End-Stage Kidney Failure in the Diabetic Patient

Introduction: The Diabetic End-Stage Kidney Disease Population

The prevalence of diabetes continues to increase in both developed and developing countries. End-stage kidney disease (ESKD) is the most recognizable result of diabetic chronic kidney disease (CKD). Patients with diabetes mellitus are at increased risk of ESKD. Diabetes mellitus continues to be a worldwide epidemic, and with it, diabetic CKD remains the single most important cause of ESKD. According to both incidence and prevalence trends, although the incidence of diabetic ESKD has stabilized, the prevalence continues to increase. The cause of ESKD may be diabetes, or the patient may have diabetes as a comorbidity. While sentinel diabetes-related complications overall have declined over time, the rates of ESKD have decreased less than rates of other complications. Although data indicate that the relative risk for ESKD due to diabetes, compared to the risk in the nondiabetic population, has fallen by half, that relative risk remains over 6. The most recent U.S. Renal Data System (USRDS) data trends for the ESKD population are shown in Table 49.1 . (In total, diabetes is present in about 60% of all U.S. dialysis patients.) USRDS prevalence numbers for years 2015–2017 in the general population rose slightly with 35% attributed to diabetes; for incidence numbers in the same years, 46% were related to diabetes. Worldwide analysis of crude incidence rates indicates a broad range of rates due to diabetic ESKD compared to the general population ( Fig. 49.1 ). Although the lifetime risk of ESKD for patients with type 1 diabetes is 10%–15%, it appears to vary in study cohorts in different countries. The incidence of ESKD due to diabetes overall is higher in males (particularly in the white population) with increasing age, and significantly higher in Black-Americans, Hispanics, Asians, and Alaskan natives, compared to whites. In the United States, reductions by half in ESKD rates over 20 years in Native Americans, who have the highest rates of diagnosed diabetes, have been attributed to improvements in clinical care. Incidence rates appear to vary over fourfold among geographic areas in the United States. It is now understood that ESKD increases the morbidity of diabetes mellitus markedly; kidney disease accounts for most of the excessive cardiovascular and all-cause mortality risk associated with diabetes. Regarding dialytic management, those with ESKD attributed to diabetes are more likely to be treated with in-center hemodialysis (HD) and less likely to have a transplant. This chapter emphasizes these and other features of special importance to the adult ESKD population with diabetes mellitus.

Table 49.1

Recent Incidence and Prevalence (Numbers of Patients) of ESKD Attributed to Diabetes

From United States Renal Data Systems. 2019 USRDS Annual Data Report. Epidemiology of Kidney disease in the United States. National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD; 2019.

Open full size image

ESKD , End-stage kidney disease.

Predialysis Care/Dialysis Initiation

Patients with diabetic CKD may develop symptoms such as gastroparesis or neuropathy, which mimic those of uremia; may be prone to malnutrition, hyperkalemia, or volume overload; or may develop uremic symptoms earlier than those without diabetes. A recent study addressed the value (hospitalization, mortality risks) of primary care physician continuity before and during the transition from CKD to maintenance dialysis. High PCP continuity was associated with a reduction in hospitalizations for patients with diabetes. Conventional wisdom has been that ESKD patients with diabetes require earlier initiation of dialysis than nondiabetics, a “preventive” strategy to slow progression of some complications of diabetes and to improve outcomes. While the trend toward early dialysis starts in the general ESKD population reversed over a decade ago, diabetic patients have continued to be among patients in whom dialysis is initiated early. Diabetes strongly predicted the early start of dialysis according to recent USRDS data reports, although the difference in eGFR at initiation was slight. Relative to other outcomes data, information concerning mortality rates and factors that influence mortality soon after initiation in incident dialysis patients is limited. Mortality risk among HD patients is highest in the first 3 to 4 months after dialysis initiation. Data on risk/benefit of early initiation in patients with diabetes, however, show mixed results. A systematic review in 2016 indicated that the effect of timing of initiation of dialysis did not differ between patients with and without diabetes.

Dysglycemia

Patients with either type 1 or type 2 diabetes and CKD/ESKD require more comprehensive laboratory monitoring, such as glucose levels and hemoglobin A _1c (HbA _1c ) testing. The maintenance of glucose homeostasis is affected both by diabetes and kidney failure. The core pathologic defects in diabetes are decreased insulin secretion (type 1 diabetes) and insulin resistance, impaired insulin secretion, increased hepatic glucose production, decreased glucose uptake by muscle, and increased renal glucose reabsorption (type 2 diabetes). When kidney impairment progresses additional alterations occur, which may further cause dysglycemia; insulin secretion by the pancreas is reduced, and insulin clearance is diminished. With a molecular weight of about 6000 daltons, insulin has a high renal clearance through filtration and secretion. Renal insulin clearance is reduced as GFR falls to 15–20 mL/min/1.73 m ² . Abnormal glucose homeostasis is further affected in uremic individuals by peripheral insulin resistance, involving defective glucose uptake and reduced muscle protein anabolism (but not cellular potassium uptake). Glucose uptake mediated by specific transporter proteins is a major action of insulin. Insulin resistance in CKD is a unique metabolic abnormality induced by factors such as uremic toxins, chronic inflammation, excess visceral fat, oxidative stress, metabolic acidosis, and deficiency of vitamin D and erythropoietin. However, the health implications of insulin resistance have not been adequately evaluated in the ESKD population. Improvement in insulin resistance with dialysis treatment is due at least in part to removal of uremic toxins. To what extent insulin resistance also contributes to protein-energy wasting through the ubiquitin-proteosome pathway to atherosclerosis and cardiovascular disease remains undetermined in ESKD patients.

The net clinical effect of these various defects in a given diabetic ESKD patient is variable; insulin resistance in advancing CKD may be offset by insulin retention and poor caloric intake; improvement in insulin resistance by dialysis may be offset by more robust caloric intake as uremia improves. A significant reduction in insulin requirements during and following an HD treatment may be anticipated. While clinical assessment of insulin resistance in diabetic ESKD is not practical, it remains a potential therapeutic target. However, insulin sensitizers, a potential treatment aimed at insulin resistance, are either sparingly used or contraindicated in ESKD.

Prediabetes

In recent years, there has been a marked increase in the number of U.S. adults with prediabetes to at least 35% of the general population. Prediabetes is conventionally identified by an HbA _1c value of 5.7%–6.4%. The risk of overt diabetes developing is not linear but increases with higher HbA _1c levels below the diabetic range. Prediabetes is associated with increased risk of all-cause and cardiovascular mortality in the general population. ESKD patients who test in the prediabetic range should have annual monitoring for the development of diabetes, lifestyle interventions, and identification/treatment of other cardiovascular risk factors, such as hypertension, hyperlipidemia, and obesity. Limited data suggest that having ESKD may be associated with a lower risk of the development of incident diabetes, although the confounding effects related to ESKD and lower HbA _1c levels described later need to be taken into account.

Assessing Glycemic Control in End-Stage Kidney Disease

Providing ESKD management for patients with diabetes mellitus requires awareness of how glycemic control is being assessed. The two conventional techniques available for assessing the effectiveness of glycemic management are self-monitoring of blood glucose and HbA _1c . Proposed glycemic control parameters for diabetic ESKD patients are fasting blood glucose of 100–140mg/dL and 2-hour postprandial blood glucose of < 200mg%. In the general diabetes population, landmark studies have established HbA _1c as the primary predictor of diabetic complications. Among ESKD patients, however, the limitations of HbA _1c have more recently become well recognized. HbA _1c reflects average blood glucose concentrations over approximately the three preceding months, has some correlation with blood glucose levels, and has provided evidence in interventional trials of the benefit of tight glycemic control. It does not inform about glycemic variability or hypoglycemia. However, reliance on HbA _1c in a given individual is limited by the fact that HbA _1c values may be widespread when plotted against average glucose levels. Current society guidelines continue to recommend an HbA _1c target of < 7.0% in the general population, with the caveat that goals must be individualized based on factors such as duration of diabetes, age and life expectancy, and comorbid conditions.

However, it should be noted that extensive trials have forced a reevaluation of the validity of the HbA _1c test in the setting of ESKD, as well as indicating only a weak correlation between the test and outcomes in this population. The major source of uncertainty regarding the role of HbA _1c -based glycemic control in ESKD at this time is the unreliability of the HbA _1c test itself. Compared to the general diabetic population, correlation with average glucose levels is weaker. Data indicate that HbA _1c measurements may be misleadingly low in patients with ESKD, resulting in an underestimation of the degree of hyperglycemia. The most likely sources of discordance from other glycemic tests in kidney patients are anemia, the use of erythrocyte stimulating agents (ESAs), and the administration of intravenous iron. ESAs stimulate the production and release of immature erythrocytes, whose hemoglobin appears to have reduced glycosylation. Although comparative data are limited, HbA _1c values in the setting of ESKD, as a rough estimate, generally measure about 1% lower than expected for given levels of serum glucose during the time preceding the HbA _1c test.

In recent years, increasing consideration has been given to glycated albumin (GA) as a replacement for HbA _1c for assessing glycemic control in ESKD patients. GA reflects glucose exposure for only the previous 2 weeks and requires more frequent testing. Unlike with HbA _1c , there is no international standardization of the GA measurement. However, GA results are not confounded by the presence of anemia or the use of ESAs. Limited data suggest good correlation with values obtained by continuous glucose monitoring (CGM).

Achieving Glycemic Control in End-Stage Kidney Disease

Treatment goals for patients with advanced CKD, at least for those with related limited life expectancy and at risk for hypoglycemia, have been extended in recent years to above 7%. However, there have been no randomized, interventional clinical trials to evaluate the beneficial effects of glycemic control in patients with diabetes and ESKD. Two frequently cited, comparably sized retrospective observational studies with methodological differences have reported on the relationship of glycemic control and outcomes in ESKD patients with diabetes, yielding somewhat conflicting conclusions. Kalantar-Zadeh et al. demonstrated, after adjustment for multiple potential confounders, an incremental relationship between higher HbA _1c values and higher mortality. A follow-up study by the same investigators indicated that poor glycemic control (HbA _1c ≥ 8.0%) was associated with high all-cause and cardiovascular deaths. Williams et al. were able to establish only a weak and variable relation, and only when follow-up was extended to a maximum of 3 years. Findings from the Dialysis Outcomes and Practice Patterns Study (DOPPS) indicated that mortality increased as HbA _1c exceeded the 7%–7.9% range. A meta-analysis of nine observational studies and one secondary analysis indicated that a higher mortality risk did not occur until the HbA _1c levels were ≥ 8.5%. Patients with very low HbA _1c levels are also at risk ( Table 49.2 ). In a post hoc analysis of the German Diabetes and Dialysis Study (4D Study), each 1% increase in baseline HbA _1c was associated with an 18% higher risk of sudden cardiac death, although myocardial infarction events were not affected. Overall, these data suggest that HbA _1c ranges of 7%–8% may be optimal. The Kidney Disease Outcome Quality Initiative (KDOQI) Clinical Practice Guidelines based on observational studies suggest a higher ESKD mortality risk with HbA _1c levels < 6.5 and > 8.0%. However, a report on DOPPS participants from Japan demonstrated that optimal HbA _1c levels may vary among countries; the association of HbA _1c levels and mortality was U shaped, but the HbA _1c level associated with the lowest mortality in diabetic HD patients in Japan (6.0%–7.0%) differed from that of the United States.

Table 49.2

Glycated Hemoglobin Levels and Mortality Risk (Adjusted) for Prevalent Patients With Diabetes Treated With Hemodialysis

From Hill CJ, Maxwell AP, Cardwell CR, et al. Glycated hemoglobin and risk of death in diabetic patients treated with hemodialysis: a meta-analysis. Am J Kidney Dis. 2013;63:84–94.

Open full size image

Adjusted for age, sex, diabetes type, dialysis vintage, and hemoglobin concentration when available.

HR , Hazard ratio.

Although not as well studied, GA has been proposed as a superior alternative to HbA _1c for determining the value of glycemic control in ESKD patients with diabetes. Clinical practice guidelines for diabetes do not recommend GA as the first-line metric of glycemic control. Its linkage to average glucose levels and prognostic significance is less well established in ESKD. GA has been shown to accurately predict the risk of hospitalization and death in patients with diabetes on HD and, with less evidence, in those on peritoneal dialysis (PD). In a recent analysis of samples from the German 4-D study, a high risk of mortality in the highest quartile of baseline GA was demonstrated. Comparative HbA _1c and GA data from a recent 1-year mortality outcomes study from Japan are shown in Fig. 49.2 .

Guideline recommendations indicate that the intensity of glycemic control must be individualized based on potential risks and benefits. An HbA _1c target using the HbA1c metric of closer to 7% for younger, incident ESKD patients is recommended, and a target closer to 8% for longstanding diabetics, older patients, longer dialysis vintage, or those with multiple comorbid conditions. For real-time assessment, CGM profiling is increasingly valued, particularly given the potential effects of the dialysis treatment on short-term glycemia. Increased glycemic variability, a possible cardiac risk factor, has been demonstrated on dialysis days and reduced insulin requirements the day after dialysis. In a recent report, closed-loop insulin delivery provided safe and effective diabetes control for in-patients on HD. Patients managed with a subcutaneous insulin pump coupled with a continuous glucose monitor by means of a computer algorithm spent almost 40% more time in the glucose target range than the control group.

Diabetic Ketoacidosis/Hyperosmolar Coma

Potentially lethal hyperglycemic crises such as diabetic ketoacidosis (DKA) and hyperosmolar coma occur infrequently in ESKD patients with diabetes but may be severe and prolonged when they do occur. DKA is a complex disorder that affects volume status, electrolytes, and acid-base balance. It usually affects patients with type 1 diabetes and rarely those with new-onset type 2 diabetes. The underlying pathophysiology is that of absolute or relative insulin deficiency, insulin resistance, and increased counter-regulatory hormones. Common precipitating causes are omission of insulin doses, myocardial ischemia or infarction, and infection. Making the appropriate diagnosis in the ESKD patient is based on hyperglycemia, positive serum ketones, metabolic acidemia, and an increased anion gap. However, one caveat is that the diagnosis may be delayed insofar as the degree of hyperglycemia may be mild, the plasma reaction for ketones may be negative, and the anion gap may not be affected. The diagnosis may also be delayed because the clinical presentation itself of severe hyperglycemia and ketoacidosis is atypical in patients with ESKD. Because osmotic diuresis does not occur, volume contraction is avoided (unless vomiting, diarrhea, and excessive insensible loss simultaneously occur). Patients may present with weight gain, thirst, pulmonary congestion, and in extreme circumstances, coma. However, because the rise in serum osmolality and sodium levels are not marked, there may be few neurologic symptoms. The total body potassium deficit, typically significant in non-ESKD DKA, is limited, and hyperkalemia may result. Because excessive urinary phosphate excretion does not occur, hypophosphatemia is not expected. Magnesium deficiency is absent. The bicarbonate deficit may be variable.

It follows that the acute management of DKA/hyperosmolar coma should be modified in ESKD patients. Insulin, frequently in low doses, is the only treatment required in many cases. Hypoglycemia should be avoided. (Current guidelines for treatment of DKA, in general, recommend against the use of bolus doses of insulin. Aggressive insulin treatment may result in severe and prolonged hypoglycemia.) Administration of large volumes of crystalloid is inappropriate, as patients have not had characteristic losses and are already at risk for volume overload. Intravenous crystalloid requirement may be increased if gastrointestinal (GI) losses have occurred. The use of parenteral bicarbonate is not indicated unless the level is critically low, particularly as it may exacerbate volume overload. No phosphate replacement is generally needed. Typical indications for urgent dialysis include pulmonary edema and hyperkalemia. While the patient is being stabilized, a cause for the episode of DKA/hyperosmolar coma should be urgently pursued.

Hypoglycemia

In recent years, there has been renewed concern regarding hypoglycemia as a determinant of adverse clinical outcomes in ESKD. Patients with ESKD, regardless of diabetes status, are prone to hypoglycemia, sometimes spontaneous, and sometimes related to hyperglycemia treatment. Likewise, among patients with hypoglycemia, ESKD is a common comorbid condition. Patients with both diabetes and ESKD are at particularly high risk of hypoglycemia, which adds to the case for avoiding intense glycemic control in this population. A recent study used International Classification of Diseases, Ninth Revision (ICD-9) codes to identify admissions for hypoglycemia as the primary admitting diagnosis, with and without ESKD, in a nationwide U.S. cohort. Almost 12% had ESKD, with higher rates in Black-Americans and Hispanics. Hypoglycemic patients with ESKD had higher hospitalization costs, longer lengths of hospital stay, and a threefold higher mortality.

The cause of hypoglycemia in ESKD patients is frequently multifactorial. Declining kidney function leads to reduced renal insulin clearance, and degradation of insulin in peripheral tissues is also decreased. In addition, defenses against hypoglycemia are not intact; renal gluconeogenesis, a protective source of glucose production from precursor molecules during starvation, is impaired. Patients with ESKD may also have diminished glycogen stores due to suboptimal nutrition.

As antidiabetic agents are likely contributors, evaluation should include assessment of the glycemic regimen and nutritional status leading up to the hypoglycemic event. In addition, causes of hypoglycemia found in the absence of diabetes or ESKD, such as adrenal insufficiency, infection/sepsis, alcohol abuse, organ failure, and insulinoma, may need to be considered. (One caveat is that fasting insulin levels may be higher in the presence of ESKD, and serum cortisol levels may be variable.) Although dextrose contained in the dialysate solution may work to offset falling glucose levels, acute hypoglycemia occurring during or immediately after the dialysis session may still require treatment with oral glucose or intravenous dextrose. If available, glucose 15–20 g should be administered; in more severe episodes, intravenous dextrose 50%, 10–50 mL. Reduction in insulin, especially after dialysis, or conversion to an oral agent with lower hypoglycemic risk, such as a thiazolidinedione or a dipeptidyl peptidase-4 inhibitor, may be necessary for consultation with the diabetologist. Glucagon kits for emergency use should be prescribed. Long-acting insulin may pose a prolonged risk. Nonselective beta-blockers may mask hypoglycemia symptoms and delay recovery from an episode.