More than 40% of all new patients starting dialysis in the United States are diabetic. Provision of maintenance dialysis for this group can be a challenging task. Morbidity and mortality are substantially higher in diabetic patients maintained on dialysis than in their nondiabetic counterparts, with cardiovascular disease and infection being the leading causes of death. In the United States, the 3-year survival rate of diabetic patients maintained on dialysis is only about 50% (USRDS, 2013).

I. WHEN TO INITIATE DIALYSIS. Early referral to nephrologists of diabetic patients with renal failure reportedly improves outcomes. Previous guidelines emphasized initiation of dialysis prior to the appearance of frank uremic manifestations (at an estimated glomerular filtration rate [eGFR] of ≤Ü15 mL/min per 1.73 m²). However, a recent randomized controlled trial that examined mortality versus time of dialysis initiation, the IDEAL study, found no difference in survival between early or late initiation of dialysis, and in that study, approximately one-third of the participants were diabetic (Cooper, 2010).

II. HEMODIALYSIS VERSUS PERITONEAL DIALYSIS. The potential problems with each form of dialysis are listed in Table 32.1. Long-term peritoneal dialysis (PD) in diabetic patients may complicate control of blood sugar because altered glucose homeostasis is stressed further by the large amount of glucose administered via the dialysis solution. In addition, glucose absorption from the abdominal cavity decreases appetite. Many PD patients have difficulty ingesting the higher recommended amount of protein for PD patients (1.2 g/kg daily). On the other hand, the incidence and severity of hypoglycemic episodes is reduced in continuous ambulatory peritoneal dialysis (CAPD) or automated peritoneal dialysis (APD) compared with hemodialysis (HD) because of the constant or near-constant presence of glucose in the abdomen. Rates of infection (peritonitis, exit-site and tunnel infections) and rates of catheter replacement are similar between diabetic and nondiabetic patients on PD. Administration of insulin intraperitoneally appears to increase slightly the risk of peritonitis in PD
patients and, though appealing on physiologic grounds, is now less commonly employed. With HD, coexisting blood vessel disease often hinders creation of an adequate, long-lasting vascular access. The survival rates of both arteriovenous (AV) fistulas and grafts are substantially reduced in diabetic compared with nondiabetic patients. A small fraction of diabetic patients develop severe hand ischemia after creation of an ipsilateral AV fistula, which can lead to gangrene and need for amputation; prompt ligation of the fistula is indicated in such instances. Because of autonomic nervous system dysfunction or cardiac diastolic dysfunction, diabetic patients are at increased risk for hypotension during HD. Problems with vascular access and increased risk of hypotension may cause diabetic patients to receive a lesser amount of dialysis (in terms of fractional urea clearance [Kt/V]) than their nondiabetic counterparts.

Lower extremity amputations are frequent in diabetic patients on either HD or PD. The rate of progression of retinopathy appears to be similar between patients treated with HD and PD. Although visual impairment impedes training for CAPD and makes it difficult for the patient to perform the exchange procedure properly, even blind diabetic patients can be trained to perform CAPD without a helper. When properly instructed, their risk of developing peritonitis is only slightly greater than the risk in sighted diabetics. A number of devices are available to help visually impaired patients connect the dialysis solution container to the peritoneal transfer set (see Chapter 22). APD is a better therapeutic choice for blind diabetic patients because many APD schedules require the performance of only one “on” and one “off” procedure daily.

Earlier reports from the U.S. Renal Data System suggested that mortality is higher in diabetic patients, and especially diabetic women, on PD than in those on HD. Patient selection biases and/or inadequate PD may have affected these observations. In a subsequent large analysis, mortality risk was actually higher with HD than PD among younger diabetics with no comorbidity but lower with HD in older diabetics, especially if they had comorbidities (Vonesh, 2004). These results are undoubtedly also affected by selection biases. Comorbidity and malnutrition have much larger effects on mortality than the dialysis modality. Meticulous management and prevention of cardiovascular and infectious morbidity may lead to substantial improvement in patient survival.

III. DIET. Whatever the mode of dialysis therapy, diabetic patients generally show evidence of wasting and malnutrition. Many factors contribute, including chronic inflammation, inadequate food intake, diabetic gastroparesis and enteropathy, and the catabolic stress associated with frequent intercurrent illness. In the event of serious illness, diabetic dialysis patients often require early and intensive nutritional support.

A. Routine dietary prescription. The diets advocated for nondiabetic HD and PD patients in Chapter 31 also apply to patients with diabetes. In an anuric diabetic patient being treated with HD, the stringent sodium, potassium, and fluid restrictions described in Chapter 31 should be applied. Special effort should be made to limit intake of simple sugars and saturated fats.

1. Carbohydrates percentage. The general recommendation for a diabetic diet is for 50%-60% of intake to be carbohydrates, with some interest in using an even lower carbohydrate diet in diabetic patients (Arora, 2005). In patients undergoing PD, the glucose calories supplied from the PD regimen (usually around 400 kcal) should be subtracted from the dietary carbohydrate prescription, and perhaps in selected patients with hypertriglyceridemia, a focus on avoiding all high-glycemic-index carbohydrates might be of benefit.

2. Dietary “glycotoxins” from advanced glycosylation end products (AGEs). Levels of AGEs are increased in food that has been cooked at high temperature, especially if food contains a high proportion of fat. Dietary AGE ingestion has been linked to adverse lipid profiles and inflammatory markers in diabetic patients (Uribarri, 2005) and to increased serum levels of AGEs in end-stage kidney disease (ESKD) patients, with perhaps increased risk of access thrombosis. Any reason to additionally restrict food in ESKD patients should be done with caution, given the high prevalence of malnutrition; however, some attention to food preparation with a focus on minimizing formation of AGEs (avoidance of deep frying, extensive heating) might be a consideration.

B. Diabetic gastroparesis and enteropathy. The diagnosis of diabetic gastroparesis is often made on the basis of symptoms of nausea, vomiting, early satiety, and postprandial fullness. Since other treatable conditions can have similar symptoms, an esophagogastroduodenoscopy should be performed before symptoms are ascribed to gastroparesis alone. The traditional “gold standard” to establish the diagnosis of gastroparesis is scintigraphic measurement of gastric emptying. However, a drawback is that scintigraphy exposes patients to radiation and is therefore not ideally suited for repeated investigations (to monitor the response to therapy). This problem can be overcome by 13C-labeled acetate and octanoic acid breath tests. Diabetic gastroparesis can be associated with poor food intake and unpredictable nutrient absorption; the result can be hypoglycemia alternating with hyperglycemia.

In such patients, small, frequent (up to six times per day) feedings may improve symptoms. The pharmacologic treatment of gastroparesis in diabetic persons on dialysis is unsatisfactory. Metoclopramide given in a small starting dose (5 mg before meals) with small increments until results are seen is usually the first drug prescribed. This drug is associated with a high incidence of extrapyramidal complications in dialysis patients, particularly at higher doses, and its effects are often temporary. Other “prokinetic” gastrointestinal motility drugs, such as domperidone, motilin agonists, or ondansetron, may be tried.

Diabetic enteropathy results from functional impairment of the enteric nervous system and can result in disordered motility of the small bowel and the colon, resulting in either prolonged or shortened bowel transit times. Diabetic enteropathy with resulting diarrhea can complicate alimentation, causing debilitation, poor food intake, and hypoglycemia. Severe cases of diabetic enteropathy can be treated with a trial of broadspectrum antimicrobials (e.g., doxycycline in a dose of 50 or 100 mg daily) to combat bacterial overgrowth in the intestine. Loperamide hydrochloride (up to 10 mg daily) to decrease bowel motility is also useful.

IV. CONTROL OF BLOOD SUGAR

A. Alteration of insulin metabolism by CKD. In uremic patients (both diabetic and nondiabetic), insulin secretion by the β cells of the pancreas is reduced, and the responsiveness of peripheral tissues (e.g., muscle) to insulin is depressed; that is, there is increased insulin resistance. Insulin resistance occurs in almost all uremic patients and results in hyperglycemia. The literature suggests that hepatic glucose production and uptake are normal in uremia and that skeletal muscle is the primary site of insulin resistance, probably via a postreceptor defect (Castellino, 1992). However, many of the actions of insulin are maintained in renal failure, including potassium uptake by the cells and inhibition of proteolysis.

The kidney is pivotal in insulin metabolism in healthy individuals. Insulin is freely filtered by the glomerulus, with 60% cleared via glomerular filtration and 40% by extraction from the peritubular vessels; less than 1% of filtered insulin is excreted in the urine unchanged. Approximately 6-8 units of insulin are degraded by the kidney each day, roughly 25% of the daily production of insulin by the pancreas. Renal metabolism is enhanced in diabetic subjects receiving exogenous insulin, since injected insulin bypasses the liver and goes directly into the systemic circulation. The rate of insulin catabolism is decreased owing to decreased renal mass, and therefore the half-life of any insulin present in the circulation is prolonged. The reduction in insulin clearance is also mediated by a decline in hepatic metabolism. All of these abnormalities are only partially corrected after institution of maintenance dialysis therapy.

1. Abnormal glucose tolerance tests in all dialysis patients. The glucose tolerance test cannot be used to diagnose diabetes in dialysis patients because the rise in serum glucose concentration will be greater and more prolonged than normal in all dialysis patients as a result of uremia-induced insulin resistance. However, fasting serum glucose concentrations are normal in nondiabetic HD patients; a high level suggests the presence of diabetes. In PD patients, a true fasting state is never achieved owing to constant absorption of glucose from the dialysis solution. In this group, unless peritonitis is present, the “fasting” serum glucose value rarely exceeds 160 mg/dL (8.9 mmol/L), even when using 4.25% dextrose dialysis solution; higher levels suggest that the patient has diabetes. In CAPD patients using icodextrin, serum glucose values may be spuriously overestimated by auto-analyzers that use the glucose dehydrogenase method of sample analysis (Tsai, 2010).

2. Increased sensitivity to insulin. In diabetic dialysis patients being treated with exogenous insulin, the importance of reduced insulin catabolism overrides the impact of insulin resistance; when exogenous insulin is administered, its effect may be intensified and prolonged. Thus,
smaller-than-usual doses should be given. Bolus administration of moderately large intravenous doses (e.g., 15 units of regular insulin), even when ketosis is present, can result in severe hypoglycemia. Hypoglycemia can occur also after administration of the longer-acting insulins, such as isophane insulin (NPH) and insulin glargine.

3. Hyperglycemia. The clinical presentation of hyperglycemia is modified when renal function is absent. The absence of the “safety valve” effect of glycosuria may result in the development of severe hyperglycemia (serum glucose level >1,000 mg/dL [56 mmol/L]). Severe hyperosmolality with accompanying alteration of mental status is unusual because of the absence of water loss induced by osmotic diuresis. Indeed, even extreme hyperglycemia is often asymptomatic in dialysis patients (Al-Kudsi, 1982). However, manifestations can include thirst, weight gain, and, on occasion, pulmonary edema or coma (Tzamaloukas, 2004). Diabetic ketoacidosis, frequently accompanied by severe hyperkalemia and coma, can develop in insulin-dependent dialysis patients. Management of hyperglycemia with or without ketoacidosis differs from that in patients without renal failure in that administration of large amounts of fluid is unnecessary and generally contraindicated. All of the clinical and laboratory abnormalities of hyperglycemia are corrected by insulin administration, which often is the only treatment needed. To manage severe hyperglycemia, one can administer a continuous infusion of low-dose regular insulin (starting at 2 units/hr) with close clinical monitoring and measurement of serum glucose and potassium concentrations at 2- to 3-hour intervals. If severe hyperkalemia is present, electrocardiography should be done. Emergency dialysis may be needed in patients with hyperglycemia and either severe pulmonary edema or lifethreatening hyperkalemia.

4. Hypoglycemia. Avoidance of hypoglycemia is often the ratelimiting step in obtaining good glucose control. Treatment for hypoglycemia can lead to rebound hyperglycemia and erratic glycemic control. There are many factors that contribute to hypoglycemia, including decreased caloric intake due to attendant anorexia, decreased insulin clearance, reduced renal gluconeogenesis due to the reduction in functioning renal mass, impaired release of the counterregulatory hormone epinephrine due to the autonomic neuropathy of renal failure, decreased hepatic metabolism of insulin, and decreased metabolism of drugs that might promote a reduction in the plasma glucose concentration such as alcohol, propranolol and other nonselective adrenergic blockers. Additionally, hypoglycemic unawareness and gastroparesis may increase the risk for hypoglycemia. In diabetic patients, HD solution should always contain about 90 mg/dL (5 mM) glucose; if glucose is not added,
severe hypoglycemia during or soon after the HD session can result (Burmeister, 2012). Higher (200 mg/dL, 11 mM) dialysate glucose levels may increase the frequency of hyperglycemia (Raimann, 2012), and may not protect against hypoglycemic episodes better than the 90 mg/dL option.

B. Insulin therapy. Achieving and maintaining reasonable glucose control while avoiding hypoglycemia is the challenge in managing patients with diabetes on dialysis. Reasonable glycemic control in diabetic patients on chronic dialysis is considered a fasting blood glucose below 140 mg/dL and a 1-hour postprandial value of less than 200 mg/dL with an HbA1c between 7% and 8%. Several large studies have found no significant correlation between glycemic control and survival, yet a much higher risk of hypoglycemia with stricter glycemic control (Williams, 2010). The HbA1c target that is associated with the best outcome in dialysis patients has not been established (KDOQI clinical practice guidelines, 2005). It has been suggested that measurement of glycated albumin more accurately assesses glycemic control in this population as it is not affected by the Hb level, but this test is not readily available. Another issue with HbA1c is that it can be affected (reduced) by administration of ESAs and iron (Ng, 2010).

Irreversibly and slowly formed compounds that are the result of nonenzymatic glycosylation of proteins, the so-called AGEs, alter the structure and function of vascular basement membranes, stimulate the production of growth factors, and alter the function of intracellular proteins. In PD patients, AGE deposition in the peritoneal membrane is associated with an increase in permeability and excessive protein losses in the dialysate (Nakamoto, 2002).

1. Insulin regimens. The following dose recommendations have been made for insulin dosage in the setting of kidney disease (Snyder, 2004):

a. No dose adjustment is required if the GFR is above 50 mL/min.

b. Reduce the insulin dose by 25% when the GFR is 10-50 mL/min.

c. Reduce the dose by 50% when the GFR is less than 10 mL/min.

2. Example using glargine and rapid-acting insulin. As an example, a common weight-based dose might be 0.6 units/kg total daily dose of insulin (Murphy, 2009). Reducing this by 50% for ESKD changes this to 0.3 units/kg total daily dose of insulin (Baldwin, 2012). Of this, half should be given as basal insulin and half as mealtime bolus insulin. This would mean that 0.15 units/kg would be given in the morning as a basal dose, and then the remainder (0.15 units/kg) divided up by the number of meals, given as rapid-acting insulin; say 0.05 units/kg at breakfast, lunch, and supper. For a 70-kg patient, the total dose of insulin would be 70 kg × 0.3 units/kg = 21 units. Half of this, or approximately 10 units,
would be given as the basal once-daily glargine, and the remaining 11 units would be given over the 3 daily meals, or 3-4 units rapid-acting insulin per meal.

3. Example using NPH plus a rapid-acting insulin. When using NPH with rapid-acting insulin, the total daily dose would be the same (21 units), but two-thirds of the daily dose (or 14 units) should be given as NPH, with two-thirds of the NPH (9 units) given at breakfast and the remaining 5 units of NPH at bedtime. The non-NPH remainder of the total daily dose (7 units) would then be given as rapid-acting insulin, giving 3 units at breakfast and 4 units at dinner. A lunchtime dose of rapid-acting insulin is not needed, as the morning dose of NPH is peaking at this time and would cover this meal.

4. Other insulin combinations. New basal and rapid-acting insulin analogs are expected to be available for clinical use in 2015, but have not been studied in patients with ESKD (Danne, 2011).

5. Timing of mealtime insulin. While mealtime insulin is generally given about 5 minutes before the meal, some patients prefer to take it just after the meal. This is somewhat safer in that it allows patients to reduce the insulin dose if the entire meal is not eaten. For example, if only 50% of the meal is eaten, only 50% of the dose is taken. Finer enhancements of the mealtime insulin dose include modification by carbohydrate counting (if the patient is willing to learn and implement this technique) and by addition to or subtraction from the mealtime insulin dose using a correction factor scale (if the patient is willing to check his or her glucose values before each meal).

6. Glucose monitoring. It is important that blood glucose levels be monitored closely and that individually appropriate dose adjustments in insulin therapy be made. Patients receiving insulin therapy at home should monitor their glucose levels at least twice daily, in the morning and at bedtime. With both of the above regimens, a reasonable “correction dose” of insulin would be 1 extra unit of daily insulin for every 50 mg/dL glucose above target (for example the target level might be 150 mg/dL) for that patient.

7. Effect of hemodialysis on insulin dose. HD has been shown to improve both tissue sensitivity to insulin and insulin secretory response to glucose (DeFronzo, 1978). The mechanism by which this occurs is unknown, but improved acid-base status may be contributing. When HD is initiated, the insulin requirement in any given patient may change, depending on the net balance between improved tissue sensitivity and improved hepatic insulin metabolism. One cannot readily predict insulin requirements in this setting, and careful observation of the patient is essential.

In HD patients, several different insulin regimens can be used to achieve glycemic control. The glargine- and NPHbased regimens described above can be used as a starting point. Some experts feel that long-acting insulin preparations
should be avoided, while others feel that such agents should be used, but there are no head-to-head comparisons of different regimens in dialysis patients.

With regard to insulin therapy on dialysis versus nondialysis days, the usual baseline dose is normally given, but the timing of mealtime doses is often changed when dialysis alters the times when food is consumed.

8. Effect of peritoneal dialysis on insulin dose. The glucose contained in peritoneal dialysate increases the need for blood glucose-lowering therapy, and more insulin is often required because of insulin resistance and the glucose load absorbed from the hypertonic dialysate. A 1.5-percent dextrose (glucose monohydrate, MW 198) dialysate solution, for example, has a glucose (MW 180) concentration of 1,500 × (180/198) = 1,364 mg (76 mmol/L), well above that in the plasma. On the other hand, in some PD patients less insulin can be required than anticipated owing to decreased carbohydrate intake and to prolongation of the duration of action of insulin resulting from reduced renal and hepatic insulin clearance.

To help maintain near-normal glycemia during PD, blood sugar in patients treated with CAPD or APD can be controlled with intraperitoneal insulin, although this is now done only rarely. Use of the intraperitoneal route has the advantages of a continuous or nearly continuous presence of insulin, elimination of the need for injections, and a more physiologic route of insulin supply to the liver via the portal vein, mimicking the way that pancreatic insulin reaches the liver (Tzamaloukas, 1991). Disadvantages are the potential for bacterial contamination of dialysate during injection of insulin into the bags, need for somewhat higher daily total insulin doses due to losses of insulin with the spent dialysate, and, perhaps of greatest concern, the risk of peritoneal fibroblastic proliferation and hepatic subcapsular steatosis (Maxwell, 1991). If intraperitoneal insulin is used, it is recommended that a long, 3.8-cm (1.5 in) needle be used to ensure that the full dose of insulin is injected into the dialysis solution container rather than being trapped in the infusion port; the dialysis solution container should be inverted several times after injection to ensure proper mixing. For intraperitoneal insulin protocols, please refer to previous editions of this Handbook.

It is important to remember that icodextrin and maltose, which are contained in some PD solutions, can interfere with, or cause falsely elevated glucose results with, some self-monitoring methods, possibly leading to inappropriate therapy (Tsai, 2010; Firanek, 2013).

9. Use of an insulin infusion pump.

Stay updated, free articles. Join our Telegram channel

Join