Skip to main content
. 2020 Jun 29;20:204. doi: 10.1186/s12876-020-01352-6

Table 3.

Accuracy of diagnostic measures and substrates for lactose malabsorption

Measurea and substrateb AUCc 95% CI p valued Cut-offe Sensitivity(%) Specificity (%)
Breath H2
 Lactose 1.000 1 < 0.001 79 ppm 100 100
 CON 0.942 0.828–1 < 0.001 21 ppm 92 100
 A2M 0.994 0.981–1 < 0.001 13 ppm 92 100
Plasma glucose
 Lactose 0.755a 0.603–0.907 0.008 1.77 mmol/L 61 100
 CON 0.535b 0.271–0.645 0.671 NA NA NA
 A2M 0.458b 0.348–0.723 0.712 NA NA NA
Urinary galactose/creatinine ratio
 Lactose 0.758 0.611–0.905 0.008 0.03 mg/mg 53 100
 CON 0.653 0.480–0.838 0.112 NA NA NA
 A2M 0.618 0.433–0.803 0.223 NA NA NA

aMaximal increase in breath H2 and plasma glucose over baseline for 3 h, urinary galactose/creatinine ratio at 3 h post lactose and milk challenge

b Lactose: 50 g lactose in 250 mL water; CON and A2M: 750 mL milk

c Values are area under the receiver operating characteristics (ROC) curve (AUC)

Different letters (a, b) represent a significant difference in the AUCs of the ROC curve between substrates calculated by ROC test

dp value between calculated ROC AUC and AUC of 0.5 (no discrimination between true positives (TP; LNP) and true negatives (TN; LP)) was computed using the Mann-Whitney U test

e Cut offs (calculated by the Youden Index) are presented, unless AUC > 0.7 and P value < 0.05

A2M a2 milk, CON conventional milk, CI confidence interval