Table 3

Summary of studies investigating predictivity of dehydration severity of laboratory investigations

AuthorStudy settingStudy designInclusion criteriaExclusion criteriaSample sizeLevel of evidence
(Oxford Centre for Evidence Based Medicine)
Key results
Hoxha et al 1 urban hospital
Prospective cohort studyChildren aged 1 month–5 years admitted with diarrhoea and/or vomiting during the 2-year study period200 children2In severe dehydration, creatinine levels (mmol/L) were significantly higher compared with other dehydration degrees (61.65±34.97 vs 41.16±7.49 in mild dehydration, 41.27±10.16 in moderate dehydration)
The venous bicarbonate levels (mmol/L) in non-dehydrated patients were 21.1±2.78, in mild dehydration 19.09±2.88, whereas in moderate and severe (>10% wt loss) dehydration, 16.31±3.16 and 12.18±3.78, respectively (p<0.001).
The base excess (BE) in severe dehydration (−18.96) compared with none (−5.9), mild (−8.57) and moderate dehydration (−12.26) decreased significantly (p<0.001).
Tam et al 1 tertiary children’s ED
Case comparison trialChildren aged <18 years with diarrhoea and vomiting who clinically required intravenous fluids for rehydration compared with minor trauma patients who required intravenous needling for conscious sedation73 cases and 143 controls2The following parameters were statistically significant (p<0.05) between the control group and the dehydrated group:
  • Urine sodium/potassium ratio (2.3 (0–56) vs 0.69 (0–4.4)).

  • Urine sodium, fractional sodium excretion (%) (0.52 (0–10.4) vs 0.19 (0–0.89)).

  • Serum bicarbonate (mmol/L) (24 (18–30) vs 20 (10–27)).

The best markers for dehydration were urine Na<90 mmol/L and serum bicarbonate<21 mmol/L (area under receiver operating characteristic curve=0.798 and 0.821, respectively; sensitivity=75% and 90%, respectively; specificity=74% and 62%, respectively)
Narchi et al 1 urban hospital
(Saudi Arabia)
Prospective cohort studyChildren aged 1 month–5 years, diarrhoea (three or more watery stools) or vomiting for less than
72 hours at presentation and dehydration
Extraintestinal infection; renal, respiratory, cardiac, neurological, or endocrine anomalies; chronic diarrhoea; malnutrition; failure to thrive; and administration of diuretics or intravenous fluids within 1 week of presentation116 children2Reduced serum bicarbonate (<22 mmol/L) occurred in 72 patients (68%), with a significant difference between the three groups (39 in mild, 23 in moderate and 10 in severe dehydration group; p=0.009).
When the concentration of bicarbonate was studied by age group, it was found to be overall significantly lower in the younger (age <1 year) group (p=0.01).
Steiner et al
  • 5 EDs

  • 6 hospitals

  • 2 gastroenteritis clinics

Meta-analysisTwenty-six reviewed studies contained original data on the precision or accuracy of findings for the diagnosis of dehydration in young children aged 1 month to 5 years.Eliminated 13 of the
26 studies because of the lack of an accepted diagnostic standard or other limitation in study design.
1246 children1
  • Blood urea nitrogen (BUN)>45 mg/dL=LR of dehydration (for at least 5% dehydration) 46.1 (95% CI 2.9 to 733).

  • Serum bicarbonate>15 mEq/L=LR of 0.18 (95% CI 0.08 to 0.37).

  • Serum bicarbonate<17 mEq/L=LR of 3.5 (95% CI 2.1 to 5.8).

  • pH<7.35=LR 2.2 (95% CI 1.2 to 4.1).

  • Anion gap>20 mmol/L=LR 1.8 (95% CI 0.8 to 4.2).

  • Uric acid>600 mmol/L=LR 1.0 (95% CI 0.3 to 3.5).

Teach et al 1 tertiary children’s ED
Cohort studyChildren aged 2 weeks–12 years with acute (<1 week) dehydrationDiabetes mellitus, diabetes insipidus, known renal insufficiency of any aetiology57 children2The following laboratory parameters were significantly correlated in simple linear regression models to fluid deficit:
  • BUN/creatinine ratio (BUN/Cr)=r 0.52 p=0.0005.

  • Uric acid=r 0.35 p=0.03.

Serum BUN/Cr>20: sensitivity 92.3%, specificity 33.3%, PPV 40%, NPV 90%
Serum BUN/Cr>30: sensitivity 61.5%, specificity 70.4%, PPV 50%, NPV 79.2%
Serum BUN/Cr>40: sensitivity 23.1%, specificity 88.8%, PPV 50%, NPV 70.1%
Serum uric acid >300µmol/L sensitivity 84.6%, specificity 40.7%, PPV 40.7%, NPV 84.6%
Serum uric acid>450µmol/L: sensitivity 30.1%, specificity 59.2%, PPV 26.7%, NPV 64%
Serum uric acid >600µmol/L: sensitivity 46.1%, specificity 77.8%, PPV 33.3%, NPV 67.7%
Yilmaz et al 1 tertiary paediatric department and 1 tertiary children’s ED
Retrospective studyChildren aged 1–21 months with acute gastroenteritis and dehydration, and treated with intravenous fluid therapyPatients younger than 1 month or older than 24 months and those with additional health problems, such as malnutrition, urinary tract infections and septicaemia1683At multiple linear regression analysis, dehydration severity correlated strongly to urea (p<0.001) and bicarbonate (p=0.01), but no to sodium (p=0.28).
Serum bicarbonate concentrations of 15 mmol/L or more exclude a severe dehydration (PPV=89.6%, NPV=28%).
Vega et al 1 tertiary children’s ED
ProspectiveChildren aged 2 weeks–15 years who have required intravenous fluid for acute dehydration972Serum bicarbonate level of <17 mEq/L was 77% sensitive for moderate dehydration and 94% sensitive for severe dehydration. When clinical impression was combined with a bicarbonate concentration of <17 mEq/L, sensitivity for prediction of severe dehydration increased to 100%.
Mackenzie et al 1 children’s hospital (Australia)Prospective cohort studyChildren under 4 years admitted with gastroenteritis102 children2The laboratory findings that pointed to dehydration of 4% or more, in a statistically significant way, were
  • Serum urea>6 mmol/L=PPV 0.63, p<0.001.

  • pH<7.35=VPP 0.62, p=0.024.

  • Base deficit≥7= PPV 0.5, p=0.103.

Increasing urea levels were associated with a higher level of dehydration (p=0.505, p<0.001). Reduced blood pH was associated with a higher level of dehydration (p=−0.453, p<0.001). Increasing base deficit was associated with a higher level of dehydration (p=−0.378, p<0.001).
Steinert et al 1 children’s ED
Prospective cohort studyChildren aged 3–36 months with gastroenteritis, clinically suspected moderate dehydration, need for intravenous rehydration79 children2Urine-specific gravity (r=−0.06, p=0.64), urine ketones (r=0.08, p=0.52) and urine output during rehydration (r=0.01, p=0.96) did not correlate with the initial degree of dehydration.
  • ED, emergency department; LR, likelihood ratio; NPV, negative predictive value; PPV, positive predictive value.