DKA and HHS: Ketones, Acidosis, and Hyperosmolality
Recognize DKA, HHS, and mixed crises from ketones, acid-base status, and osmolality; sequence fluids, potassium, and insulin; confirm safe resolution.
Two adults arrive dehydrated with glucose above 600 mg/dL. One has marked ketonemia and acidemia; the other has severe hyperosmolality with little ketosis. The glucose concentration alone cannot choose their treatment pathway. Measure ketones, acid-base status, electrolytes, and osmolality while assessing circulation and mental status.
DKA and HHS are overlapping consequences of inadequate effective insulin action. They are not exclusive labels for type 1 and type 2 diabetes. The adult thresholds below follow the 2024 multisociety consensus; children require a pediatric protocol. [1]
Separate ketone production from water deficit
When insulin is insufficient and counterregulatory hormones rise, hepatic glucose production increases and peripheral glucose use falls. Glucose in urine draws water and electrolytes with it. Progressive volume loss reduces kidney perfusion and glucose clearance, worsening hyperglycemia. Access to water, renal reserve, infection, and delayed recognition influence how far this process progresses.
In DKA, insufficient insulin action also permits adipose lipolysis. Fatty acids reach the liver and support ketone production. Beta-hydroxybutyrate and acetoacetate contribute to the acid load. The liver exports ketones; extrahepatic tissues use them as fuel. Deep rapid breathing is respiratory compensation for metabolic acidosis, not a requirement for diagnosis. Nausea, vomiting, abdominal pain, and a fruity odor can occur, but their absence does not exclude DKA. Acetone is the volatile ketone responsible for the odor; unlike the ketoacids, it does not itself supply the acid load.
Two dimensions, with an important overlap
DKA dimension
Insulin deficit permits lipolysis. Ketone production increases. Bicarbonate falls and metabolic acidosis develops.
The immediate measurements are beta-hydroxybutyrate and venous pH or bicarbonate.
HHS dimension
Marked hyperglycemia sustains osmotic diuresis. Water deficit and hyperosmolality become profound. Residual insulin action often limits major ketogenesis.
The immediate measurements are glucose, sodium, osmolality, and circulating volume.
Mixed crisis
Hyperosmolality and significant ketoacidosis coexist. Use DKA-level insulin treatment while respecting HHS limits on osmotic correction.
Neither the diabetes label nor the highest glucose value can replace these measurements.
HHS often develops more gradually than DKA and may present with profound weakness, confusion, or focal neurological findings. Infection and insulin omission are frequent precipitants of hyperglycemic crises, but myocardial infarction, stroke, pancreatitis, medicines, and previously unrecognized diabetes also matter. Evaluate the individual trigger instead of assuming a universal ranking. Sepsis can coexist with DKA, and negative cultures do not independently exclude infection. [1]
Try it here · Checkpoint 1 of 3
Make your prediction before reading the choices. A first attempt is just a starting point.
Case 1
Show answer and explanations for case 1
A. Isolated respiratory acidosis. (Why this does not fit)
Low bicarbonate and marked ketonemia support a metabolic ketoacid process.
B. HHS without ketoacidosis. (Why this does not fit)
Glucose is below the HHS threshold and significant ketonemia with acidosis is present.
C. Starvation ketosis established by glucose below 250. (Why this does not fit)
Glucose below 250 does not exclude DKA, and the supplied findings meet current diagnostic components.
D. DKA under current adult criteria. (Best answer)
Glucose exceeds 200, ketones exceed 3, and acidosis is present; the older 250 glucose threshold should not exclude it.
Takeaway: Use the updated glucose threshold together with ketones and acidosis.
Use current diagnostic criteria and explain discordant results
Adult DKA requires three components: glucose at least 200 mg/dL or a prior history of diabetes regardless of presenting glucose; significant ketosis, preferably beta-hydroxybutyrate at least 3.0 mmol/L or urine ketones at least 2+ when blood testing is unavailable; and metabolic acidosis, with pH below 7.3 or bicarbonate below 18 mmol/L. A patient with known diabetes can therefore have DKA below the old glucose threshold of 250. Euglycemic DKA in the consensus refers to glucose below 200 mg/dL with the other components present. [1]
HHS requires glucose at least 600 mg/dL; calculated effective osmolality above 300 mOsm/kg or total osmolality above 320; absence of significant ketonemia, with beta-hydroxybutyrate below 3.0 mmol/L; and no major acidosis, with pH at least 7.3 and bicarbonate at least 15 mmol/L. Mild ketonemia can occur. The diagnostic bicarbonate boundary is not identical to the insulin-pathway boundary: a hyperosmolar patient with bicarbonate below 18 or other significant acidosis needs assessment for a mixed process and higher-intensity treatment as indicated.
Total calculated osmolality ≈ effective osmolality + BUN (mg/dL)/2.8.
The anion gap = sodium − (chloride + bicarbonate). Urea contributes to total osmolality but much less to effective tonicity because it crosses cell membranes relatively freely. These equations are estimates; use measured values and clinical trends when available.
Blood beta-hydroxybutyrate is preferred over urine ketones for diagnosis and monitoring. Nitroprusside urine testing mainly detects acetoacetate, so it can underestimate early DKA when beta-hydroxybutyrate predominates and remain positive during recovery as the balance shifts. Vomiting-induced alkalosis or respiratory alkalosis can partially mask acidemia; reconcile pH, bicarbonate, ketones, and the gap instead of using pH alone to dismiss a mixed acid-base disorder.
SGLT2 inhibitors can permit ketoacidosis with less striking hyperglycemia, particularly during illness, fasting, dehydration, or insulin reduction. Stop the drug when ketoacidosis is suspected and evaluate regardless of glucose. The empagliflozin label advises withholding it for at least three days before surgery or procedures involving prolonged fasting when possible, with resumption after clinical stability and oral intake return.
Agent-specific instructions matter. This perioperative resumption instruction does not establish routine safety after an episode of DKA. For type 2 diabetes after DKA resolution, the 2024 consensus does not routinely recommend initiating or continuing an SGLT2 inhibitor because safety data are lacking. [1][2]
Not every elevated gap is DKA. Measure lactate when shock, tissue hypoperfusion, or metformin-associated lactic acidosis is plausible. A person taking only metformin can still develop DKA, so the medication list does not settle the cause. Alcohol-associated or starvation ketosis and toxic exposures belong in the differential. A normal osmolar gap cannot exclude a late toxic-alcohol presentation after metabolism has progressed. Severe suspected metformin toxicity with profound lactic acidosis, shock, or renal failure warrants urgent toxicology and dialysis assessment using EXTRIP guidance. [3][8]
Coordinate fluids, potassium, and insulin
Initial care assesses airway, circulation, mental state, precipitating illness, and the required monitoring setting. Severe DKA, HHS, critical precipitating illness, or altered mental status generally needs intensive care. Selected uncomplicated mild or moderate DKA can be managed with closely monitored subcutaneous rapid-acting insulin at one- to two-hour intervals under an established protocol; this is not an appropriate shortcut for severe DKA or HHS. [1]
In adults without cardiac or renal compromise, isotonic saline or balanced crystalloid may start at 500-1000 mL/hour over the first 2-4 hours, followed by adjustment to circulation, sodium, fluid balance, and osmolality. Older adults and patients with heart failure or advanced kidney disease may need smaller boluses, such as 250 mL, with frequent reassessment. Do not prescribe one fixed total volume to everyone. Balanced fluids are an accepted option and can reduce chloride-related acidosis; randomized-trial subgroup evidence supports their use in DKA. [1][4]
Serum potassium can be high even when whole-body potassium is depleted because insulin deficiency and hypertonicity shift potassium outward while osmotic diuresis loses it in urine. Insulin and correction of the crisis can then lower serum potassium rapidly. Kidney failure changes this assumption: an anuric patient may retain potassium and should not receive automatic replacement.
If potassium is below 3.5 mmol/L, replace potassium and delay insulin until it rises above 3.5. When potassium falls below 5.0, replacement is generally added in patients who can excrete potassium, aiming for 4-5 mmol/L. Common replacement uses 20-30 mmol per liter of fluid, but the actual rate depends on the deficit, route, kidney function, and monitoring. Recheck potassium about two hours after insulin begins and at least every four hours thereafter, more often when needed. [1]
Once potassium permits, intravenous insulin for DKA or mixed DKA/HHS commonly starts at 0.1 units/kg/hour. HHS without significant ketosis or acidosis generally uses 0.05 units/kg/hour. When DKA glucose falls below 250 mg/dL, add 5-10% dextrose and usually reduce insulin to 0.05 units/kg/hour while continuing to clear ketones. Giving glucose at this stage is purposeful: it permits ongoing insulin treatment without hypoglycemia. In euglycemic DKA, dextrose may be needed from the outset.
Try it here · Checkpoint 2 of 3
Make your prediction before reading the choices. A first attempt is just a starting point.
Case 15
Show answer and explanations for case 15
A. Delay insulin until oral intake resumes. (Why this does not fit)
Ketoacidosis requires insulin treatment even when the patient cannot yet eat; dextrose can support ongoing treatment as needed.
B. Frequent rapid-acting subcutaneous insulin via a validated, monitored protocol. (Best answer)
Selected uncomplicated mild or moderate cases can be managed this way.
C. Use subcutaneous correction insulin at routine mealtimes as the sole insulin dosing schedule. (Why this does not fit)
Mild DKA still needs frequent protocolized dosing and monitoring; routine mealtime correction alone is insufficient.
D. No insulin because mild DKA resolves from fluids alone. (Why this does not fit)
Insulin is needed to suppress ongoing ketogenesis.
Takeaway: Subcutaneous DKA treatment is a monitored option for selected patients, not a universal substitute.
Control the rate of correction and recognize complications
Check bedside glucose every 1-2 hours during treatment and serial electrolytes, creatinine, beta-hydroxybutyrate, and venous pH, commonly every four hours. In HHS, follow osmolality as well. The adult consensus recommends limiting HHS glucose decline to 90-120 mg/dL/hour, osmolality decline to 3-8 mOsm/kg/hour, and sodium decline to no more than 10 mmol/L in 24 hours. These are monitoring constraints, not targets to chase without evaluating the patient. [1]
Sodium commonly rises initially as glucose falls because water shifts back into cells. A 100 mg/dL glucose reduction can produce an approximately 1.6 mmol/L sodium rise. That rise alone does not justify switching to hypotonic fluid. Consider 0.45% saline when osmolality is not declining despite adequate positive fluid balance and appropriate insulin, integrating the whole trajectory rather than sodium in isolation.
Hypoglycemia, hypokalemia, fluid overload, acute kidney injury, and thrombotic events need active surveillance. HHS is a prothrombotic setting, and prophylactic anticoagulation is generally appropriate unless contraindicated; full-dose treatment requires a separate indication. New focal deficits or reduced consciousness require urgent reassessment for hypoglycemia, cerebral injury, stroke, thrombosis, or other causes. Do not label every neurological change as cerebral edema without evaluation.
Bicarbonate is not routine DKA treatment. The adult consensus permits consideration when pH is below 7.0, with appropriate monitoring. Routine phosphate replacement has not improved outcomes; consider it when low phosphate accompanies clinically important muscle weakness, respiratory compromise, or cardiac dysfunction. These adjuncts do not replace insulin and fluid treatment.
Pediatric cerebral injury is not adequately explained as a simple consequence of giving any fluid quickly. Perfusion and inflammatory processes also matter. The PECARN randomized trial found no significant differences in its neurological outcomes between the studied fluid rates and saline concentrations. This supports evidence-based pediatric resuscitation, not unrestricted fluids or importing the adult regimen into children. Follow ISPAD-based pediatric monitoring and emergency treatment for suspected cerebral injury. [5][6]
Prove resolution before transitioning treatment
DKA resolution requires beta-hydroxybutyrate below 0.6 mmol/L and either venous pH at least 7.3 or bicarbonate at least 18 mmol/L; glucose should ideally also be below 200 mg/dL. Do not stop insulin solely because glucose normalized or the anion gap closed. Conversely, saline-related hyperchloremic acidosis can leave bicarbonate low after ketones clear, so the gap and bicarbonate need contextual interpretation. Urine ketones are not the resolution endpoint. [1]
The consensus notes that HHS resolution has no universally accepted definition. Its proposed recovery assessment includes osmolality below 300 mOsm/kg, glucose below 250 mg/dL, urine output above 0.5 mL/kg/hour, and improved cognition. Urine output must be interpreted per body weight and over time; one fixed volume such as 200 mL/hour is not a universal target. Persistent oliguria may reflect unresolved volume deficit or kidney injury and requires assessment.
When transitioning from intravenous to subcutaneous insulin, provide adequate overlap so the infusion does not end before basal coverage is established. The 2024 crisis consensus describes 1-2 hours; the DKA section of the 2026 ADA hospital standard specifies basal insulin 2-4 hours before stopping intravenous insulin. Follow the current institutional protocol and the insulin preparation's onset.
Choose the regimen using prior treatment, nutritional intake, renal function, hypoglycemia risk, and the likely ongoing insulin requirement. Confirmed type 1 diabetes requires ongoing insulin, including basal coverage during illness. Some people with ketosis-prone type 2 diabetes may have a different longer-term requirement after recovery; do not infer permanent beta-cell loss from the crisis alone. [1][7]
Discharge planning should address the actual precipitant and practical barriers: affordable insulin, supplies, a usable glucose and ketone plan, pump troubleshooting and backup injections, illness instructions, follow-up, and mental-health support when needed. Ask respectfully about missed doses and access. Education cannot guarantee a fixed recurrence rate; prevention requires a plan the patient can carry out.
Try it here · Checkpoint 3 of 3
Make your prediction before reading the choices. A first attempt is just a starting point.
Case 27
Show answer and explanations for case 27
A. These findings support HHS resolution. (Best answer)
Urine output exceeds 0.5 mL/kg/hour, and the other supplied recovery measures meet the consensus framework.
B. Normal pH alone would establish recovery even if osmolality remained high. (Why this does not fit)
HHS is chiefly a hyperosmolar disorder, so pH cannot replace its recovery criteria.
C. Glucose below 250 alone is sufficient regardless of cognition. (Why this does not fit)
Cognition and osmolality must also improve.
D. Recovery is impossible until urine output exceeds 200 mL/hour. (Why this does not fit)
The output threshold is weight-based, not a universal 200 mL/hour target.
Takeaway: Assess HHS recovery across metabolic, renal, and neurological domains.
A. No ketoacidosis because pH is above 7.3. (Why this does not fit)
The bicarbonate and ketone findings cannot be dismissed by pH alone.
B. Isolated vomiting-induced alkalosis explains every value. (Why this does not fit)
Alkalosis alone does not explain marked ketonemia with low bicarbonate and a high gap.
C. DKA with a mixed acid-base disorder masking acidemia. (Best answer)
Ketosis and bicarbonate below 18 satisfy the metabolic components of DKA. The gap is 136 - (96 + 16) = 24. PaCO2 is below the expected compensation of about 32 +/- 2 mmHg, indicating an additional respiratory alkalosis. Vomiting can also contribute metabolic alkalosis; a near-normal pH does not exclude the ketoacidosis.
D. HHS without a ketoacidotic component. (Why this does not fit)
The glucose and biochemical pattern do not fit isolated HHS.
Takeaway: A near-normal pH can conceal simultaneous acidifying and alkalinizing processes.
A. Use maintenance fluids only because reduced ejection fraction excludes intravascular depletion. (Why this does not fit)
Heart failure does not exclude intravascular depletion from HHS. Assess circulation and titrate resuscitation carefully.
B. Use the standard 500-1000 mL/hour initial schedule without modification. (Why this does not fit)
This schedule is intended for adults without cardiac or renal compromise. Age, low ejection fraction, and CKD warrant smaller increments and reassessment.
C. Smaller isotonic boluses, such as 250 mL, with frequent circulatory and pulmonary reassessment. (Best answer)
Cardiorenal vulnerability requires individualized resuscitation rather than an unmodified standard large-volume schedule.
D. Use routine furosemide before evaluating circulating volume. (Why this does not fit)
Diuresis may worsen depletion and should be based on the actual volume state.
Takeaway: Fluid need and fluid tolerance must be assessed together.
HHS without significant ketoacidosis generally uses the lower fixed-rate infusion.
B. 0.01 units/kg/hour as the standard HHS infusion rate. (Why this does not fit)
The consensus pathway generally starts at 0.05 units/kg/hour in HHS without significant ketosis or acidosis, with individual adjustment and monitoring.
C. 0.1 units/kg/hour solely because glucose is above 600. (Why this does not fit)
The higher rate is used for DKA or significant mixed ketoacidosis rather than glucose alone.
D. 0.5 units/kg/hour. (Why this does not fit)
This is ten times the usual HHS starting infusion rate and would risk excessive glucose and potassium decline.
Takeaway: Use ketones and acid-base status to distinguish the insulin pathways.
A. Increase insulin to prioritize glucose correction without following osmolality. (Why this does not fit)
HHS management must follow osmotic trends, because glucose alone cannot establish a safe correction rate.
B. Hypotonic saline solely whenever any sodium result exceeds baseline. (Why this does not fit)
Sodium alone is insufficient; the stem supplies the more relevant persistent osmolality problem.
C. Use of 0.45% saline when the overall trajectory supports it. (Best answer)
The consensus identifies failure of osmolality to decline despite adequate treatment as a context for hypotonic saline.
D. Give another large isotonic bolus without reassessing the established positive balance. (Why this does not fit)
Persistent hyperosmolality does not automatically mean more isotonic volume is needed when balance and circulation have improved. Reassess water replacement and the whole trajectory.
Takeaway: Hypotonic fluid decisions require evidence about volume balance and osmolality, not an isolated sodium value.
A. Withhold phosphate because trials found no benefit from routine replacement. (Why this does not fit)
Routine replacement lacks benefit, but respiratory muscle weakness with phosphate below 1.0 mmol/L is a selective indication to consider treatment.
B. Carefully monitored phosphate replacement. (Best answer)
Low phosphate with respiratory compromise is a context in which replacement may be appropriate.
C. Routine phosphate in every DKA patient regardless of level or symptoms. (Why this does not fit)
The indication here is selective; routine replacement has not improved outcomes.
D. Stop insulin and wait for phosphate to recover without replacement. (Why this does not fit)
Withholding the treatment for ketoacidosis does not address the symptomatic phosphate deficit; coordinate monitored replacement with ongoing crisis care.
Takeaway: Replace phosphate selectively when the deficit has clinically important consequences.
