes receiving maintenance hemodialysis in the U.S. (2001–2003), HbA1c >10% was associated with higher risks of all-cause and cardiovascular death compared with HbA1c 5–6%; hazard ratios (HR) were 1.41 (95% CI, 1.25–1.60) and 1.73 (95% CI, 1.44–2.08), respectively. Risk of all-cause mortality increased incrementally with progressively higher HbA1c (7–7.9%, 8–8.9%, 9–9.9%, and ≥10% vs. 5.0–5.9%) (10). Other observational studies, including studies conducted outside the U.S. and studies of people receiving peritoneal dialysis, also showed associations between higher HbA1c levels and poorer outcomes (11–14). For example, in a German study of 1,255 people with diabetes receiving hemodialysis, HbA1c >8% was associated with a >2-fold risk of sudden death compared with HbA1c ≤6% (HR 2.14 [95% CI, 1.33–3.44]) (15). Similarly, time-averaged HbA1c ≥8% (vs. HbA1c 6.0–6.9%) was associated with higher risk of all-cause mortality in people treated with peritoneal dialysis (11). These findings should be interpreted with caution, as the studies used varied comparator HbA1c levels, several of which used HbA1c <6%, which is lower than the recommended HbA1c targets in the general population.
Observational data also suggest that very low HbA1c levels associate with adverse outcomes (12,16–19). Among 9,201 people with diabetes receiving hemodialysis from 12 countries, mortality risk increased incrementally at lower and higher HbA1c levels from the comparator HbA1c, 7–7.9%. The observed risks were higher for HbA1c <5% (HR 1.25 [95% CI, 1.09–1.67]) and >9% (HR 1.38 [95% CI, 1.11–1.71]) compared with 7–7.9% (16). In a 2013 meta-analysis of 10 international studies (9 observational and 1 secondary analysis of an RCT; n = 83,684 patients), HbA1c ≥8.5% (vs. 6.5–7.4%) was associated with higher mortality (HR 1.14 [95% CI, 1.09–1.19]). HbA1c ≤5.4% (vs. 6.5–7.4%) trended toward an association with higher mortality, but the association did not reach statistical significance (19). More recent observational studies showed similar findings (18,20).
Taken together, these data suggest that HbA1c levels at the extremes associate with mortality in the dialysis population. However, interpretation of the data is complicated by inaccuracies in HbA1c measurement and potential confounding from markers of advanced disease, such as underlying cardiovascular disease burden and/or frailty. As such, the optimal glycemic target range in the dialysis population has not been established.
Glucose Targets in Dialysis-Dependent Kidney Failure Are Ambiguous
Individualization of therapy goals is the cornerstone of contemporary guideline-directed diabetes care in the nondialysis diabetes population. Diabetes organizations, including the ADA, EASD, and AACE, recommend intensive glucose control (HbA1c <6.5–7%) to minimize the risk of microvascular complications among most people with diabetes (21–23). Less stringent targets (HbA1c <8%) are recommended for people with higher risk of hypoglycemia and people with shorter life expectancies, although glucose-lowering agents with cardiovascular benefits are indicated irrespective of HbA1c targets. While the 2024 ADA Standards of Care in Diabetes do not specify a target HbA1c range for individuals with dialysis-dependent kidney disease or cite dialysis population–specific evidence, they do classify kidney failure as a major risk factor for hypoglycemia, implying that HbA1c targets should be less stringent among people receiving dialysis.
Major kidney health organizations recommend HbA1c testing as a part of routine diabetes care among people with dialysis-dependent kidney failure but do not specify HbA1c targets, citing lack of population-specific evidence (24–26). Kidney Disease Improving Global Outcomes (KDIGO) recommends an individualized HbA1c target between <6.5% and <8% for people living with diabetes and any stage of chronic kidney disease (CKD) not treated with dialysis while recognizing the limitations of HbA1c accuracy and precision in advanced CKD (24,27,28). The Joint British Diabetes Societies and U.K. Kidney Association offer the most comprehensive guideline for diabetes care in people receiving dialysis, suggesting that HbA1c <6% or >9.5% represents insufficient control (29). Table 1 summarizes relevant key points from the current guidelines.
Guideline recommendations for diabetes care in individuals with dialysis-dependent kidney failure
Organization (year and reference no.) . | HbA1c target (%) . | Dialysis-related guidance or commentary . | Guideline body–identified evidence gaps . | ||
---|---|---|---|---|---|
Recommended for most people not treated with dialysis . | Recommended for people treated with dialysis . | Implied for people treated with dialysis . | |||
ADA and EASD (2022) (21,22) | <7 (nonpregnant adults with at least 10 years’ life expectancy) | No target specified | Less stringent goals for those with shortened life expectancy, risks associated with hypoglycemia, important comorbidities, or vascular complications | Acknowledgment of limited data to guide best practices in people treated with dialysis. | Trials of therapies that provide cardiovascular benefit (GLP-1 RA, GLP-1/GIP RA, and SGLT2i) to identify avenues for mortality reduction in the dialysis population. |
AACE (2022) (23) | ≤6.5 (most nonpregnant adults, if it can be achieved safely) | No target specified | 7–8 in people with advanced kidney disease | No specific guidance on how to achieve targets in people treated with dialysis. | |
KDIGO (2022) (24) | 6.5–8 (people with diabetes and CKD not treated with dialysis) | No target specified | Targets are unknown | GLP-1 RA use for the treatment of obesity in advanced CKD should be considered. | Evidence to support HbA1c targets and use of GLP-1RA in dialysis. |
Kidney Disease Outcomes Quality Initiative (2022, 2024) (25,26)† | 6.5–8 (people with diabetes and CKD not treated with dialysis) | No target specified | Targets are unknown | Strong support for use of incretin therapy to treat obesity in CKD. | Evidence to support HbA1c targets and use of GLP-1RA in dialysis. |
Joint British Diabetes Societies/U.K. Kidney Association (2022) (29) | No target specified | No target specified | HbA1c noted to be unreliable in people treated with dialysis, but identifies HbA1c <6 or >9.5 as poor control | Cautious use of GLP-1RA in dialysis is suggested, citing some safety evidence with the acknowledgment that GLP-1 RA are not licensed for use in people with eGFR <15 mL/min/1.73 m2 in the U.K. | Evidence about management of diabetes in people treated with peritoneal dialysis. |
Organization (year and reference no.) . | HbA1c target (%) . | Dialysis-related guidance or commentary . | Guideline body–identified evidence gaps . | ||
---|---|---|---|---|---|
Recommended for most people not treated with dialysis . | Recommended for people treated with dialysis . | Implied for people treated with dialysis . | |||
ADA and EASD (2022) (21,22) | <7 (nonpregnant adults with at least 10 years’ life expectancy) | No target specified | Less stringent goals for those with shortened life expectancy, risks associated with hypoglycemia, important comorbidities, or vascular complications | Acknowledgment of limited data to guide best practices in people treated with dialysis. | Trials of therapies that provide cardiovascular benefit (GLP-1 RA, GLP-1/GIP RA, and SGLT2i) to identify avenues for mortality reduction in the dialysis population. |
AACE (2022) (23) | ≤6.5 (most nonpregnant adults, if it can be achieved safely) | No target specified | 7–8 in people with advanced kidney disease | No specific guidance on how to achieve targets in people treated with dialysis. | |
KDIGO (2022) (24) | 6.5–8 (people with diabetes and CKD not treated with dialysis) | No target specified | Targets are unknown | GLP-1 RA use for the treatment of obesity in advanced CKD should be considered. | Evidence to support HbA1c targets and use of GLP-1RA in dialysis. |
Kidney Disease Outcomes Quality Initiative (2022, 2024) (25,26)† | 6.5–8 (people with diabetes and CKD not treated with dialysis) | No target specified | Targets are unknown | Strong support for use of incretin therapy to treat obesity in CKD. | Evidence to support HbA1c targets and use of GLP-1RA in dialysis. |
Joint British Diabetes Societies/U.K. Kidney Association (2022) (29) | No target specified | No target specified | HbA1c noted to be unreliable in people treated with dialysis, but identifies HbA1c <6 or >9.5 as poor control | Cautious use of GLP-1RA in dialysis is suggested, citing some safety evidence with the acknowledgment that GLP-1 RA are not licensed for use in people with eGFR <15 mL/min/1.73 m2 in the U.K. | Evidence about management of diabetes in people treated with peritoneal dialysis. |
RA, receptor agonist; SGLT2i, SGLT2 inhibitor.
†KDOQI is a program of the National Kidney Foundation.
Alternative glycemic measures, including glycated albumin, fructosamine, and continuous glucose monitoring (CGM), have been reviewed in detail elsewhere (30–32). In brief, neither glycated albumin nor fructosamine is recommended for assessing glycemic control in the setting of dialysis due to lack of validation and, similar to HbA1c, pitfalls in accuracy. Although no large studies have evaluated the accuracy of CGM in the dialysis population, small studies suggest promise for CGM among people receiving dialysis (30–33). Some organizations, such as the Joint British Diabetes Societies and U.K. Kidney Association, advocate for CGM in addition to routine HbA1c measurement to prevent hypoglycemia, although neither organization addresses how to interpret potential discrepancies in the two measures.
CGM display of real-time glucose levels allows for response to rapid changes in glycemia and impending hypoglycemia, which is particularly important in the dialysis population as glycemic crises are common (34). CGM metrics (e.g., time in range [TIR] 70–180 mg/dL and time below range <70 mg/dL) can inform adjustments to diabetes care plans. In the general diabetes population, HbA1c strongly correlates with the percent TIR, and a goal of >70% TIR and <4% time below range is recommended for most people living with diabetes. Whether the same CGM targets should be applied to people receiving dialysis has not been established, but CGM can reveal glycemic trends, including different glycemic profiles on dialysis versus nondialysis days, that may guide medication adjustments. However, CGM data should be interpreted with caution, as dialysis-associated fluid shifts may contribute to CGM inaccuracy (32). When CGM is employed, use of capillary glucose checks should be encouraged in situations of potential discordance, such as symptomatic hypoglycemia with normal CGM glucose values or CGM glucose <70 mg/dL without symptoms.
While most people with dialysis-dependent kidney failure likely require less intensive glycemic control than is recommended for people not receiving dialysis, fixation on the prevention of hypoglycemia may result in harmful hyperglycemia or inhibit use of disease-modifying therapies that are indicated irrespective of HbA1c. Further, it is plausible that younger individuals, people who have been treated with dialysis for >4 years (i.e., have survived longer than median survival times), and individuals awaiting kidney transplant may benefit from more stringent glycemic targets. In 2020, nearly 30% of the U.S. dialysis population was under 34 years old (35), >40% of the dialysis population had been treated with dialysis for >4 years (35), and >70,000 people were listed for kidney transplant (2). Other factors, such as burden of cardiovascular disease, dementia status, retinopathy, neuropathy, and frailty, should be considered when determining optimal diabetes treatment regimens at the individual level (36,37). Individualization of target HbA1c levels, like that recommended by the ADA for the general population, is prudent (Fig. 1).
Characteristics that may inform individualized glycemic management in the dialysis population. Characteristics on the left may justify more stringent targets, and characteristics on the right may justify less stringent HbA1c targets (21).
Characteristics that may inform individualized glycemic management in the dialysis population. Characteristics on the left may justify more stringent targets, and characteristics on the right may justify less stringent HbA1c targets (21).
FDA Approval and Labeling of Glucose-Lowering Medications
Table 2 displays the glucose-lowering medications that are approved by the U.S. Food and Drug Administration (FDA) for use in the setting of dialysis-dependent kidney failure. Diabetes medications excreted by the kidneys, such as metformin, are contraindicated in the setting of dialysis-dependent kidney failure (38,39). Sodium–glucose cotransporter 2 (SGLT2) inhibitors may be continued in people with type 2 diabetes and advanced CKD who transition to dialysis but are not recommended for initiation in people treated with dialysis (24). However, medications not excreted by the kidneys are presumed by the FDA to have similar safety and effectiveness profiles among individuals with and without kidney dysfunction, and dedicated studies in people with impaired kidney function are not required (40). Therefore, regardless of the availability of population-specific safety and effectiveness data, medications not excreted by the kidneys, such as the most potent glucagon-like peptide 1 (GLP-1) receptor agonists (e.g., semaglutide and tirzepatide), do not have dialysis-specific warnings or contraindications (40). However, it is plausible that the safety and effectiveness profiles of these medications are altered in the setting of kidney failure irrespective of the excretion pathway. For example, impaired gastrointestinal drug absorption and/or hypoalbuminemia-induced decrements in nonkidney drug clearance may increase free concentrations of drugs, changing drug bioavailability and safety profiles (41,42). These factors, among others, suggest that drug safety and effectiveness differ between individuals with and without dialysis-dependent kidney failure and thus require dedicated study in the dialysis population (4,5).
Package insert instructions for glucose-lowering medications approved by the FDA for use in dialysis-dependent kidney failure
Medication . | Dose adjustment in the setting of dialysis dependence . | Package insert instructions for use in severe kidney impairment and/or dialysis . | Dialysis-specific clinical studies cited in package insert? . |
---|---|---|---|
Insulins | |||
Insulin | Titrate to individualized goals | Frequent glucose monitoring and dose adjustment | No |
Incretin therapy | |||
Dulaglutide (90) | None | Use with caution | No |
Liraglutide (91) | None | Use caution when initiating and escalating; cites limited experience in dialysis-dependent kidney disease | No |
Semaglutide (92) | None | Not addressed | Yes, pharmacokinetic studies |
Tirzepatide (93) | None | Not addressed | Yes, pharmacokinetic studies |
DPP-4 inhibitor | |||
Alogliptin (94) | Maximum 6.25 mg daily | May be administered without regard to timing of dialysis | Yes, pharmacokinetic studies |
Linagliptin (95) | No dose adjustment required | Use with insulin may associate with higher rate of hypoglycemia; reduction of insulin or insulin secretagogue may be required when used in combination with linagliptin | Yes, pharmacokinetic studies |
Saxagliptin (96) | Maximum 2.5 mg daily | Administer after hemodialysis | Yes, pharmacokinetic studies plus statement that saxagliptin has not been studied in peritoneal dialysis |
Sitagliptin (97) | Maximum 25 mg once daily | May be administered without regard to timing of dialysis | Yes (see Table 3 for details) |
Sulfonylurea | |||
Glimepiride (98) | None | To minimize risk of hypoglycemia, start with 1-mg dose and titrate to individualized goals | No |
Glipizide (99) | None | To minimize risk of hypoglycemia, start with 2.5-mg dose and titrate to individualized goals | No |
TZD | |||
Pioglitazone (100) | No dose adjustment required | Not addressed | No |
Medication . | Dose adjustment in the setting of dialysis dependence . | Package insert instructions for use in severe kidney impairment and/or dialysis . | Dialysis-specific clinical studies cited in package insert? . |
---|---|---|---|
Insulins | |||
Insulin | Titrate to individualized goals | Frequent glucose monitoring and dose adjustment | No |
Incretin therapy | |||
Dulaglutide (90) | None | Use with caution | No |
Liraglutide (91) | None | Use caution when initiating and escalating; cites limited experience in dialysis-dependent kidney disease | No |
Semaglutide (92) | None | Not addressed | Yes, pharmacokinetic studies |
Tirzepatide (93) | None | Not addressed | Yes, pharmacokinetic studies |
DPP-4 inhibitor | |||
Alogliptin (94) | Maximum 6.25 mg daily | May be administered without regard to timing of dialysis | Yes, pharmacokinetic studies |
Linagliptin (95) | No dose adjustment required | Use with insulin may associate with higher rate of hypoglycemia; reduction of insulin or insulin secretagogue may be required when used in combination with linagliptin | Yes, pharmacokinetic studies |
Saxagliptin (96) | Maximum 2.5 mg daily | Administer after hemodialysis | Yes, pharmacokinetic studies plus statement that saxagliptin has not been studied in peritoneal dialysis |
Sitagliptin (97) | Maximum 25 mg once daily | May be administered without regard to timing of dialysis | Yes (see Table 3 for details) |
Sulfonylurea | |||
Glimepiride (98) | None | To minimize risk of hypoglycemia, start with 1-mg dose and titrate to individualized goals | No |
Glipizide (99) | None | To minimize risk of hypoglycemia, start with 2.5-mg dose and titrate to individualized goals | No |
TZD | |||
Pioglitazone (100) | No dose adjustment required | Not addressed | No |
Evidence for Use of Specific Glucose-Lowering Medications in the Dialysis Population
Table 3 summarizes published clinical trials of glucose lowering medications that included individuals with dialysis-dependent kidney failure.
Prospective clinical trials of glucose-lowering medications approved by the FDA for use in dialysis-dependent kidney failure*
Medication class and country of origin (publication year and reference no.) . | Primary outcome . | Study group drug treatment (no. of individuals treated with dialysis) . | Control group drug treatment (no. of individuals treated with dialysis) . | Design . | Efficacy . | Safety . |
---|---|---|---|---|---|---|
TZD | ||||||
U.S. (2005) (51) | Difference in preprandial SMBG, HbA1c, and insulin use at month 6 | Troglitazone (6) | Placebo (6) | Randomized, open label | • No difference in SMBG or HbA1c • Reduction in insulin use | No change in weight or liver function tests |
Japan (2008) (52) | HbA1c at week 12 | Pioglitazone (20) | Not applicable | Open label | • Reduction in HbA1c and plasma glucose at 4 weeks (0.7%) | No increase in adverse events; no difference in interdialytic weight gain |
DPP-4 inhibitor | ||||||
Multinational (2008) (54) | • HbA1c at week 12 (sitagliptin vs. placebo) • HbA1c at week 54 (sitagliptin vs. placebo /glipizide) | Sitagliptin (8, HD; 4, PD) | Placebo, 0–12 weeks; glipizide, 13–54 weeks (3, HD; 2, PD) | Randomized, double blind | • Reduction in HbA1c was 0.6% with sitagliptin and 0.2% with placebo in total study population (n = 91, eGFR <50 mL/min/1.73 m2) at 12 weeks • Reduction in HbA1c (0.7% for sitagliptin and 1% for placebo/glipizide) in total study population (n = 91, eGFR <50 mL/min/1.73 m2) at 52 weeks • No subanalysis of participants with ESRD | Adverse events similar between study arms in the first 12 weeks (sitagliptin vs. placebo); a higher proportion of participants in the control arm who switched to glipizide after 12 weeks experienced hypoglycemia over 54 weeks |
Multinational (2013) (48) | HbA1c at week 54 | Sitagliptin (64) | Glipizide (65) | Randomized, double blind | • Similar reduction in HbA1c with sitagliptin compared with glipizide | Nonstatistically significant increase in hypoglycemia with glipizide |
Japan (2016) (57) | GA at week 24 | Saxagliptin (41) | Usual care (41) | Open label | • Reduction in GA (3.4%) | No evidence of hypoglycemia or liver dysfunction |
Multinational (2011) (56) | HbA1c at week 12 | Saxagliptin (19) | Placebo (20) | Randomized, double blind | • No difference in HbA1c between arms in participants treated with dialysis | Adverse events similar between study arms |
Japan (2020) (55) | HbA1c at week 12 | Trelagliptin (40) | Placebo (39) | Randomized, double blind | • Least-square mean difference in HbA1c for trelagliptin compared with placebo was −0.72% in total study population (n = 107, eGFR <30 mL/min/1.73 m2) • No subanalysis of participants with ESRD | Numerically more hypoglycemia was observed in the trelagliptin group |
GLP-1 receptor agonist | ||||||
Japan (2015) (58) | HbA1c and GA at week 52 | Liraglutide (0.6–0.9 mg; 15, PD) | Usual care (15, PD) | Prospective observational cohort | • No change in HbA1c or GA in either group • Reduction in postpostprandial and mean daily blood glucose at 6 and 12 months compared with baseline | Gastrointestinal adverse events |
Denmark (2016) (60,61) | Dose-corrected trough concentration of liraglutide | Liraglutide (14) | Placebo (10) | Randomized, double blind | • Plasma concentration of liraglutide increased by 49% • Trend toward improved glycemia | Increased hypoglycemia; increased gastrointestinal adverse events; weight loss |
Japan (2018) (59) | Body composition via bioimpedance analysis | Dulaglutide (11) | Tenegliptin (10) | Open label | • Decreased fat mass • Improved glycemia | Decreased lean muscle mass |
Medication class and country of origin (publication year and reference no.) . | Primary outcome . | Study group drug treatment (no. of individuals treated with dialysis) . | Control group drug treatment (no. of individuals treated with dialysis) . | Design . | Efficacy . | Safety . |
---|---|---|---|---|---|---|
TZD | ||||||
U.S. (2005) (51) | Difference in preprandial SMBG, HbA1c, and insulin use at month 6 | Troglitazone (6) | Placebo (6) | Randomized, open label | • No difference in SMBG or HbA1c • Reduction in insulin use | No change in weight or liver function tests |
Japan (2008) (52) | HbA1c at week 12 | Pioglitazone (20) | Not applicable | Open label | • Reduction in HbA1c and plasma glucose at 4 weeks (0.7%) | No increase in adverse events; no difference in interdialytic weight gain |
DPP-4 inhibitor | ||||||
Multinational (2008) (54) | • HbA1c at week 12 (sitagliptin vs. placebo) • HbA1c at week 54 (sitagliptin vs. placebo /glipizide) | Sitagliptin (8, HD; 4, PD) | Placebo, 0–12 weeks; glipizide, 13–54 weeks (3, HD; 2, PD) | Randomized, double blind | • Reduction in HbA1c was 0.6% with sitagliptin and 0.2% with placebo in total study population (n = 91, eGFR <50 mL/min/1.73 m2) at 12 weeks • Reduction in HbA1c (0.7% for sitagliptin and 1% for placebo/glipizide) in total study population (n = 91, eGFR <50 mL/min/1.73 m2) at 52 weeks • No subanalysis of participants with ESRD | Adverse events similar between study arms in the first 12 weeks (sitagliptin vs. placebo); a higher proportion of participants in the control arm who switched to glipizide after 12 weeks experienced hypoglycemia over 54 weeks |
Multinational (2013) (48) | HbA1c at week 54 | Sitagliptin (64) | Glipizide (65) | Randomized, double blind | • Similar reduction in HbA1c with sitagliptin compared with glipizide | Nonstatistically significant increase in hypoglycemia with glipizide |
Japan (2016) (57) | GA at week 24 | Saxagliptin (41) | Usual care (41) | Open label | • Reduction in GA (3.4%) | No evidence of hypoglycemia or liver dysfunction |
Multinational (2011) (56) | HbA1c at week 12 | Saxagliptin (19) | Placebo (20) | Randomized, double blind | • No difference in HbA1c between arms in participants treated with dialysis | Adverse events similar between study arms |
Japan (2020) (55) | HbA1c at week 12 | Trelagliptin (40) | Placebo (39) | Randomized, double blind | • Least-square mean difference in HbA1c for trelagliptin compared with placebo was −0.72% in total study population (n = 107, eGFR <30 mL/min/1.73 m2) • No subanalysis of participants with ESRD | Numerically more hypoglycemia was observed in the trelagliptin group |
GLP-1 receptor agonist | ||||||
Japan (2015) (58) | HbA1c and GA at week 52 | Liraglutide (0.6–0.9 mg; 15, PD) | Usual care (15, PD) | Prospective observational cohort | • No change in HbA1c or GA in either group • Reduction in postpostprandial and mean daily blood glucose at 6 and 12 months compared with baseline | Gastrointestinal adverse events |
Denmark (2016) (60,61) | Dose-corrected trough concentration of liraglutide | Liraglutide (14) | Placebo (10) | Randomized, double blind | • Plasma concentration of liraglutide increased by 49% • Trend toward improved glycemia | Increased hypoglycemia; increased gastrointestinal adverse events; weight loss |
Japan (2018) (59) | Body composition via bioimpedance analysis | Dulaglutide (11) | Tenegliptin (10) | Open label | • Decreased fat mass • Improved glycemia | Decreased lean muscle mass |
ESRD, end-stage renal disease; GA, glycated albumin; HD, hemodialysis; PD, peritoneal dialysis; SMBG, self-monitored blood glucose.
*Studies enrolled people treated with dialysis at study initiation. Only prospective, outpatient studies were included. Studies that only examined drug pharmacokinetics were excluded.
Insulin
Insulin was the most prescribed glucose-lowering medication in the dialysis population in the U.S. in 2021 (9). Among people living with type 2 diabetes, ∼42% of people receiving hemodialysis and ∼53% of people receiving peritoneal dialysis were prescribed insulin (9). While there are no published RCTs or cohort studies evaluating the safety and effectiveness of insulin compared with other glucose-lowering agents among people receiving maintenance dialysis, recent data suggest that the type of insulin influences outcomes. In a study among people receiving dialysis who had type 2 diabetes in seven European countries, use of analog (vs. human) insulin was associated with lower risks of all-cause mortality (HR 0.81 [95% CI, 0.66–0.99]), major adverse cardiovascular events (HR 0.82 [95% CI, 0.68–0.98]), and hospitalization (HR 0.76 [95% CI, 0.67–0.86]) (43). However, this information should be interpreted with caution due to the observational nature of the study and potential for confounding from unmeasured factors, such as regional differences in care and patient case mix (44). Such an association has not been shown in the diabetes population not receiving dialysis (45), raising the possibility of differences in safety profiles across the general and dialysis populations.
Insulin-induced hypoglycemia is common in the dialysis population, with individuals receiving insulin and maintenance dialysis experiencing a 32% higher risk of hypoglycemia than those on noninsulin glucose-lowering medications (34). Unique dosing challenges may contribute to these findings. During hemodialysis, the dialysis membrane (i.e., the dialyzer) absorbs insulin, lowering plasma insulin concentrations (46). While higher doses of insulin may be required on hemodialysis treatment days, lower doses may be required on non–hemodialysis treatment days due to dialysis-induced insulin sensitivity (31). One euglycemic clamp study reported a 25% reduction in basal insulin requirement the day following hemodialysis among people with type 2 diabetes (47), but there is no standardized guidance for changing insulin dose based on dialysis day.
Sulfonylureas
Sulfonylureas were the second most prescribed glucose-lowering medications in the U.S. dialysis population in 2021 (7.5% of people with type 2 diabetes receiving hemodialysis and ∼11% of people with type 2 diabetes receiving peritoneal dialysis) (9). As sulfonylureas stimulate non-glucose–dependent insulin secretion, they increase the risk of hypoglycemia. Small studies (described below) have suggested that sulfonylurea use (vs. dipeptidyl peptidase 4 [DPP-4] inhibitor use) is associated with lower blood glucose and hypoglycemia (48). To minimize the risk of hypoglycemia, some experts recommend the use of short-acting glipizide because of its hepatic metabolism and inactive or weakly active metabolites (49,50).
TZD
Thiazolidinediones (TZD), oral peroxisome proliferator–activated receptor γ agonists that improve glycemia via insulin sensitization, are exclusively cleared by the liver and do not increase the risk of hypoglycemia. Few studies have examined their use in dialysis-dependent kidney failure, and they are prescribed infrequently (∼1% of people with type 2 diabetes receiving hemodialysis and ∼2% of people with type 2 diabetes receiving peritoneal dialysis in the U.S. in 2021) (9). An RCT of 12 people with type 2 diabetes receiving hemodialysis showed no reduction in HbA1c at 6 months with add-on therapy with troglitazone versus placebo, but participants randomized to troglitazone required less insulin (troglitazone, −8.4 ± 4.12 units of insulin/day vs. placebo, +4.4 ± 5.47 units of insulin/day; P < 0.05) (51). No notable safety signals were reported, including no evidence of new-onset or worsening heart failure, liver dysfunction, or hypoglycemia. Other studies suggest similar safety and efficacy of TZD among people receiving maintenance dialysis (52).
DPP-4 Inhibitors
From 2012 to 2021, prescription of DPP-4 inhibitors to people with type 2 diabetes receiving dialysis in the U.S. increased from 3% to 8%. (53). Small RCTs evaluating the effectiveness of DPP-4 inhibitors among people receiving dialysis have yielded mixed results (48,54–57). In the largest RCT (n = 129 patients receiving dialysis), sitagliptin and glipizide demonstrated comparable reductions in HbA1c with sitagliptin (−0.72%; 95% CI, −0.95 to 0.48) and with glipizide (−0.87%; 95% CI, 1.11–0.63), with a trend toward less hypoglycemia with sitagliptin (48). Another RCT (n = 170 patients with estimated glomerular filtration rate [eGFR] <50 mL/min/1.73 m2 that included people receiving dialysis), saxagliptin compared with placebo reduced HbA1c at 12 weeks (between-group difference −0.42% [95% CI, −0.71 to −0.12%]) overall but not among the 39 people with dialysis-dependent kidney failure (−0.84% for saxagliptin [n = 19] and −0.87% for placebo [n = 20]). While no studies have raised major safety concerns about DPP-4 inhibitor use in the dialysis population, limited glycemic efficacy and lack of cardiovascular benefit render DPP-4 inhibitors less attractive than newer agents.
GLP-1 Receptor Agonists and Combination Incretin Therapies
GLP-1 receptor agonists improve glycemia without risk of hypoglycemia and reduce the risk of cardiovascular events in people with type 2 diabetes, including people with CKD. The FLOW (Evaluate Renal Function With Semaglutide Once Weekly) trial demonstrated a 24% reduction in the primary outcome of major kidney disease events and an 18% reduction in risk of a confirmatory secondary outcome of major adverse cardiovascular events with use of semaglutide (vs. placebo) among patients with moderate to severe CKD (eGFR 25–75 mL/min/1.73 m2 and albuminuria) and type 2 diabetes. The strength of the benefit was not modified by degree of kidney dysfunction. However, the study excluded patients with dialysis-dependent kidney failure.
In the U.S. in 2021, GLP-1 receptor agonists were prescribed to 3.4% of people with type 2 diabetes receiving hemodialysis and 5% of those receiving peritoneal dialysis. Use of two of the earliest GLP-1 receptor agonists, exenatide and lixisenatide, is contraindicated in dialysis due to their renal clearance. Cautious use is recommended for liraglutide and dulaglutide, once-daily and once-weekly GLP-1 receptor agonists, respectively. Newer agents, including semaglutide and tirzepatide, have no specific recommendations for dose reduction or use in people treated with dialysis (Table 2).
To date, GLP-1 receptor agonist use among individuals receiving dialysis has been investigated in only a few small studies that yielded conflicting results (58–62). The only published RCT that included patients receiving dialysis (n = 47 patients, 24 of whom received hemodialysis or peritoneal dialysis) demonstrated that plasma liraglutide concentrations increased by 49% in patients receiving dialysis compared with control individuals with normal kidney function (61). Liraglutide-randomized patients receiving dialysis (n = 14) had more adverse gastrointestinal events than all other groups (P < 0.04) (61). In a post hoc analysis, glycemic variability and HbA1c were similar between liraglutide and placebo in participants treated with dialysis, but liraglutide treatment was associated with a higher risk of CGM-measured hypoglycemia (time in hypoglycemia defined as blood glucose <70 mg/dL, 5.2% ± 2.6%; control time in hypoglycemia, 0.2% ± 0.1%; P < 0.01) (60). Studies from Japan also suggested some safety concerns (58,59). Compared with tenegliptin, the once-weekly GLP-1 receptor agonist dulaglutide had similar reductions in glycated albumin (−14.7% [95% CI, −20.4 to −3.3] vs.−12.4% [95% CI, −19.5 to −0.6]; P = 0.65), but only the dulaglutide group experienced fat and skeletal muscle mass loss (21.9–18.9 kg, P = 0.04, and 21.0–20.2 kg, P = 0.01, respectively).
Liraglutide- and dulaglutide-induced weight loss raises concerns for use of more potent incretin combinations like tirzepatide, a GLP-1 and gastric inhibitory peptide (GIP) receptor agonist that induces more weight loss than dulaglutide among all users. A single retrospective observational study of 14 people in Japan with type 2 diabetes (average BMI 25.7 kg/m2) who transitioned from dulaglutide to tirzepatide showed improvement in CGM-measured time in range with tirzepatide (50.8% time in glucose range 70–180 mg/dL with tirzepatide versus 42.7% with dulaglutide use; P = 0.02). Body weight decreased 3 months after the individual switched to tirzepatide (−1.5 kg ± 0.3 kg, P < 0.001) (63). Incretin-induced weight loss may be detrimental to people receiving dialysis with compromised nutritional status, as observational data suggest that lighter (vs. heavier) body weight associates with worse survival (64–66). However, it is plausible that for some patients, such as patients with obesity seeking to qualify for kidney transplant, incretin-induced weight loss may be beneficial.
Other patients, such as those with high-risk cardiovascular profiles, may benefit from incretin therapy regardless of diabetes status, as shown among nondialysis patients in the SELECT (Semaglutide Effects on Cardiovascular Outcomes in People With Overweight or Obesity) and STEP-HFpEF DM (Semaglutide Treatment Effect in People With Obesity and Heart Failure With Preserved Ejection Fraction and Diabetes Mellitus) trials (67,68). As novel and more potent incretin therapy combinations (GLP-1 plus GIP, glucagon, amylin analogs, etc.) are being developed, understanding their efficacy and safety, particularly regarding the impact of incretin-induced weight loss on people treated with dialysis, is essential.
The tolerability of GLP-1 receptor agonists by individuals receiving dialysis also has not been established. Gastrointestinal intolerance led to discontinuation of GLP-1 receptor agonist therapy in 10–15% of participants in studies conducted in the nondialysis population (69–71). Data suggest that 70–90% of people treated with dialysis experience gastrointestinal symptoms, potentially unfavorably impacting GLP-1 receptor agonist tolerability, adherence, and, thus, effectiveness (72,73). Limited by lack of population-specific evidence, most major guidelines suggest but do not strongly recommend GLP-1 receptor agonist use among patients receiving dialysis. However, guidelines issued by KDIGO (24,27,28) and the Kidney Disease Outcomes Quality Initiative (25,26) cite the potential for benefit of GLP-1 receptor agonists for individuals with obesity preparing for kidney transplant (Table 1).
SGLT2 Inhibitors
While the glycemic benefit of SGLT2 inhibitors declines as kidney function declines, SGLT2 inhibitors reduce the risk of cardiovascular complications (major adverse cardiovascular events and heart failure) and slow progression of CKD among people with eGFR as low as 20 mL/min/1.73 m2. Initiation of SGLT2 inhibitors is not recommended for people with eGFR <20 mL/min/1.73 m2 (9), but there is considerable interest in SGLT2 inhibitor use in such individuals (74,75). Preserved residual kidney function among people receiving dialysis is an important predictor of lower mortality risk (75). Individuals with dialysis-dependent kidney failure who continue to produce urine might benefit from SGLT2 inhibitors’ proven effectiveness for slowing the decline of kidney function (74). Further, people with dialysis-dependent kidney failure have high burdens of cardiovascular disease (77%, 45%, and 44% of people treated with hemodialysis have known cardiovascular disease, congestive heart failure, and coronary artery disease, respectively [9]) and may benefit from the cardiovascular effects of SGLT2 inhibitors. Off-label use of SGLT2 inhibitors in the dialysis population is becoming increasingly common. According to the U.S. Renal Data System, only 0.2% of individuals receiving hemodialysis and peritoneal dialysis were prescribed an SGLT2 inhibitor in 2021, but this percentage may increase in coming years. With several ongoing RCTs examining the safety and efficacy of SGLT2 inhibitors (including the impact of SGLT2 inhibitors on cardiovascular outcomes, residual kidney function, and glucose homeostasis) in people treated with dialysis, the available evidence is rapidly evolving.
Considerations for Diabetes Management in Special Populations
Type 1 Diabetes
The general type 1 diabetes population is at high risk for both hypo- and hyperglycemia, including severe hypoglycemia and diabetic ketoacidosis (DKA). Kidney failure exacerbates these risks because of loss of renal gluconeogenesis, decreased insulin clearance, and dialysis-related changes in insulin sensitivity (31). In the U.S. dialysis population (2013–2017), young adults with diabetes aged 18–44 years, individuals who are more likely to have type 1 diabetes, experienced threefold higher hypoglycemic crises and more than 50-fold higher hyperglycemic crises compared with people aged ≥75 years. However, in 2017, only 3.5% of people with type 1 diabetes and dialysis-dependent kidney failure received a prescription for glucagon (76). Data from one center suggest that people receiving dialysis who are hospitalized with DKA remain in the hospital twice as long as people with DKA and eGFR >60 mL/min/1.73 m2 (median 7 vs. 3 days, P < 0.001) (77). To our knowledge, there have been no RCTs of promising type 1 diabetes management approaches, such as CGM paired with automated insulin delivery (78), among people with type 1 diabetes and dialysis-dependent kidney failure. Data from the type 1 diabetes population not receiving dialysis suggest that such automated insulin delivery systems increase time in range and limit glycemic variability (78), both important in optimizing diabetes management among people receiving dialysis. In a crossover study of 26 individuals living with type 2 diabetes treated with hemodialysis, the proportion of time in range was 52.8% ± 12.5% with closed-loop treatment versus 37.7% ± 20.5% with control treatment (mean difference, 15.1 percentage points [95% CI, 8.0–22.2]) (79). Whether this extends to people living with type 1 diabetes is unknown, but emerging reports suggest that use of closed-loop systems may be safe and effective in people living with type 1 diabetes and dialysis-dependent kidney failure (80,81). Nonetheless, given the potential for peridialysis changes in insulin sensitivity and clearance rates that affect automated insulin delivery algorithms (32), larger studies of automated insulin delivery systems in people treated with dialysis are required to ensure that the algorithms are equipped to manage the day-to-day glycemic variability experienced by people receiving dialysis. In the meantime, automated insulin delivery systems should be used with caution. Capillary glucose checks are helpful in cases of potential CGM inaccuracy, and an off-pump plan should be established in case of automated insulin delivery system failure (32).
Peritoneal Dialysis
There are two types of dialysis: hemodialysis and peritoneal dialysis. Hemodialysis is used by over 90% of people with kidney failure worldwide, but use of peritoneal dialysis is increasing. Hemodialysis solutions contain physiologic levels of dextrose (on average, 100 mg/dL) and have, if any, a glucose-lowering effect. In peritoneal dialysis, a typically glucose-based dialysis solution is instilled in the abdomen, and transperitoneal membrane diffusive and osmotic forces facilitate toxin and fluid removal and electrolyte homeostasis (82). Higher concentrations of dextrose are used to enhance fluid removal, and solution dextrose concentrations range from 1.5% (76 mmol/L) to 4.25% (214 mmol/L) (82). Systemic dextrose absorption can cause or worsen hyperglycemia. Intraperitoneally administered insulin has been recommended in the past to mitigate hyperglycemia associated with peritoneal dialysis but is no longer preferred due to risk of worsening lipids, hepatic subcapsular steatosis, and peritonitis compared with subcutaneous insulin administration (83). Peritoneal dialysis can be performed as regular, manual exchanges over the course of a day (continuous ambulatory peritoneal dialysis) or as machine-automated exchanges over a shorter period of time (most often at night, automated peritoneal dialysis). Associated variations in time and frequency of exposure to the glucose-containing dialysis solution contribute to glycemic variability. Among patients with high fluid removal needs and dialysis solution–induced hyperglycemia, icodextrin, a slow-resorbing glucose polymer, can be used in place of or in addition to glucose-based solutions. However, icodextrin is known to react with glucose sensor electrodes using the enzymes glucose dehydrogenase pyrroloquinoline quinone and glucose oxidase and can falsely raise glucose readings (31,32). Understanding how dialysis solutions and the peritoneal membrane impact glycemia, glucose-lowering medications, and glucose monitoring is important to overall glycemic management. Evidence for diabetes technology and management specifically in people receiving peritoneal dialysis is limited and requires further study.
Opportunities for Improving Diabetes Care in the Dialysis Setting
Providing Holistic, Person-Centered Care
The paucity of data to guide diabetes treatment decisions in the dialysis population is striking but should not be a hindrance to providing person-centered, goal-directed diabetes care to people with diabetes and dialysis-dependent kidney failure. In its Standards of Care in Diabetes, the ADA recommends a diabetes care plan that not only considers an individual’s comorbidities and prognoses but also prioritizes the individual’s preferences and values. In selecting pharmacotherapy for type 2 diabetes, the ADA recommends using shared decision-making and considering the agent’s effect on cardiovascular and kidney comorbidities, glycemic effectiveness, impact on weight, adverse event profile, and cost (84). In people with established atherosclerotic cardiovascular disease, heart failure, or CKD, treatment plans include agents that reduce cardiovascular and kidney disease risks independent of HbA1c (84). In type 1 diabetes, the ADA recommends early use of CGM and consideration of automated insulin delivery systems for all adults with type 1 diabetes (84).
The same standard of diabetes care should be applied to the dialysis population, and lack of population-specific evidence should not always limit consideration of newer diabetes interventions. Since people living with type 2 diabetes treated with dialysis are at high risk for cardiovascular disease, consideration of GLP-1 receptor agonists is important, with the understanding that titration schedules may need to be slowed to enhance gastrointestinal tolerability in people receiving dialysis and maximal tolerated doses may be lower than expected. GLP-1 receptor agonists may be particularly beneficial for individuals receiving dialysis with a history of atherosclerotic cardiovascular disease, those who have longer predicted survival (e.g., young or less frail), or those who require weight management, but the decision about whether to initiate GLP-1 receptor agonists must reflect individual priorities. Similarly, SGLT2 inhibitors can be considered in high-risk individuals (e.g., high risk for heart failure hospitalization). CGM and automated insulin delivery systems can be considered and offered to people treated with dialysis, as they have the potential to prevent glycemic crises and reduce the burden of capillary glucose testing and insulin injections. With shared decision-making, diabetes care treatment plans should integrate treatment approaches to diabetes and comorbid conditions while prioritizing individual preferences. Truly holistic care must also consider the social and economic factors that could prohibit people receiving dialysis from accessing these disease-modifying drugs. Newer medications are often expensive and difficult to obtain. Care delivery models will need to consider economic burden and accessibility. At the same time, prospective dialysis-specific RCTs should be conducted to inform how glucose-lowering and disease-modifying therapies can be applied to support individualized care for people receiving maintenance dialysis.
Further, the U.S. dialysis system offers a unique opportunity to implement an organized, centralized approach to delivering diabetes care to people receiving maintenance dialysis (85). The time burden of thrice-weekly dialysis therapy and dialysis-related symptoms such as postdialysis fatigue render some patients unable or unwilling to attend other medical appointments. As such, people receiving in-center hemodialysis often rely on dialysis clinicians for oversight of nonkidney conditions, including diabetes (85). The extent to which dialysis clinicians feel comfortable managing and individualizing diabetes care, particularly with contemporary agents and advanced monitoring approaches, is unknown. Similarly, neither the optimal clinician to assume the role of primary prescriber of glucose-lowering, disease-modifying medications nor the ideal way to deliver diabetes care in the dialysis setting has been established. Studies examining how diabetes care can be best implemented in the dialysis setting are needed. Decades-old studies suggest that performing aspects of diabetes care (e.g., foot checks and diabetes education) in dialysis clinics improves outcomes (86,87). However, other than routine foot checks, no aspects of diabetes care have been systematically implemented in U.S. dialysis care. Strategies that bring informed diabetes care to the dialysis clinic, such as telehealth provided by endocrinologists, may be particularly promising.
Knowledge Gaps and Research Needs
Recent pharmacologic and technological advances have revolutionized diabetes care in the general diabetes population and have potential to reduce the morbidity and mortality associated with diabetes among individuals receiving maintenance dialysis. However, gaps in population-specific data preclude the development of evidence-based diabetes management guidelines among people receiving maintenance dialysis. RCTs that investigate critical diabetes intervention effectiveness questions while also elucidating their tolerability and acceptability among people with dialysis-dependent kidney failure are critical to determining whether guideline bodies should recommend use of these agents in the dialysis population. Moreover, investigating different care delivery models to identify the optimal clinician to assume diabetes management and address barriers to access to medications and advanced glucose monitoring approaches (e.g., incorporation of pharmacy assistance programs and home delivery of prescriptions) is essential. Figure 2 summarizes research needs in diabetes care among individuals with dialysis-dependent kidney failure.
Research needs in diabetes care among individuals with dialysis-dependent kidney failure. FBG, fasting blood glucose.
Research needs in diabetes care among individuals with dialysis-dependent kidney failure. FBG, fasting blood glucose.
To fill these dialysis population–specific knowledge gaps, barriers to research in the dialysis setting must be overcome. Barriers include patient and clinician knowledge gaps, competing personal (e.g., high burden of comorbid conditions and transportation constraints) and professional (e.g., competing time demands) priorities, and research mistrust (88). Research inconvenience impedes participation of many people treated with dialysis. Cultivating a dialysis clinic culture where research and clinical care are synergistic is essential to research success in the dialysis environment (89). Input from dialysis stakeholders (patients, clinicians, dialysis clinic personnel, and dialysis delivery organization leaders) must be incorporated from clinical trial design to implementation to facilitate feasible and acceptable research. Communication between stakeholders is particularly important in the dialysis setting, not only to foster trust between patients, clinicians, and researchers but also to prevent communication gaps that obscure the realities of the challenges of delivering dialysis care and contribute untenable research procedures and poor adherence to protocols (89). There is no doubt that advancing research capacity in the dialysis setting faces unique challenges, but innovative strategies that foster interdisciplinary partnerships have the potential to break down these barriers and facilitate research that will improve outcomes in the dialysis population.
Conclusions
In summary, there is a compelling need to improve diabetes and cardiovascular outcomes among individuals living with dialysis-dependent kidney failure and diabetes. Major advances in diabetes interventions that improve outcomes in the nondialysis population offer optimism that rigorous population-specific studies could yield evidence that would support realization of similar outcome improvements in the dialysis population. Regardless, absence of dialysis-specific evidence should not detract from the importance of providing holistic, person-centered diabetes care that is tailored to integrate treatments for diabetes and comorbid medical conditions to people with dialysis-dependent kidney failure.
This article is featured in a podcast available at diabetesjournals.org/care/pages/diabetes_care_on_air.
Article Information
Acknowledgments. The authors thank Elizabeth Moreton, health sciences librarian at the University of North Carolina Health Sciences Library, for her assistance with the literature search.
Funding and Duality of Interest. K.R.K. is supported by the University of North Carolina Department of Medicine and School of Medicine Physician Scientist Training Program and the National Center for Advancing Translational Sciences, National Institutes of Health, through grant K12TR004416. K.R.K. has received personal compensation for consultation from Novo Nordisk. Additionally, K.R.K. has received research-related contracts (paid to the institution) from the National Center for Advancing Translational Sciences, Bayer, Boehringer-Ingelheim, Carmot, Diasome, Eli Lilly, Novo Nordisk, Rhythm Pharmaceuticals, and vTv Therapeutics. I.L. has received research funding (paid to the institution) from Novo Nordisk, Sanofi, and Boehringer-Ingelheim. I.L. has received research-related consulting fees (paid to the institution) from Novo Nordisk. I.L. has received advisory/consulting fees and/or other support from AbbVie, Altimmune, AstraZeneca, Bayer, Biomea, Boehringer-Ingelheim, Carmot, Cytoki Pharma, Eli Lilly, Intercept, Janssen/J&J, Mediflix, Merck, Metsera, Novo Nordisk, Pharmaventures, Pfizer, Regeneron, Sanofi, Shionogi, Structure Therapeutics, TARGET RWE, TERNS Pharma, The Comm Group, WebMD, and Zealand Pharma. K.R.T. receives research contracts (paid to the institution) from the National Institute of Diabetes and Digestive and Kidney Diseases, National Heart, Lung, and Blood Institute, National Institute on Minority Health and Health Disparities, National Center for Advancing Translational Sciences, National Institutes of Health Director’s Office, Centers for Disease Control and Prevention, Travere, and Bayer. K.R.T. has received advisory/consulting fees from Eli Lilly, Boehringer Ingelheim, AstraZeneca, Bayer, Novo Nordisk, ProKidney, and Pfizer. No other potential conflicts of interest relevant to this article were reported.
Handling Editors. The journal editors responsible for overseeing the review of the manuscript were Steven E. Kahn and Csaba P. Kovesdy.