DIABETIC KETOACIDOSIS


I. Review of normal lipid metabolism

  1. Triglycerides in adipose ==lipolysis==> Long-chain FAs
  2. Long-chain FAs==hepatic beta-oxidation==>Acetyl CoA
  3. Acetyl CoA==hepatic ketogenesis==>ketone bodies
  4. Ketone bodies are Beta-hydroxybutyrate and Acetoacetate
  1. Beta-OHB is oxidized to AcAc-; their relative concentrations depend on redox state of cell; Beta-OHB predominates in situation favoring reductive metabolism (e.g. decreased tissue perfusion, met. acidosis, catabolic states--like DKA!)
  2. Typical ratio Beta-OHB:AcAc- is 3:1; us. increases in DKA

II. Hormonal influences on glucose and lipid metabolism

  1. Insulin
  1. In liver, increases glu uptake from portal blood; stimulates glycogenesis, inhibits glycogenolysis and gluconeogenesis
  2. In skeletal muscle, increases glu uptake from blood, stimulates protein synth, inhibits proteolysis
  3. In adipose tissue, required for glu and lipoprotein uptake from blood; stimulates lipogenesis, inhibits lipolysis
  4. Tissues which don't require insulin to transport glucose into cells: brain, renal medulla, formed blood elements
  1. Counterregulatory hormones: glucagon (major player in DKA), epi/norepi, cortisol, growth hormone (no acute effects, only over days-weeks)
  1. Glucagon: increases hepatic beta-oxidation, ketogenesis, gluconeogenesis and glycogenolysis; decreases hepatic FA synth.
  2. Epi/Norepi: increase hepatic gluconeogenesis & glycogenolysis; increases adipose lipolysis; decreases peripheral glu utilization
  3. Cortisol: major effect is decreased peripheral glu utiliz; little effect on production
  4. Growth hormone: increases hepatic gluconeogenesis and glycogenolysis; increases adipose lipolysis
  5. In high counterreg. hormone states (see above), require high levels of insulin to avoid progressive hyperglycemia and ketoacidosis--glucagon levels in DKA are 5-6 x nl*

III. Pathophysiology of DKA--summary

  1. Results from relative insulin deficiency in setting of elevated counterregulatory hormones.
  1. Both are necessary; insulin deficiency alone causes just hyperglycemia and ketosis, not DKA. Dietary indiscretion causes only hyperglycemia.
  2. Low insulin favors lipolysis; high glucagon/insulin ratio stimulates beta-oxidation; both processes drive forward ketogenesis
  1. Causes of insulin deficiency: insulin withdrawal, Beta-cell failure, increased insulin resistance.
  2. Causes of counterregulatory hormone excess: fasting, dehydration, acidosis, fever, surgery, pregnancy, infection, ?psychol. stress?
  3. Summary of DKA metabolic state:
  1. Increase in serum glu, amino acids, & FFA; glycogen depletion
  2. Increase in serum ketones and resultant metabolic acidosis

V. Pathophysiology of DKA--details

  1. Hyperglycemia leads to osmotic diuresis
  1. Loss of free water causes hyperosmolarity of extracellular fluids, mostly due to high glucose concentration. creating high osmotic gradient between intracellular & extracellular compartments
  1. Typical free water loss is 100-150 ml/kg; very high serum glucose indicates greater deficit
  1. Lack of insulin prevents glucose entry into cells which might correct this osmotic gradient
  2. Result is fluid shift to from intracellular to extracellular compartments
  1. This shift tends to mask severity of volume depletion in by minimizing its hemodynamic consequence
  2. When treatment (e.g. insulin) lowers serum osmolarity, can get abrupt drop in circulating volume from return of fluid to intracellular compartment, esp. if insufficient vol. replacement is concurrently given.
  1. Passive electrolyte depletion also results from osmotic diuresis. Typical losses:
  1. 5-13 mEq/kg Na
  2. 4-10 mEq/kg K
  3. 0.5-4 mMol/kg PO4
  1. If pt can take adequate PO fluids to replace losses from osmotic diuresis, and has nl renal function, will keep peeing out glucose as it's formed, so will have a somewhat high but stable serum glucose.
  2. If pt cannot take adequate PO fluids (because a little kid who can't communicate thirst, or vomiting, or obtunded), can develop severe hypovolemia
  1. Renal perfusion and GFR fall, so
  1. Can't pee out glu at rate at which it's being formed, so get progressive hyperglycemia, exacerbating the cycl
  1. Because of this, often can reduce serum glu from >600 mg/dl (which means >10% vol. depletion and decreased renal fn) to about 300 with IVF alone
  1. Can't excrete organic acids, exacerbating metabolic acidosis from ketones
  1. Reduced perfusion of other tissues adds lactic acidosis to exacerbate acidosis from ketones
  1. Usual sequence of events:
  1. Hypoinsulinemia results in hyperglycemia, causing
  2. Osmotic diuresis, leading to
  3. High osmotic gradient between intracellular & extracellular compartments, and
  4. Fluid shift from intracellular to extracellular compartment
  1. Sodium metabolism
  1. Large Na losses ensue in DKA
  2. Urinary [Na] is us. 40-100 mEq/l
  3. Free water is lost in excess of Na, but serum [Na] is usually low, because hypertonic serum draws water from intracellular space
  4. For every increase of 100mg/dl in serum [glu] beyond 100mg/dl, serum [Na] falls about 1.6 mEq/l
  5. Hypertriglyceridemia, common in severe DKA, can cause factitiously low readings of serum [Na]; suspect if <120
  1. Potassium metabolism
  1. Very large losses ensue in DKA (us. 3-5mmol/kg), due to:
  1. K movement from intra- to extracellular space in response to acidosis and also to protein catabolism
  2. Vomiting
  3. Kaliuresis, which results from
  1. Osmotic diuresis; "drags" electrolytes along with water
  2. Volume depletion and reactive hyperaldosteronemia
  1. Still, serum [K] is us. nl/high before Tx initiated, due to intra- to extracellular shift and low GFR (see above)
  1. Can get syndrome of hyperkalemia even with nl serum [K], because of the low intracellular [K] and thus high concentration gradient
  1. Peaked T's & widened QRS's on EKG
  2. Ventr. arrhythmias, asystole
  1. Often get acute hypokalemia (U waves on EKG; ventr. ectopy) upon initiating treatment (lowest about 4h after), because of
  1. Dilution of contracted vasc. compartment
  2. Kaliuresis upon improved GFR
  3. K movement to intracellular space upon resolution of acidosis and in response to exogenous insulin
  1. Can withold K administration for initial 1-2h of Tx, but should start K replacement after that even if serum [K] is nl/high
  1. Phosphate metabolism
  1. Us. severely depleted (0.5-1.5mmol/kg): moves out of intracellular space & lost in urine
  2. But us. nl/sl. high serum [PO4] before initiating Tx
  3. Serum [PO4] usually falls after starting Tx
  4. PO4 depletion leads to a fall of [2,3-DPG] inside RBCs, which theoretically results in tighter O2 binding to HGb, lowering tissue O2 delivery
  5. Also theoretically, acidosis has the opposite effect on O2-HGb binding
  6. It is controversial whether PO4 replacement is indicated in DKA:
  1. Serum [PO4] us. does not plummet during DKA Tx, and it's unclear whether, even if falls to abnormally low levels, has any clinical effects
  2. PO4 replacement apparently doesn't alter clinical course, and can reportedly--though rarely--cause falls in serum [Ca] and [Mg]
  1. Bicarbonate metabolism
  1. Serum [HCO3] always low in DKA due to met. acidosis
  2. But there is no depletion of HCO3, because KA's & lactate made during DKA are ultimately metabolized to HCO3 when insulin is given
  3. So, HCO3 administration is rarely needed to bring about resolution of met. acidosis in DKA
  1. Other metabolic abnormalitis
  1. Azotemia--Occurs frequently; due to low GFR and increased protein catabolism
  2. *Spuriously high serum [Cr] can result from the presence of AcAc- in the serum, which is picked up by a common lab assay for Cr ("alkaline picrate assay")
  3. "Idiogenic Osmoles"--unidentified (poss. glucose polymers), osmotically active chemicaly produced intracellularly in brain during prolonged periods of hypertonicity; apparently an adaptive mech. to avoid brain shrinkage and tearing of bridging veins; may be involved in idiopathic cerebral edema in DKA (see below)

VI. Derangements of organ systems

  1. Respiratory: Increased resp. drive due to acidosis leads to "Kussmaul" respirations and hypocapnia, but can't compensate for pH <7.1
  1. CV:
  1. Met. acidosis causes overall decrease in circulatory tone through the following mechanisms:
  1. Depression of brainstem vasomotor center
  2. Decreased arteriolar SM tone
  3. Direct decrease in myocardial contractility
  1. Usually compensated for by high catechol. state and high serum [Ca], but can get circulatory collapse and shock
  1. Tx: IVF & HCO3
  2. Especially if doesn't respond to those measures, think gm-negative sepsis or silent MI
  1. HEME:
  1. Thrombosis is an occas. complication in adults
  1. Presumably due to dehydration and increased blood viscosity, low CO, frequent presence of atherosclerosis
  2. Cerebral vessels most commonly affected
  3. Us. presents hours-days after admission
  1. GI: Acidosis and K depletion cause
  1. Paralytic ileus (abd. distension, pain, absent BS)
  2. Gastric dilatation
  3. Decreased splanchnic blood flow
  1. CNS
  1. Commonly only limited derangement, due to hypocapnia:
  1. Paradoxically nl/high CNS pH, because protons & HCO3 don't cross blood-brain barrier, though CO2 does--so rarely obtunded
  2. Reflex cerebral vasoconstriction and decreased blood flow; unknown significance or relation to cerebral edema
  1. Idiopathic cerebral edema--a rare but unpredictable, unpreventable, and usually rapidly fatal complication of DKA
  1. Seen more often in kids than adults; unknown incidence
  2. No warning signs or known predictive factors
  3. Us. presents sev. hrs post-onset of tx, when blood vol. restored & serum [glu]& acidosis coming under control
  1. Sudden-onset decrease in level of consc.
  2. Neuro signs (pup. dilat., decerebrate or decorticate posturing)
  3. Rapid progression to coma & death, often with uncal herniation; sometimes resolves spontaneously with/without sequelae
  1. Theories of pathogenesis
  1. "Osmotic dysequilibrium"--idiogenic osmoles (see above) create increased intracellular oncotic pressure in CNS, maintaining brain size during acute DKA but causing swelling when hypertonicity resolved; some supportive evidence exists
  2. Insulin theory--only known cases occurred with insulin therapy--?causative role for insulin or rapid resolution of hyperglycemia?
  3. Paradoxical CNS acidosis due to HCO3 administration (see below) seems to have no connection
  4. Series of 119 pts 13mo-30y found that neurol. complications more likely to occur when Na doesn't rise as glu fals during tx of DKA. Concluded that excess free H2O administration puts pts at risk for cerebral edema & suggest giving IVF such that Na rises as glu falls (J. Peds 117:22, 1990)
  1. Dx: Head CT; diff includes saggital sinus or cerebral artery thrombosis
  2. Tx: Mostly supportive; NICU care
  1. Intubation & hyperventilation; reduces IC pressure by cerebral vasoconstriction
  2. Mannitol (low dose, i.e. 0.25g/kg) may reduce edema, but may also
  1. Aggravate hypertonicity
  2. Be nephrotoxic
  3. Cause rebound increase in IC pressure
  1. Steroids are not effective and would cause undesirable metabolic changes

VII. Presentation of DKA

  1. Dx is usually easy, but may be tricky in little kids in whom DKA is frequently the first noticed manifestation of DM
  2. Often misdx'd as bacterial sepsis because of pallor, lethargy, tachypnea, dehydration; routine labs reveal dx*
  3. Hx:
  1. Lethargy
  2. Abd. pain (think urinary process!)
  3. Polyuria
  4. Dyspnea
  5. Missed insulin dose
  6. Weakness
  7. Vomiting
  8. Thirst
  9. Change in MS
  10. Complaints suggestive of infectious illness
  11. Search for precipitating condition/underlying cause: infection is most common(most frequently urinary, pneumonia, or septicemia; rarely, mucormycosis), fever, surgery/trauma, fasting, dehydration, pregnancy, pancreatitis, MI, CVA, Coma of any cause, poss. psych. stress--no cause identified in 50% of cases*
  1. Px
  1. Fruity breath
  2. Kussmaul's (but can have nl resp. rate at pH <7.2!)
  3. Tachycardia
  4. BP us. nl (hypotension ass'd with rel. risk of death >2)
  5. Often afebrile, even with infection; fever strongly suggests infection
  6. Abd. distension and absent BS (if with ileus)
  7. Change in MS (20% show no change; 10% are unconscious)

VIII. Lab workup

  1. Appropriate admit labs:
  1. SMA7, Ca, PO4--calculate serum osmolarity (2(Na + K) + Glu/18 + BUN/2.8); avg. osmol. on pres. 310-316mOsm
  2. ABG; avg pH on pres. about 7.1
    1. Venous blood gas for initial evaluation in pts with DKA--very close correlation in pH (mean diff. between arterial and venous pH was 0.03)--ditto for arterial vs. venous HCO3 and CO2. Accurary might be less in situations where mixed acid-base disorder or hypoxemis are present (Ann. Emerg. Med. 31:459, 1998--AFP)
  3. U/A--Urine ketones have high negative predictive value for DKA (Ann. Emerg. Med. 34:342, 1999--AFP); note that captoprol and other free sulfhydryl drugs can cause false-positive urine ketone tests
  4. CBC if indicated
  1. Glu is on avg. about 500, but can be neal nl; level of glu is more reflective of degree of hypovolemia than of rate of glu production
  2. Avg. BUN on pres. 25-30 mg/dl
  3. Us. have high WBC, even if no infection present
  4. High serum [amylase] is common without corresponding GI abnormality
  5. Ketones: typical test for KA's ("Acetest" tablets or "Ketastix") uses the Nitroprusside Rxn, which only detects true ketones, i.e. AcAc-
  1. But Beta-OHB is the major species of KA present during DKA, and tends to change much more than AcAc- during development & resolution
  2. So, routine checking for serum ketones is of little use; no good for evaluating response to Tx (e.g., could be lower before Tx because mainly Beta-OHB, and rise during therapy as Beta-OHB converted to AcAc-, even though total serum [KA] is decreasing!)

IX. Other diagnostic monitoring

  1. Admit to ICU if: very young, pH <7.1, obtunded, or hemodynamically unstable
  2. Check cardiac monitor (lead II) for evidence of hyperkalemia, keep pt on it if
  1. Abnl serum [K]
  2. Administering HCO3
  3. Administering high-dose insulin
  1. A-line (optional) for continuous BP monitoring and blood sampling
  2. Foley (if obtunded) for accurate urine output measurement Q1h
  3. Labs after admit labs:
  1. Glu & pH studies Q1h initially, Q2-4h after (Q1h while on insulin drip)
  2. Electrolytes Q2-4h PRN (remember K lowest about 4h after starting Tx)
  1. Frequent anion gap determination is the best way to monitor changes in ketone (+ lactate) content
  1. Follow serum [Ca] if giving PO4 replacement
  2. Measurement of serum "ketones" of little value (see above)

X. Treatment of DKA

  1. Flow-sheets are helpful in keeping track of Tx as you go along. See Ped.Clin.N.A. 31:591, 1984
  2. Tx should begin immediately on dx; generally takes 12-24h to get back to nl metabolically
  3. Resuscitation phase
  1. Airway management:
  1. If obtunded and diminished airway reflexes, intubate and place nasogastric tube (for gastric decompression to prevent aspiration)
  2. If gastric dilatation or ileus is suspected, should get nasogastric regardless of mental status
  1. Correction of hypovolemia with isotonic solutions--will increase GFR and decrease serum glucose
  1. Estimation of vol. deficit: us. about 3-5l in adults; best to compare current & usual wt, if latter is known
  2. This stage of IVF therapy should be given as IVNS or Lactated Ringer's @ 10-20ml/kg/h; less if in ok shape & peeing well.
  3. LR generally better because:
  1. Gives less Cl which theoretically reduces risk of "hypochloremic acidosis," which occ. occurs during recovery from DKA
  2. Gives some lactate (28mEq/l)which is metabolized to HCO3 over several hours
  3. K content (4mEq/l) can help avoid hypokalemia secondary to Tx (see above); ok even in mild hyperkalemia
  4. But LR's may exacerbate severe hyperkalemia in oliguric pts
  5. Can substitute 1/2NS if get hypernatremic with IVF
  1. This IVF will decrease serum [glu] but not serum ketoacids without exogenous insulin
  1. Insulin may be delayed no more than 2h after this phase of Tx to see response of serum [glu] and [K] (see below)
  1. Post-resuscitation phase
  1. Insulin therapy:
  1. High vs. Low dose
  1. Used to give huge doses, but low-dose, descibed here, just as good in correcting hyperglycemia & acidosis,
  2. Though maybe no lower incidence of hypoglycemia or hypokalemia like initially claimed
  1. Series of 220 adult pts, 30% had 1 or more glu <50; couldn't identify risk factors in 1st 48h of hosp; after, sig. risk w/fvr, NPO status, hepatic & renal derangement (Arch Int. Med 152:2472, 1992)
  1. SOME pts, inc. those never tx'd with insulin before, will require high-dose regimen due to temporary rise in ins. resistance; suspect bact. inf. in these
  1. Endpoint of insulin Tx is correction of acidosis, not of hyperglycemia! Glu will fall to nl earlier, sometimes just with IVF
  1. If serum [glu] falls to nl and still acidotic, must supplement insulin & IVF with IV glucose (see below)
  1. Insulin Tx should start at latest 1-2h after fluid resuscitation
  2. Tx should be with regular insulin and given IV or, if no venous access, IM; SQ administration is associated with poor absorption when pt is dehydrated
  3. SQ ultra-short-acting insulin for DKA
    1. In a study in 40 pts with DKA randomized to insulin Lispro (0.3U/kg then 0.1U/kg Q1h until correction of hypoglycemia) vs. continuous IV regular insulin, there were no sig. differences between the groups in time to resolution of hyperglycemia and ketoacidosis, or total length of hospital stay. (Am. J. Med. 117:291, 2004--AFP)
  4. Bolus 0.1U/kg
  5. Drip 0.1U/kg/h
  1. Flush tubing with 50-100ml sol'n before attaching to pt to saturate insulin binding sites on walls of IV tubing
  2. IV pump is mandatory for giving insulin IV
  3. Check serum [glu] Q1h while on drip
  1. Can double insulin dose if response is inadequate (e.g. no rise in pH or slow fall in serum [glu] in 1st 3-4h of Tx)
  2.  
  3. When glu <300, add D5W or D10W to maintenance fluids (see below)
  1. rate of Dextrose administration should be calculated to match glu losses: rate of change ofserum [glu] x "glu space"
  2. "glu space" = vol. in which glu is distributed, approx. 0.3l/kg body weight
  1. At the same time, if acidosis improving, can halve the insulin drip, but be careful for recurrence of acidosis
  2. Change to SQ insulin when all the following obtain; keep IV insulin going for first few hrs after 1st SQ dose:
  1. Serum [glu] <250
  2. pH >7.3
  3. Serum [HCO3] >15
  1. For new diabetics follow ADA guidelines for initial insulin regimen
  2. Make sure to start long-acting insulin soon after stopping IV insulin, in order to avoid wild swings in blood sugar
  1. Correction of free water deficits with hypotonic IVF
  1. Us. 100-150ml/kg deficit
  2. Base replacement on this estimate + ongoing urine losses + maintenance
  3. Replace 1/2 deficit over 1st 8-12h and other 1/2 over next 16-24h (for the 2nd 1/2, don't need to monitor/replace urine output)
  4. Unlike initial fluid resuscitation with isotonic fluids to restore circulating vol., subsequent fluids should be hypotonic (1/2 NS)
  1. However, run risk of hyponatremia with the theoretical possibility of triggering idiopathic cerebral edema, so
  2. Follow serum [Na] closely and change to NS if drops
  1. Potassium replacement--ALWAYS required!
  1. If hyperkalemic, can wait 1-2h post-initiating Tx to start; otherwise start with fluid resuscitation (see above)
  2. After volume replaced, add 40mEq/l K to IVF (1/2 as KCl and 1/2 as KPO4)
  1. See above for theoretical benefits of PO4 admin; n.b. must monitor serum [Ca] & [Mg] if administering (can fall)
  2. Can give up to 60mEq/l K in documented hypokalemia, but no more thru peripheral IV
  3. If severely hypokalemic (rare), give conc. K via central line at max. 1mEq/kg/h, with continuous heart monitor
  1. HCO3--CONTROVERSIAL!
  1. Arguments in favor:
  1. Impaired myocardial performance (inc. hypotension) and response to catecholamines at pH<7.0 (both rare in kids)
  2. Impaired ventr. response to acidosis at pH <7.2
  1. Arguments against:
  1. Serum ketones will eventually be metabolized to HCO3 anyway
  2. Rapid correction of acidosis could theoretically impede O2 dissoc. from HGb and lower tissue O2 delivery
  1. No evidence of clinical harm even when occurs; seems only to occur with very large HCO3 doses
  1. Can cause paradoxical CNS acidosis due to CO2 diffusion into CNS (see "Derangements of organ systems" above)
  1. Shown to occur but unclear whether clinically significant
  1. Can cause decrease in CNS oxygenation (unknown mech; seems only to occur with very large HCO3 doses)
  2. Rapid correction of acidosis could theoretically produce hypokalemia
  3. Might accentuate Na load and increase serum osmolarity
  1. Controlled studies of HCO3 administration in DKA:
  1. No difference in clinical outcome, although hypokalemia 6x more common in group receiving HCO3
  2. No difference IV bolus vs. drip
  1. Authors of Ped. Clin N.A. review say no evidence of benefits & compelling theoretical risks, so limit HCO3 to:
  1. Symptomatic hyperkalemia
  2. Hemodynamic instability
  1. Authors of NEJM review say to use only for
  1. pH<7.0; stop administration when reach pH of 7.2

Sources: Krane, Elliot J. "Diabetic Ketoacidosis: Biochemistry, Physiology, Treatment, and Prevention" Ped. Clin. N. A. 34:935-60, Aug 1987; Foster, Daniel L. and McGarry, J. Denis, "The Metabolic Derangements and Treatment of Diabetic Ketoacidosis" NEJM 309: 159-169, Jul '83)