Variability of biomarkers used for the classification of metabolic syndrome: A repeated measurements study.
Level 3 - non-randomized controlled study
Prospective longitudinal repeated-measures cohort study
PubMed 35469729 · doi:10.1016/j.numecd.2022.03.022
What was done
Twenty-five employees of University Medicine Greifswald aged 22–70 years were evaluated once a month for 12 months. Monthly assessments included anthropometric measurements, blood pressure, and non-fasting blood draws measuring glucose, HDL cholesterol, LDL cholesterol, and triglycerides. Metabolic syndrome was classified using International Diabetes Federation criteria adapted for non-fasting samples. Intra-individual variability was evaluated using coefficients of variation (CV), intra-class correlation coefficients (ICC), and Cohen's kappa.
What was found
Eight participants (32%) met criteria for metabolic syndrome at least once during the 12-month period, but none met the diagnostic criteria consistently across all monthly follow-ups. Diagnostic agreement across visits showed a Cohen's kappa of 0.57. CV was highest for triglycerides (27.5%), followed by glucose (10.1%), LDL cholesterol (9.5%), and HDL cholesterol (8.6%). ICCs were lowest for glucose (0.51), triglycerides (0.65), systolic blood pressure (0.68), and diastolic blood pressure (0.69).
Why it matters
Because metabolic syndrome biomarkers fluctuate substantially month-to-month, single-point measurements lead to significant diagnostic misclassification. Epidemiological studies and clinical assessments may misestimate prevalence or individual risk without accounting for intra-individual variability.
Limits
The study had a very small sample size (n = 25) limited to employees at a single medical institution. Measurements relied on non-fasting blood samples, which likely increased variance in glucose and lipid markers compared to standard fasting protocols. Clinical outcomes associated with fluctuating diagnoses were not assessed.
Cited by
- supports Triglyceride levels exhibit substantially greater day-to-day variability and measurement instability over time than LDL cholesterol and HDL cholesterol.