Urine dipsticks are not accurate for detecting mild ketosis during a severely energy restricted diet.
Level 3 - non-randomized controlled study
Prospective repeated-measures diagnostic accuracy study in a single non-randomized cohort
PubMed 33082996 · doi:10.1002/osp4.432
What was done
Diagnostic accuracy study evaluating urine dipsticks (Ketostix, measuring acetoacetate) against a portable blood ketone monitor (FreeStyle Optium, measuring β-hydroxybutyrate) as the reference standard. 263 paired fasting urine and blood samples were collected from 50 women (mean age 58.0 ± 4.3 years, BMI 34.3 ± 2.4 kg/m²) before, during, and up to 10 weeks after a 16-week total meal replacement severe energy-restricted diet. Sensitivity, specificity, and area under the receiver operating characteristic (ROC) curves were calculated across three blood β-hydroxybutyrate thresholds: ≥0.3 mM, ≥0.5 mM, and ≥1.0 mM.
What was found
At blood β-hydroxybutyrate thresholds of ≥0.3 mM, ≥0.5 mM, and ≥1.0 mM: - Sensitivity of urine dipsticks was 35%, 52%, and 76%, respectively (failing to detect ketosis in 65%, 48%, and 24% of instances). - Specificity was 100%, 97%, and 78%, respectively. - Area under the ROC curves was 0.67, 0.74, and 0.77, respectively.
Why it matters
Urine dipsticks lack sufficient sensitivity to reliably detect the mild ketosis induced by severe energy restriction. When clinical monitoring of dietary adherence via ketosis is required, blood ketone monitors should be used instead.
Limits
The study included only 50 older women with obesity, limiting generalizability to men, younger adults, or other dietary contexts. The 263 sample pairs were repeated longitudinal measures from the same cohort rather than independent subjects. The reference method used was a handheld blood ketone monitor rather than a formal laboratory reference standard.
Cited by
- supports Urine ketone strips are not very accurate and provide only semi-quantitative measurement of ketosis.