The absence of osmotic diuresis modifies the effects of hyperglycemia on body fluids in patients with advanced renal failure. To determine the relationship between clinical manifestations and abnormalities in tonicity and extracellular volume in such patients, we analyzed 43 episodes of severe dialysis-associated hyperglycemia (serum glucose exceeding 600 mg/dL) treated only with insulin. The main manifestations were dyspnea in 22 cases (pulmonary edema in 19), nausea and vomiting in 15, coma in 13 and seizures in 3, while 5 patients had no symptoms. Treatment with insulin resulted in a decrease in serum glucose value from 913 ± 197 mg/dL to 170 ± 78 mg/dL, an increase in serum sodium level from 125 ± 5 to 136 ± 5 mmol/L, and a fall in calculated serum tonicity value from 300 ± 13 to 282 ± 11 mmol/kg (all at p < 0.001). The ratio of the change in serum sodium level over change in serum glucose concentration was –1.50 ± 0.22 mmol/L per 100 mg/dL. The percent increase in extracellular volume secondary to hyperglycemia developing from the prior euglycemic state and calculated from changes in serum sodium and chloride concentrations, was 10.9% ± 4.6% (1.5% ± 0.6% per 100 mg/dL increase in serum glucose level). All clinical manifestations dissipated after correction of hyperglycemia in 42 patients. One woman developed during treatment a fatal myocardial infarction. Dialysis patients with severe hyperglycemia may develop symptoms as a result of hypertonicity and extracellular expansion. Insulin alone may be sufficient treatment for these symptoms. The changes in serum tonicity and electrolytes during treatment are consistent with theoretical predictions.

We analyzed the changes in serum potassium concentration ([K]) and acid-base parameters in 43 episodes of dialysis-associated hyperglycemia (serum glucose level > 33.3 mmol/L), 22 of which were characterized as diabetic ketoacidosis (DKA) and the remaining 21 as nonketotic hyperglycemia (NKH). All episodes were treated with insulin therapy only. Age, gender, initial and final serum values of glucose, sodium, chloride, tonicity and osmolality did not differ between DKA and NKH. At presentation, serum values of [K] (DKA 6.2 ± 1.3 mmol/L; NKH 5.2 ± 1.5 mmol/L) and anion gap [AG] (DKA 27.2 ± 6.4 mEq/L; NKH 15.4 ± 3.5 mEq/L) were higher in DKA, whereas serum total carbon dioxide content [TCO2] (DKA 12.0 ± 4.6 mmol/L; NKH 22.5 ± 3.1 mmol/L), arterial blood pH (DKA 7.15 ± 0.09; NKH 7.43 ± 0.07) and arterial blood PaCO2 (DKA 26.2 ± 12.3 mm Hg; NKH 34.5 ± 6.7 mm Hg) were higher in NKH. At the end of insulin treatment, serum values of [K] (DKA 4.0 ± 0.7 mmol/L, NKH 4.0 ± 0.5 mmol/L), [AG] (DKA 16.3 ± 5.4 mEq/L, NKH 14.9 ± 3.0 mEq/L), [TCO2] (DKA 23.5 ± 5.0 mmol/L, NKH 24.1 ± 4.2 mmol/L), arterial blood pH (DKA 7.42 ± 0.09, NKH 7.51 ± 0.14) and arterial blood PaCO2 (DKA 31.8 ± 6.7 mm Hg, NKH 34.2 ± 8.3 mm Hg) did not differ between the two groups. Linear regression of the decrease in serum [K] value during treatment, (Δ[K]), on the presenting serum [K] concentration,([K]2), was: DKA, Δ[K] = 2.78 – 0.81 × [K]2, r = −0.85, p < 0.001; NKH, Δ[K] = 2.44 – 0.71 × [K]2, r = −0.90, p < 0.001. The slopes of the regressions were not significantly different. Stepwise logistic regression including both DKA and NKH cases identified the presenting serum [K] level and the change in serum [TCO2] value during treatment as the predictors of Δ[K] (R2 = 0.81). Hyperkalemia is a feature of severe hyperglycemia (DKA or NKH) occurring in patients on dialysis. Insulin administration brings about correction of DKA and return of serum [K] concentration to the normal range in the majority of the hyperglycemic episodes without the need for other measures. The initial serum [K] value and the change in serum [TCO2] level during treatment influence the decrease in serum [K] value during treatment of dialysis-associated hyperglycemia with insulin.