Diabetic ketoacidosis: a negative urine dipstick for ketones made it unlikely.

Clinical bottom line (level 4)

  1. Urine ketone dipstick was very sensitive for diabetic ketoacidosis or diabetic ketosis (97% sensitivity - SnNout).
Hendey et al: Annals of Emergency Medicine 1997; 29: 735-738
Expires Octoboer 2003

The study

Setting: emergency department, urban teaching hospital, USA

114 patients (aged range 14 to 80 years; mean 38, ?% male) with 146 episodes of diabetic ketoacidosis (68%) or diabetic ketosis

Excluded if
  • Charts could not be traced
  • Isolated hyperglycaemia
  • Serum ketone test negative
  • Readmissions or transfers
  • Urinalysis or serum ketone not performed or delayed > 4 hours between tests



    • Independent unblinded reference standard, applied in all patients from a consecutive appropriate spectrum.
    Reference standard:
    • all of:
      • serum glucose > 14mmol./l or a history of diabetes
      • metabolic acidosis with corrected pH < 7.30 or serum bicarbonate < 15mM
      • ketonaemia, positive at 1:2 dilution or above
      • diabetic ketosis: as DKA but without the acidosis
    Diagnostic test: urine dip test with Multistix - positive test result if a trace detected

    The evidence


    diagnostic test number of patients sensitivity for
    DKA and DK
    (95% CI)    		 LR+ LR-
    urine dip test 142 97%
    (95% to 99%)
    total 146


    diagnostic test number of patients sensitivity for
    DKA
    (95% CI)    		 LR+ LR-
    urine dip test 96 97%
    (94% to 100%)
    total 99


    diagnostic test number of patients sensitivity for
    DK
    (95% CI)    		 LR+ LR-
    urine test dip 46 98%
    (94% to 100%)
    total 47

    Comments

    1. 19% had no prior diagnosis of diabetes mellitus.
    2. Authors do not give details of patients who had dip test (and possibly ketonuria) but who did not meet the diagnostic criteria for DK or DKA. It is not possible to calculate pre or post test probabilities (or likelihood ratios) since specificity cannot be estimated.
    3. Since staff may have been using negative urine dip test to rule out DKA in their clinical practice there is the possibility that some patients could have had undiagnosed DKA and negative urine dip tests. The reported sensitivity will therefore be an overestimate.
    4. Inclusion criteria did not mention signs of symptoms of diabetic patients who presented to the ER dept, so makes it difficult to know how far this study can be generalised.

    Citation

    1. Hendey GW, Schwab T, Soliz T: Urine ketone dip test as a screen for ketonaemia in diabetic ketoacidosis and ketosis in the emergency department. Annals of Emergency Medicine 1997; 29: 735-738
    Contributor: Richard Hardem and Chris Ball, October 2000
    Reviewer: Wai-Lam Chan

    Clinical Question.
    Patient diabetic ketoacidosis
    Intervention or Exposure urine ketone dipstick
    Outcome diagnosis