Estimating habitual iodine intake and prevalence of inadequacy from spot urine in cross-sectional studies: a modeling analysis to determine the required sample size.
Arns-Glaser L., Zihlmann R., Gessler S., Verkaik-Kloosterman J., Zandberg L., Assey VD., Rigutto-Farebrother J., Braegger CP., Zimmermann MB., Andersson M.
BACKGROUND: The habitual/usual iodine intake and the prevalence of iodine inadequacy may be estimated from spot urinary iodine concentrations in cross-sectional studies by collecting a repeat spot urine in a subgroup of the study population and accounting for within-person variability in iodine intake. However, guidance on the required overall sample size (N) and the replicate rate (n) is lacking. OBJECTIVES: To determine the sample size (N) and replicate rate (n) needed to estimate the prevalence of iodine inadequacy in cross-sectional studies. METHODS: We used data from local observational studies conducted in women 17-49 y old in Switzerland (N = 308), South Africa (N = 154), and Tanzania (N = 190). All participants collected 2 spot urine samples. We calculated the iodine intake using urinary iodine concentrations and accounted for urine volume using urinary creatinine concentration. For each study population, we estimated the habitual iodine intake distribution and determined the prevalence of iodine intake below the average requirement using the Statistical Program to Assess habitual Dietary Exposure (SPADE). We used the obtained model parameters in power analyzes and estimated the prevalence of iodine inadequacy for different sample sizes (N = 400, 600, and 900) and replicate rates (n = 50, 100, 200, 400, 600, and 900). RESULTS: The estimated prevalence (95% CI) of inadequate iodine intake was 21% (15, 28%), 5.1% (1.3, 8.7%), and 8.2% (3.4, 13%) for Swiss, South African, and Tanzanian women, respectively. An N of 400 women, with a repeated measure (n) in 100 women, achieved a satisfactory precision of the prevalence estimate in all study populations. Increasing the replicate rate (n) improved the precision more effectively than increasing the N of the study. CONCLUSIONS: The sample size for cross-sectional studies aiming to assess the prevalence of inadequate iodine intake depend on the expected prevalence, the overall variance in intake, and the study design. However, an N of 400 participants with a repeated measure of 25% may be used as guidance when planning observational studies applying simple random sampling. This trial was registered at clinicaltrials.gov as NCT03731312.