Hyperosmolar hyperglycemic nonketotic syndrome. Report of 22 cases and brief review

The Corrected Serum Sodium Concentration in Hyperglycemic Crises: Computation and Clinical Applications

In hyperglycemia, hypertonicity results from solute (glucose) gain and loss of water in excess of sodium plus potassium through osmotic diuresis. Patients with stage 5 chronic kidney disease (CKD) and hyperglycemia have minimal or no osmotic diuresis; patients with preserved renal function and diabetic ketoacidosis (DKA) or hyperosmolar hyperglycemic state (HHS) have often large osmotic diuresis. Hypertonicity from glucose gain is reversed with normalization of serum glucose ([Glu]); hypertonicity due to osmotic diuresis requires infusion of hypotonic solutions. Prediction of the serum sodium after [Glu] normalization (the corrected [Na]) estimates the part of hypertonicity caused by osmotic diuresis. Theoretical methods calculating the corrected [Na] and clinical reports allowing its calculation were reviewed. Corrected [Na] was computed separately in reports of DKA, HHS and hyperglycemia in CKD stage 5. The theoretical prediction of [Na] increase by 1.6 mmol/L per 5.6 mmol/L decrease in [Glu] in most clinical settings, except in extreme hyperglycemia or profound hypervolemia, was supported by studies of hyperglycemia in CKD stage 5 treated only with insulin. Mean corrected [Na] was 139.0 mmol/L in 772 hyperglycemic episodes in CKD stage 5 patients. In patients with preserved renal function, mean corrected [Na] was within the eunatremic range (141.1 mmol/L) in 7,812 DKA cases, and in the range of severe hypernatremia (160.8 mmol/L) in 755 cases of HHS. However, in DKA corrected [Na] was in the hypernatremic range in several reports and rose during treatment with adverse neurological consequences in other reports. The corrected [Na], computed as [Na] increase by 1.6 mmol/L per 5.6 mmol/L decrease in [Glu], provides a reasonable estimate of the degree of hypertonicity due to losses of hypotonic fluids through osmotic diuresis at presentation of DKH or HHS and should guide the tonicity of replacement solutions. However, the corrected [Na] may change during treatment because of ongoing fluid losses and should be monitored during treatment.

Abnormalities of serum potassium concentration in dialysis-associated hyperglycemia and their correction with insulin: review of published reports

The main difference between dialysis-associated hyperglycemia (DH) and diabetic ketoacidosis (DKA) or nonketotic hyperglycemia (NKH) occurring in patients with preserved renal function is the absence of osmotic diuresis in DH, which eliminates the need for large fluid and solute (including potassium) replacement. We analyzed published reports of serum potassium (K(+)) abnormalities and their treatment in DH. Hyperkalemia was often present at presentation of DH with higher frequency and severity than in hyperglycemic syndromes in patients with preserved renal function. The frequency and severity of hyperkalemia were higher in DH episodes with DKA than those with NKH in both hemodialysis and peritoneal dialysis. For DKA, the frequency and severity of hyperkalemia were similar in hemodialysis and peritoneal dialysis. For NKH, hyperkalemia was more severe and frequent in hemodialysis than in peritoneal dialysis. Insulin infusion corrected the hyperkalemia of DH in most cases. Additional measures for the management of hyperkalemia or modest potassium infusions for hypokalemia were needed in a few DH episodes. The predictors of the decrease in serum K(+) during treatment of DH with insulin included the starting serum K(+) level, the decreases in serum values of glucose concentration and tonicity, and the increase in serum total carbon dioxide level. DH represents a risk factor for hyperkalemia. Insulin infusion is the only treatment for hyperkalemia usually required.

Abnormalities of serum potassium concentration in dialysis-associated hyperglycemia and their correction with insulin: review of published reports

The main difference between dialysis-associated hyperglycemia (DH) and diabetic ketoacidosis (DKA) or nonketotic hyperglycemia (NKH) occurring in patients with preserved renal function is the absence of osmotic diuresis in DH, which eliminates the need for large fluid and solute (including potassium) replacement. We analyzed published reports of serum potassium (K(+)) abnormalities and their treatment in DH. Hyperkalemia was often present at presentation of DH with higher frequency and severity than in hyperglycemic syndromes in patients with preserved renal function. The frequency and severity of hyperkalemia were higher in DH episodes with DKA than those with NKH in both hemodialysis and peritoneal dialysis. For DKA, the frequency and severity of hyperkalemia were similar in hemodialysis and peritoneal dialysis. For NKH, hyperkalemia was more severe and frequent in hemodialysis than in peritoneal dialysis. Insulin infusion corrected the hyperkalemia of DH in most cases. Additional measures for the management of hyperkalemia or modest potassium infusions for hypokalemia were needed in a few DH episodes. The predictors of the decrease in serum K(+) during treatment of DH with insulin included the starting serum K(+) level, the decreases in serum values of glucose concentration and tonicity, and the increase in serum total carbon dioxide level. DH represents a risk factor for hyperkalemia. Insulin infusion is the only treatment for hyperkalemia usually required.

Body fluid abnormalities in severe hyperglycemia in patients on chronic dialysis: review of published reports

Reports of dialysis-associated hyperglycemia (DH) were compared to reports of diabetic ketoacidosis (DKA) and nonketotic hyperglycemia (NKH) in patients with preserved renal function. Average serum values in DH (491 observations), DKA (1036 observations), and NKH (403 observations) were as follows, respectively: glucose, 772, 649, and 961 mg/dl; sodium, 127, 134, and 149, mmol/l; and tonicity, 298, 304, and 355 mOsm/kg. Assuming that euglycemic (serum glucose, 90 mg/dl) values were the same (sodium, 140 mmol/l; tonicity, 285 mOsm/kg) for all three states, the hyperglycemic rise in the average serum tonicity value per 100-mg/dl rise in serum glucose concentration was 1.9 mOsm/kg in DH, 3.5 mOsm/kg in DKA, and 8.1 mOsm/kg in NKH. Neurological manifestations in DH patients were caused by coexisting conditions (ketoacidosis, sepsis, and neurological disease) in most instances, and by severe hypertonicity (>320 mOsm/kg), with clearing after insulin administration, in a few instances. In 148 episodes of DH corrected with insulin only, the mean increase in serum sodium per 100-mg/dl decrease in serum glucose (Delta[Na]/Delta[Glu]) was -1.61 mmol/l. In agreement with theoretical predictions, Delta[Na]/Delta[Glu] was numerically smaller in patients with edema than in those with euvolemia. The average hyperglycemic increase in extracellular volume, calculated from changes in serum sodium concentration during correction of DH using insulin alone, was 0.013 l/l per 100-mg/dl increase in serum glucose concentration. A small number of DH patients presented with pulmonary edema rectified by insulin alone. DH causes modest hypertonicity, with few patients having neurological manifestations caused usually by other coexisting conditions. In contrast to DKA or NKH, which usually presents with hypovolemia, DH causes hypervolemia manifested occasionally by pulmonary edema. Insulin is adequate treatment for DH.

Acute lower limb ischemia is a frequent complication of severe diabetic hyperosmolarity

AIM

To describe the outcome of intensive care unit (ICU) patients admitted with a hyperglycaemic hyperosmolar non-ketotic syndrome (HHNS), with a specific analysis of precipitating conditions and complications including lower limb ischemia.

METHODS

Retrospective review of patients admitted in a university-hospital ICU for HHNS.

RESULTS

Seventeen consecutive patients (9F/8M, age: 75 years [57-81] (median [25-75% percentiles], Glasgow Coma score: 13 [12-14]) were admitted for HHNS over an 8-year period (1998-2005). On admission, the blood glucose level was 40.0 mmol/l [26.3-60.8], the corrected serum sodium concentration 167 mmol/l [158-174], and the calculated plasma osmolarity 384 mosmol/l [365-405]. All the patients presented with renal failure due to severe dehydration. An infection was identified as the precipitating factor in 8/17 cases. Three (18%) patients died in the ICU. Non-survivors were significantly older than survivors (P=0.02). Using univariate analysis, no other parameter measured on admission was related to mortality. Four patients (24%) presented with lower limb ischemia. They had a significantly more elevated blood urea nitrogen (P=0.03), creatinine phosphokinase level (P=0.04), and leukocyte count (P=0.02). The bilateral, symmetrical, and distal extremity involvement suggested diminished blood flow due to hyperviscosity, hypotension, vasoconstrictors, or cholesterol emboli rather than a proximal arterial obstruction as causative mechanisms. No patient was treated surgically. Ischemia reversed with fluid loading and resulted in toe dry digital necrosis.

CONCLUSION

HHNS is a rare but life-threatening cause of ICU admission. There is a high incidence of lower limb ischemia in HHNS patients, which may be related to dehydration and blood hyperviscosity.

Nonketotic hyperosmolar coma in a patient with type 1 diabetes-related diabetic nephropathy: case report

Nonketotic hyperosmolar coma (NHC) is characterized by severe hyperglycemia; absence of, or only slight ketosis; nonketotic acidosis; severe dehydration; depressed sensorium or frank coma; and various neurologic signs. This condition is uncommon in type 1 diabetes. Because of little or no osmotic diuresis in patients with diabetic nephropathy, increases in plasma osmolality and therefore the likelihood of neurologic symptoms are limited. A 20-year-old male patient with type 1 diabetes with chronic kidney disease on conservative treatment (glomerular filtration rate [GFR], 18 mL/dk) presented with acute nonketotic hyperosmolar syndrome. The patient was admitted presenting with thirst, fatigue, and drowsiness. Blood biochemistry levels were urea 87 mg/dL, creatinine 5.09 mg/dL, glucose 830 mg/dL, glycosylated hemoglobin (HbA1c) 8%, C peptide <0.3 ng/mL, sodium 131 mmol/L, chloride 93 mmol/L, potassium 5.2 mmol/L, and calculated serum osmolality 385 mOsm/kg. The presumptive diagnosis on admission was nonketotic hyperosmolar syndrome precipitated by urinary infection. This is the first case report of hyperosmolar coma in a patient with type 1 diabetes with chronic kidney disease.

Hyperglycemic crises in diabetes mellitus type 2

Hyperglycemic crises in type 2 diabetes are not rare and are becoming increasingly recognized as part of the spectrum of the presentation of previously undiagnosed diabetes mellitus and the decompensation of established diabetes mellitus. Contributing factors and associations are being elucidated but remain far from clear, particularly in DKA states. Medications commonly used in the treatment of many comorbid illnesses in patients with diabetes can themselves predispose to HHS. Endocrinopathies can contribute to insulin resistance and directly increase the glycemic load, leading to hyperglycemia. Medications such as the protease inhibitors may in the future lead to a better understanding of the pathophysiology of the metabolic derangements seen in the development of type 2 diabetes mellitus.

Hyperosmolar nonketotic coma precipitated by lithium-induced nephrogenic diabetes insipidus

A 45-year-old man, with a 10-year history of manic depression treated with lithium, was admitted with hyperosmolar, nonketotic coma. He gave a five-year history of polyuria and polydipsia, during which time urinalysis had been negative for glucose. After recovery from hyperglycaemia, he remained polyuric despite normal blood glucose concentrations; water deprivation testing indicated nephrogenic diabetes insipidus, likely to be lithium-induced. We hypothesize that when this man developed type 2 diabetes, chronic polyuria due to nephrogenic diabetes insipidus was sufficient to precipitate hyperosmolar dehydration.

Pro: glucose priming solutions should be used for cardiopulmonary bypass

Current knowledge suggests that risks of glucose-containing solutions for patients undergoing CPB are hypothetical at best. Instead, patients may benefit from reduced perioperative fluid requirements bestowed by intraoperative glucose-containing solutions. This risk-benefit analysis does not apply to patients undergoing circulatory arrest. This population endures the certain risk of global CNS ischemia; furthermore, no studies address the effect of glucose on fluid requirements in this population. As the benefit of glucose during operations requiring circulatory arrest is unknown and the probable risk of exacerbating global CNS is high, deliberate hyperglycemia in this population is probably unwise. For patients undergoing CPB without circulatory arrest, the risk-benefit balance falls in favor of adding glucose. The contention that hyperglycemia worsens CNS deficits after cardiac operation is undocumented and may not be true. For their patients undergoing CPB, clinicians should seriously consider using glucose-containing priming solutions.

Treatment of diabetic ketoacidosis and non-ketotic hyperosmolar diabetic coma

Although mortality of diabetic ketoacidosis (KA) has decreased during the past 20 yr to 1-2%, hyperosmolar non-ketotic coma (HNC) is still lethal in 20-30% of cases due to severe underlying conditions or to complications. The most frequent causes of death are infections and thromboembolic disorders. The strategies of initial treatment of KA and HNC are similar; in KA, insulin, fluid and electrolyte replacement have first priority. In HNC, rehydration and electrolyte administration are of primary importance. It is now generally recognized that insulin therapy is best performed using low doses (4-8 units/h); after institution of insulin treatment and rehydration there are rapid changes of fluid and electrolytes from the extra- into the intravascular space. In this situation it is a major therapeutic challenge to avoid complications due to hypokalaemia, hypophosphataemia, hypomagnesaemia and hypovolaemia. These complications should be avoided by adequate replacement, and particularly by regular clinical and laboratory monitoring. When blood glucose concentrations decrease below 14 mmols/l, blood glucose concentrations should initially be maintained at this level because rapid lowering below this level may increase the risk of brain oedema. Too-vigorous fluid replacement with crystalline solutions also increases the risk that brain oedema or complications like the adult respiratory distress syndrome will develop. If hypovolaemia persists in spite of adequate crystalloid solutions, colloid-containing fluids such as albumin should be administered. It is not established whether replacements of phosphate and magnesium have clinical benefits. Nevertheless, it is probably justified to administer phosphate and magnesium when their serum concentrations are below the normal range, particularly if the clinical situation is critical. Mortality from diabetic coma in industrialized countries may only be decreased by prophylaxis, i.e. by education of all diabetic patients and physicians to detect metabolic decompensation early.

Concepts of fluid therapy in diabetic ketoacidosis and hyperosmolar hyperglycemic nonketotic coma

Despite many advances in the overall treatment of type I diabetes mellitus during the last few years, no major advance has been made in decreasing the mortality rate of diabetic ketoacidosis or hyperosmolar hyperglycemic nonketotic coma. A major concern in both of these disease states is the development of cerebral edema during treatment. The guiding principles of therapy in both disease states are rehydration, electrolyte replacement, insulin therapy, and treatment of any underlying illnesses. If the patient is hypotensive, therapy begins with colloid or normal saline administration to support blood pressure. Fluid and electrolyte deficits should be calculated and replaced during 48 hours. Low-dose insulin therapy is employed for treatment of hyperglycemia. Neurologic function should be carefully monitored and mannitol administered if a change in neurologic function occurs.

Deaths associated with diabetic ketoacidosis and hyperosmolar coma. 1973-1988

We assessed the possibility of improvements in the management of the potentially fatal acute hyperglycaemic complications of diabetes by a review of all deaths in patients who presented to the Alfred Hospital, Melbourne, with diabetic ketoacidosis or hyperosmolar coma during the 16 years, 1973-1988. All late deaths of patients during hospitalization were included in the mortality data. In the 610 episodes of diabetic ketoacidosis (pH, 7.30 or lower) or hyperosmolar coma (osmolality, 350 mOsmol/kg or greater), only one death occurred as a result of the acute metabolic disturbance--in a patient who had suffered a cardiac arrest before admission to hospital. The over-all mortality rate was 6.2% (38 deaths). The mortality rate was 4.9% (26 deaths) for 528 episodes of diabetic ketoacidosis and 14.6% (12 deaths) for 82 episodes of hyperosmolar coma. Patients with diabetic ketoacidosis who died were older than were those who survived (64 +/- 13 years compared with 40 +/- 21 years, respectively; P less than 0.001). Mortality in patients with hyperosmolar coma did not relate to age, initial blood-glucose level or osmolality. Twelve deaths resulted from bacterial pneumonia and two deaths resulted from aspiration pneumonia. Other major causes of death were mesenteric and iliac thromboses (six cases), myocardial infarction (eight cases) and cerebral haemorrhage (two cases). Of the 26 deaths that were associated with diabetic ketoacidosis, only two deaths--as a result of aspiration pneumonia and bowel infarction, respectively--were assessed as potentially avoidable after the patient's admission to hospital. Eight of the 12 hyperosmolar-coma-associated deaths occurred in newly recognized diabetic patients in whom there were avoidable delays in diagnosis. We conclude that further improvements in outcome will be difficult to achieve, but that efforts should be directed towards the earlier diagnosis of diabetes and the earlier recognition and treatment of associated acute pulmonary and vascular complications.

Prognostic factors in the diabetic hyperosmolar state

To evaluate the current outcome of patients hospitalized with diabetic hyperosmolar state (DHS), we retrospectively studied 135 patients admitted to two general hospitals over an 11-year period. Mortality was 17%. Patients who died had a mean age of 77 years, compared to 68 years for the survivors (P = 0.008). They were also more likely to be nursing home residents (48 versus 23%, P = 0.01). Additionally, mean serum osmolality was significantly higher among those who died (383 versus 358 mosm/L, P less than 0.0001) as was blood urea nitrogen (81.3 versus 62.3 mg/dl, P = 0.006) and sodium (148 versus 137.4 mEq/L, P less than 0.001). However, mean glucose level and anion gap were similar among patients who died and patients who survived (1068 versus 1092 mg%; 23 versus 24 mEq/L, respectively). The presence of a chronic disease or an acute comorbid illness was not associated with mortality. Diminished physiologic reserve, attendant comorbidity, or functional disability may explain the effect of age and nursing home residence. High osmolality may indicate a greater water deficit and a more advanced stage of DHS at the time of diagnosis.

The hyperglycaemic, hyperosmolar non-ketotic syndrome: some aspects of management

Hyperosmolar non-ketotic coma is a relatively uncommon but important medical emergency. It is associated with a high mortality, which has changed little over the past twenty years. Approximately half of all patients give no prior history of diabetes and since patients often present with neurological abnormalities, resembling a cerebrovascular accident, the diagnosis can sometimes be delayed or missed. Therefore every patient presenting to hospital with a state of clouded consciousness or objective signs of neurological abnormality, should have a blood glucose estimation performed at an early stage. It has been suggested that a regimen of less aggressive early fluid replacement, with more attention being directed to the associated or underlying problems, may be beneficial to the patient. Claims that the continuing high mortality in non-ketotic coma can be improved by such measures await confirmation.