HHNC: A Potentially Fatal Complication of Diabetes
By Glenn C. Freas, MD, JD, FACEP
Hyperosmolar hyperglycemic nonketotic coma (HHNC) is a misnomer because the syndrome does not always cause coma. Nonetheless, the name has stuck over the years, and HHNC remains a serious complication of diabetes mellitus, with high potential for significant morbidity and mortality. HHNC is characterized by remarkable increases in serum glucose and osmolarity (> 600 mg/dL and 350 mOsm/kg water, respectively), lack of ketoacidosis, and major alterations in mental status (stupor, coma, or seizures). HHNC occurs roughly one-sixth as frequently as does diabetic ketoacidosis (DKA). Up to two-thirds of the patients have no prior history of diabetes, and HHNC is the presenting sign of diabetes. The vast majority of those who do have a history of diabetes have mild Type II disease.1
Mortality rates have frequently been quoted to be high and with a wide range, typically reported as 20-70%. More recent mortality estimates are in the lower end of that range.2 Mortality is related to precipitating conditions, coexisting disease, and the older age of the patients who get HHNC; however, there has been no conclusive correlation between the degree of hyperglycemia or hyperosmolarity and mortality. Precipitating factors include: infection (particularly gram-negative infections), renal insufficiency, GI hemorrhage, stroke, myocardial infarction, and pancreatitis. Certain drugs have been implicated as precipitating factors in the development of HHNC. They include thiazide diuretics, furosemide, phenytoin, glucocorticoids, cimetidine, chlorpromazine, propranolol, chlorthalidone, and ethacrynic acid. Trauma, burns, surgery, dialysis, and hyperalimentation have been described as contributing to the development of HHNC.3
Because of our frequent contact with diabetic patients, and because of the severity of this complication, it is imperative that emergency physicians be vigilant about recognizing this syndrome and aggressive in treating it and the associated comorbid conditions. Accordingly, this feature reviews the pathophysiology, diagnosis, and treatment of HHNC.
HHNC can be thought of as a combination of pancreatic and renal insufficiency precipitated by the causes listed above. Ordinarily, patients with hyperglycemia respond with an increase in insulin secretion. Since older patients are more likely to have pancreatic beta-cell insufficiency, this response may be inadequate. Physiologic stressors or precipitants of HHNC cause an increase in circulating catecholamines, which inhibits pancreatic insulin secretion and increases peripheral resistance to insulin. This combination of the inability to increase insulin secretion and the catecholamine-induced depression of insulin secretion and function leads to unchecked hyperglycemia. Because of compromised renal function from pre-existing renal disease and/or dehydration, glucose excretion is impaired, further aggravating hyperglycemia. An osmotic (hypotonic) diuresis occurs, worsening the dehydration and the body’s inability to excrete glucose. This vicious cycle leads to profound hyperglycemia, with serum glucose often above 1000 mg/dL.2
A persistent question is why ketosis does not develop in this setting of hyperglycemia. The answer is not clear. Simplistically put, there appears to be enough insulin secretion to block ketogenesis but not enough to prevent hyperglycemia. It may also be that the hyperosmolarity itself suppresses ketone formation, as does the lower level of lipolytic hormones in patients with HHNC.4 Whatever the reasons, patients with HHNC will typically have, at worst, a mild ketosis (perhaps from starvation as well as hyperglycemia) and mild accompanying acidosis (perhaps from lactic acidosis). Additionally, there are unusual patients who exhibit signs, symptoms, and laboratory parameters consistent with both DKA and HHNC.
Cellular dehydration and total body electrolyte losses result from internal shifts in body fluids and impaired ability to reabsorb water and electrolytes. Large amounts of glucose in the extracellular compartment draw free water into that compartment from the intracellular space. This preservation of the circulating volume of the extracellular space means that hypotension is not likely to develop until very late in the course of HHNC. However, because of the osmotic diuresis, a large amount of total body water is lost. Fluid losses in cases of HHNC typically range from 8-12 L.4 Although the rate of water loss exceeds sodium loss, there is a significant depletion of total body sodium. The osmotic diuresis also causes a loss of potassium. Because the osmotic diuresis is more prolonged than that of DKA, potassium deficits are even more pronounced with HHNC. Patients frequently require 400-1000 mEq of potassium supplementation to restore what was lost.1
Clinical Features and Diagnosis
The diagnosis of HHNC is relatively straightforward. Pitfalls include the lack of a history of diabetes and the lack of ketosis. Change in mental status, including seizures, and the work-up and treatment of the precipitating disease may temporarily obscure HHNC. Nonetheless, as sick as these patients are (they almost always look sick enough to get lab tests), and with the ready availability of bedside glucose determinations (registering as "HHH" or "high" in these patients), the diagnosis should be strongly suspected in any dehydrated patient, particularly the elderly, with a remarkably elevated glucose and a change in mental status.
Patients will have the classic symptoms of hyperglycemia, if they are able to communicate. In addition, they may have symptoms of their underlying disease. Typically, patients have a more prolonged time course of symptoms than do those with DKA. A change in mental status is what prompts an ED visit for most of these patients.5 On examination, patients with HHNC should have clear signs of severe dehydration, although some have cautioned that mucous membrane examination and skin turgor assessment may be unreliable in elderly patients.1 Shock may be present in a significant number of patients, particularly those with an infectious precipitant. Neurologic findings are the rule. Profound depression in the level of consciousness is common. Obtundation is correlated with the level of hyperosmolarity. Focal deficits are possible. Seizures in general are common, as are focal motor seizures.5
Laboratory findings confirm the diagnosis. A CBC, serum electrolytes, renal function studies, blood glucose, and an ABG are universally recommended. Other diagnostic studies are targeted to search for underlying illness and a precipitating cause. Serum sodium and potassium values may be low, normal, or high. When interpreting lab results, recall that there is a decrease in the serum sodium of 1.6 mEq/L for every 100 mg/dL increase in the blood glucose. Renal function is abnormal; in one series, the average presenting values for BUN and creatinine were 87 and 5.5 mg/dL, respectively.5 Blood glucose is over 500 mg/dL and frequently close to, or over, 1000 mg/dL. Blood gases may show a mild metabolic acidosis.
After addressing any airway and breathing problems, and treating seizures with benzodiazepines if necessary, the next priority is to establish reliable intravenous access that will allow for aggressive volume resuscitation. Whether they present in shock or not, patients with HHNC are severely volume depleted, and virtually all should get normal saline solution as their initial IV fluid. Meticulous attention to all measurable parameters, and documentation of those parameters on a flow sheet, are mandatory if the treating physician is to make reasonable judgments about the complexities of fluid and electrolyte therapy after the initial volume replacement establishes stable vital signs and adequate urine output. Changing the IV fluid to 0.45% normal saline may be warranted based on calculations of total body water deficits and total body sodium deficits. One-half the estimated total body water deficit should be replaced in the first 12 hours. Potassium replacement should begin as soon as adequate renal function has been confirmed.
Insulin requirements in HHNC are relatively small. Rehydration can go a long way toward reversing the pathophysiology described above. Some authors recommend small doses of regular insulin IV after vital signs are stable and urine output is adequate. One recommended regimen is 0.15 units/kg as an initial bolus, followed by a low-dose continuous infusion.6 Insulin administration will further aggravate hypokalemia, so it is advisable to begin potassium supplementation before insulin administration. It has been argued that some patients with HHNC do not require supplemental insulin.7 Arguments for low-dose insulin seem to outnumber those in favor of withholding insulin completely. In cases where HHNC coexists with severe acidosis, renal failure, or hyperkalemia, insulin is necessary.
An important predictor of complications of therapy is failure to continuously monitor response to treatment. Admission to an intensive care unit is mandatory. Frequent determinations of serum electrolytes and blood glucose should be used to recalculate and monitor water, sodium, and potassium replacement. Blood glucose should not fall below the 250-300 mg/dL range. Cerebral edema and congestive heart failure are uncommon, but reported, complications of treatment. An aggressive search for precipitating conditions and prompt treatment of those problems can decrease mortality.
HHNC is a severe complication of diabetes mellitus. Diagnosis is predicated on recognizing the typical clinical presentation and confirmed by laboratory findings of profound hyperglycemia and hyperosmolarity, combined with an absence of significant ketosis or acidosis. Therapy is directed at aggressive fluid and electrolyte replacement and treatment of precipitating causes.
1. Pope DW, et al. Hyperosmolar hyperglycemic nonketotic coma. Emerg Med Clin North Am 1989;7:849-857.
2. Podolsky S. Hyperosmolar nonketotic coma in the elderly diabetic. Med Clin North Am 1978;4:815-828.
3. Brody GM. DKA and HHNC. Top Emerg Med 1992;14:18-20.
4. Gerich JE, et al. Clinical and metabolic characteristics of hyperosmolar nonketotic coma. Diabetes 1971;20:228-238.
5. Arieff AI, et al. Nonketotic hyperosmolar coma with hyperglycemia: Clinical features, pathophysiology, renal function, acid-base balance, plasma-cerebrospinal fluid equilibria and the effects of therapy in 37 cases. Medicine 1972;51:73-94.
6. Kitabchi AE, et al. Optimal insulin delivery in DKA and HHNC. Diabetes Care 1982:Suppl 1;78-87.
7. West ML, et al. Quantitative analysis of glucose loss during acute therapy for hyperglycemia hyperosmolar syndrome. Diabetes Care 1986;9:465-471.