CONTENTS
- Rapid Reference 🚀
- Pathophysiology
- Definition & diagnosis
- Evaluation
- Treatment
- Podcast
- Questions & discussion
- Pitfalls
pathogenesis of hyperosmolar hyperglycemic state (HHS)
- HHS is often triggered by an acute stressor, which increases levels of cortisol and catecholamines (thereby reducing insulin sensitivity).
- HHS occurs in patients with enough insulin to prevent ketoacidosis, but not enough insulin to control hyperglycemia.
- Higher levels of insulin are required to control hyperglycemia, compared to the amount required to prevent ketogenesis (as seen in the way that many patients in the ICU develop mild insulin resistance and hyperglycemia – without developing ketoacidosis).
- Uncontrolled hyperglycemia causes an osmotic diuresis, with loss of water.
- Patients fail to compensate adequately for water loss by increasing oral water intake (e.g., due to baseline debility, bed-bound status, or a relatively insensitive central drive to maintain normal tonicity).
- Over a period of several days, uncontrolled water loss leads to a hypertonic state. This may lead to altered mental status (which exacerbates the patient's inability to drink an adequate amount of water).
osmolality versus tonicity
- Osmolality is a measurement of the number of particles in a solution.
- This is a physical chemistry property of any water-based solution.
- Osmolality can be measured in the laboratory (based on freezing point depression).
- Tonicity refers to how much the particles in a solution pull water across a semi-permeable biological membrane. This depends on the number of effective osmoles in the solution. For example:
- Sodium is an effective osmole, because it cannot cross the semi-permeable membrane. Therefore, sodium will tend to pull water across the membrane. If you bolus a patient with concentrated sodium solutions, this affects the volume of fluid in their cells.
- Urea is an ineffective osmole, because urea can diffuse freely into cells. Therefore, the urea cannot pull water across the membrane (the urea itself will just freely diffuse into the cells). Thus, if you bolus a patient with concentrated urea, this will not affect the volume of fluid in their cells.
- Glucose is sometimes an effective osmole, but sometimes it's not!
- When a patient first presents with HHS, low insulin causes glucose to stay outside the cells. This causes glucose to function as an effective osmole.
- During treatment with insulin, glucose enters the cells. At this point, glucose is an ineffective osmole.
- Since glucose doesn't freely diffuse across membranes this is a bit of an over-simplification – but hopefully a useful one.
- Tonicity is more clinically important than osmolality, because tonicity is what actually determines whether cells swell or shrink.
- Key principle: When we give patients with HHS insulin, we are dropping their tonicity (without necessarily affecting their osmolality). This explains why insulin shouldn't be given too rapidly – it may cause cellular swelling.
There is a maddening degree of variation between different definitions for HHS. The following definition is based most closely on the United Kingdom position statement, although even this document concedes that “a precise definition of HHS does not exist.” (25980647)
one working definition of HHS
- (1) High osmolality (>320 mOsm).
- This typically doesn't need to be measured directly.
- 🧮 Osmolality can be calculated using MDCalc.
- (2) Hyperglycemia (usually >>600 mg/dL or >>33 mM)
- (3) Absence of substantial ketoacidosis, for example,
- Beta-hydroxybutyrate <3 mM (Beta-hydroxybutyrate is the most precise way to define serum ketone levels, but not always necessary).
- <or>
- Anion gap below ~15 mM (anion gap is often sufficient, in the absence of uremia or other factors which affect anion gap).
- (4) Evidence of being clinically unwell.
- A patient with poorly controlled diabetes who just ate a large amount of carbohydrates could satisfy the criteria above without truly having HHS. For example, some patients with initial glucose values in the 600-800 mg/dL range can be safely discharged home from the emergency department (figure below). Therefore, to have true HHS, at least one of the following criteria should also be present:
- The patient appears clinically unwell (especially if the patient has altered mental status)
- <or>
- Laboratory abnormalities are profoundly deranged (e.g., glucose over ~1,000 mg/dL or ~55 mM)
- A patient with poorly controlled diabetes who just ate a large amount of carbohydrates could satisfy the criteria above without truly having HHS. For example, some patients with initial glucose values in the 600-800 mg/dL range can be safely discharged home from the emergency department (figure below). Therefore, to have true HHS, at least one of the following criteria should also be present:
diagnosis of HHS
- HHS generally isn't difficult to diagnose, given that essentially all ill patients will receive a glucose measurement upon evaluation in the emergency department.
- 🔑 Profound hyperglycemia should always raise a question of whether the patient has HHS.
DKA-HHS overlap
- Occasionally, patients may present with features of both DKA and HHS. Specifically, they may have a combination of:
- (1) Substantial ketoacidosis (e.g. beta-hydroxybutyrate >3 mM and substantial anion gap elevation)
- (2) Extreme hyperglycemia with marked hypertonicity (e.g., glucose >>1000 mg/dL or >>55 mM, with serum osmolality >> 320 mOsm)
- Overall, such patients should be treated based on the therapy algorithm for DKA (discussed here). The only difference is that if the patient is <40 years old, there may be an increased risk of cerebral edema if the osmolality decreases too rapidly. Thus, osmolality and sodium levels should be monitored as discussed below for HHS.
evaluate for an underlying process
- Perhaps the most important aspect of treating DKA or HHS is ensuring that no underlying process is missed (e.g., sepsis, pancreatitis, myocardial infarction, CVA). HHS carries a high mortality (often quoted ~15%) – but this is largely due to comorbid problems and triggers, rather than the HHS itself.(31142480)
- Common causes of HHS (25342831)
- Infection in about half of patients (common sources include pneumonia, urinary tract)
- Medication changes (glucocorticoid, thiazides, phenytoin, beta-blockers, anti-calcineurin immunosuppressives, HIV protease inhibitors, antipsychotics)
- Insulin nonadherence or inadequate dosing
- Stroke, MI
- Pancreatitis, trauma, heat exposure
- Depending on the history and physical examination, evaluation for the underlying process might include:
- Infectious workup (chest X-ray, urinalysis, blood cultures)
- CT scan of the abdomen and pelvis, lipase
- Evaluation for myocardial ischemia (troponin is indicated only if there is clinical suspicion for type-I MI)
- Stroke evaluation
evaluate the HHS
- Basic laboratory tests, including glucose, electrolytes, Ca/Mg/Phos
- Lactate level (may reflect sepsis or profound hypovolemia)
- Beta-hydroxybutyrate level (most precise way to quantify the presence and severity of ketoacidosis)
- Serum osmolality
- Not truly necessary, as formulas will generally estimate this fairly well in HHS (e.g., 🧮 MDCalc here).
Many sources suggest that the treatment of HHS and DKA are identical. This is a mistake. There are enormous physiological differences between these conditions that mandate different management. Lumping both conditions together into a single treatment algorithm will cause HHS to be treated overly aggressively. The strategy discussed here is based on the 2015 UK guidelines for HHS.(25980647) However, please note that there is no high-level evidence on HHS, so none of these protocols is rigorously evidence-based.
key principle: slow down!
- HHS is a deranged state which develops gradually over days to weeks. However, these patients generally adapt to their new state and often tolerate it relatively well.
- As a general rule of thumb, if an abnormal state develops gradually then it may be treated gradually.
- The primary risk of treating HHS is overly aggressive therapy, which may cause dangerous swings in electrolyte levels and osmolality.
- When in doubt, the safest approach to HHS is generally to correct abnormalities slowly.
- (This is unlike DKA, which often develops rapidly and requires more urgent therapy to correct acidosis.)
volume resuscitation
- The first step is gradual volume resuscitation using an isotonic fluid.
- Balanced crystalloids may usually be preferred (e.g., lactated Ringers or Plasmalyte).
- For a patient with uremic acidosis or NAGMA, isotonic bicarbonate could be considered.
- More discussion of pH-guided fluid selection here.
- Infuse fluids in a controlled fashion (e.g., 250-500 ml/hour). The rate and quantity of volume will depend on the clinical context and hemodynamic assessment.
- Effects of volume resuscitation typically include the following:
- Glucose decreases, ideally by ~70-90 mg/dL per hour or 4-5 mM/hour (due to dilution).
- Serum osmolality should decrease slightly.
- Serum sodium increases (due to osmotic shifting of water out of the vascular space).
- Monitor electrolytes and glucose intermittently.
- Switch to half-normal saline (0.45% sodium chloride) if the osmolality is increasing despite a positive fluid balance.
aggressive electrolyte repletion
- Potassium
- Potassium should be aggressively repleted as in DKA, with a target potassium >5.3 mM (in patients with normal renal function).
- Magnesium
- Hypomagnesemia is common. Magnesium should be aggressively repleted.
- Maintaining a magnesium level on the high end will tend to prevent Torsade de Pointes if the potassium level falls.
- Phosphate
- Phosphate should be repleted as necessary.
basal insulin following initial fluid resuscitation
- Unlike in DKA, there is less risk of deterioration after stopping the insulin infusion in HHS patients. Thus, insulin may be dosed in a more conservative fashion than in DKA.
- For patients previously on basal insulin, their home-dose basal insulin should be resumed.
- For patients not previously on basal insulin, initiation of weight-based, long-acting insulin should be considered (~ 0.3 units/kg glargine daily).
- Initiation of long-acting insulin as early as possible may facilitate a smooth transition off the insulin infusion (with reduced incidence of rebound hyperglycemia). In some cases, immediate long-acting insulin plus some doses of short-acting insulin may avoid the need for an insulin infusion at all.
what is the optimal glucose target?
- Acutely “normalizing” glucose in a patient with longstanding diabetes may induce a stress response, which is potentially detrimental. Furthermore, inducing hypoglycemia in this context is probably extremely dangerous.
- The initial glucose target over the first day may be 🎯 180-270 mg/dL (10-15 mM). Over subsequent days, this may be gradually lowered further.
indications: who needs an insulin infusion?
- Not every patient with HHS necessarily requires an insulin infusion. In many patients, volume resuscitation plus subcutaneous insulin will be perfectly adequate to achieve glycemic control.
- For patients with a glucose which remains very high after volume resuscitation (e.g., >600 mg/dL or >33 mM), an insulin infusion is reasonable.
- There's no immediate necessity to rapidly drop glucose levels. Thus, in situations where it is unclear whether an insulin infusion is necessary, it may be reasonable to first trial subcutaneous insulin.
contraindications to insulin infusion
- It's generally advisable to delay insulin until the glucose has already been reduced by dilution with isotonic crystalloid (volume resuscitation step above).
- Hypokalemia is a contraindication to insulin infusion.
dose the insulin infusion conservatively
- Don't use a bolus.
- The starting dose is 0.05 U/kg/hr (half of the initial dose used in DKA).
- The target should be to reduce the glucose by ~40-80 mg/dL per hour (2.2-4.4 mM). However, in practice, the glucose will often end up dropping faster than this.
- Stop the insulin when glucose approaches ~300 mg/dL
- Dropping the glucose below 🎯 180-270 mg/dL (10-15 mM) may increase the risk of cerebral edema.
- In HHS (without ketoacidosis), there is no mandate to overlap the insulin infusion with basal insulin. However, it may nonetheless be wise to initiate basal insulin as early as possible (to avoid rebound hyperglycemia when the insulin infusion is discontinued).
- Following discontinuation of the insulin infusion, if the glucose increases then it may ideally be treated with short-acting insulin. Additionally, if the patient is eating, then meal-associated insulin may also be helpful.
- If the patient's glucose falls <200 mg/dL (11 mM), stop insulin and initiate a D5W or D10W infusion. Avoid allowing the glucose to fall below 🎯 180-270 mg/dL (10-15 mM) during the first day of therapy.(31142480)
understanding the transition from hyperglycemia to hypernatremia
- Volume resuscitation and insulin administration generally cause the following changes:
- Serum sodium increases.
- Serum glucose decreases.
- Serum osmolality decreases.
- Tissue may swell slightly (due to reduction in effective tonicity, as discussed above).
- Don't interpret the rising sodium to necessarily be an indication to give free water! Although the sodium is increasing, the effective tonicity is decreasing.
- Development of hypernatremia represents an un-masking of underlying hypertonicity.
hypertonicity management for patients over ~40 years old
- The risk of cerebral edema due to rapid reduction in tonicity seems to be extremely low in these patients.
- A reasonable target might be to reduce the serum osmolality by ~20 mOsm/L per day (possibly faster in patients with altered mental status). However, if the osmolality decreases faster than intended, it will probably be safe.
- Administration of free water will often be required to reduce the osmolality appropriately. To estimate the amount of water required, calculate the volume of water required to reduce the serum sodium by 10 mEq/L (🧮 using MDCalc).
hypertonicity management for patients below ~40 years old
- The risk of cerebral edema is greater in younger patients.
- Target a reduction in serum osmolality by ~20 mOsm/day.
- During the initial resuscitation with crystalloid and insulin, serum osmolality is the best target. This may be either calculated or measured, but using a consistent methodology is important.
- After the glucose has normalized, then serum sodium will be an accurate reflection of serum osmolality. At this point, it's often easier to simply follow the serum sodium and target a reduction of 10 mEq/L per day in the sodium level.
- Management here is very similar to the management of chronic hypernatremia:
- Restrict free water intake.
- Calculate and gradually administer the amount of free water required to reduce the sodium by 10 mEq/L each day (🧮 using MDCalc).
- Follow electrolytes and adjust as needed.
- For more detail, see the chapter on hypernatremia.
- Patients with severe HHS often require a few days to gradually reduce their sodium and osmolality to normal.
- This may be caused by HHS.
- Serial creatinine kinase levels may be considered for patients with more dramatic hyperglycemia.
- More on rhabdomyolysis here.
Follow us on iTunes
The Podcast Episode
Want to Download the Episode?
Right Click Here and Choose Save-As
To keep this page small and fast, questions & discussion about this post can be found on another page here.
- Hyperglycemia rarely causes mental status changes unless the serum osmolality is >320 mOsm. Thus, if the serum osmolality is <320 mOsm and mental status is significantly abnormal, look for an alternative explanation (Anna 2015).
- Patients with glucose >600 mg/dL (>33 mM) don't necessarily have HHS, nor do they necessarily need an insulin infusion. Don't assume that every patient with severe hyperglycemia requires ICU admission. As with everything in medicine, the context is king (what else is going on? does the patient appear sick or well?).
- True HHS develops slowly and should be corrected slowly. When in doubt, make small adjustments.
- The morbidity of HHS is due largely to underlying triggers, so search carefully for them (e.g., infection or infarction).
- Don't forget to pay attention to sodium concentration and serum osmolality. Younger patients with HHS are at risk for cerebral edema if their tonicity is reduced too rapidly.
Guide to emoji hyperlinks
- = Link to online calculator.
- = Link to Medscape monograph about a drug.
- = Link to IBCC section about a drug.
- = Link to IBCC section covering that topic.
- = Link to FOAMed site with related information.
- = Link to supplemental media.
Going further
- Hyperglycemic Hyperosmolar Syndrome (HHS) – emDocs by Anand Swaminathan
- IBCC chapter on hypernatremia
References
- 11119020 Liamis G, Gianoutsos C, Elisaf MS. Hyperosmolar nonketotic syndrome with hypernatremia: how can we monitor treatment?. Diabetes Metab. 2000;26(5):403-405. [PubMed]
- 11285047 Milionis HJ, Liamis G, Elisaf MS. Appropriate treatment of hypernatraemia in diabetic hyperglycaemic hyperosmolar syndrome. J Intern Med. 2001;249(3):273-276. doi:10.1046/j.1365-2796.2001.0799a.x [PubMed]
- 25342831 Pasquel FJ, Umpierrez GE. Hyperosmolar hyperglycemic state: a historic review of the clinical presentation, diagnosis, and treatment. Diabetes Care. 2014;37(11):3124-3131. doi:10.2337/dc14-0984 [PubMed]
- 25905210 Milanesi A, Weinreb JE. Hyperglycemic Hyperosmolar State. In: Feingold KR, Anawalt B, Boyce A, et al., eds. Endotext. South Dartmouth (MA): MDText.com, Inc.; August 1, 2018. [PubMed]
- 25980647 Scott AR; Joint British Diabetes Societies (JBDS) for Inpatient Care; JBDS hyperosmolar hyperglycaemic guidelines group. Management of hyperosmolar hyperglycaemic state in adults with diabetes. Diabet Med. 2015;32(6):714-724. doi:10.1111/dme.12757 [PubMed]
- 28364357 Dhatariya KK, Vellanki P. Treatment of Diabetic Ketoacidosis (DKA)/Hyperglycemic Hyperosmolar State (HHS): Novel Advances in the Management of Hyperglycemic Crises (UK Versus USA). Curr Diab Rep. 2017;17(5):33. doi:10.1007/s11892-017-0857-4 [PubMed]
- 31142480 Karslioglu French E, Donihi AC, Korytkowski MT. Diabetic ketoacidosis and hyperosmolar hyperglycemic syndrome: review of acute decompensated diabetes in adult patients. BMJ. 2019;365:l1114. Published 2019 May 29. doi:10.1136/bmj.l1114 [PubMed]