Abstract
In this edition, Huri et al. (BJOG 2022) have added to the growing literature defining pregnancy-specific reference ranges for total serum bile acid (TSBA) concentrations.
Reference intervals in clinical pathology are typically calculated using an ‘indirect resource’, such as stored laboratory samples. Results are partitioned into biologically relevant groups (typically age and sex), outliers are excluded to remove anomalous results due to disease, and the central 95% of the population is calculated for each group. Two large studies, that of Huri et al. and our own (Mitchell et al. BJOG 2021;128:1635–44), have recently calculated reference intervals for non-fasting TSBA concentrations in the third trimester of pregnancy, finding strikingly similar results for the upper limit (20.2 and 18.3 μmol/l, respectively). Both studies excluded samples with cholestatic pathology before analysis, treating the cohorts as a ‘direct resource’ and obviating exclusion of outliers.
However, the results demonstrate outliers within each dataset. In the study by Huri et al. this may have been the result of the selected participants having other gestational diseases (for example, gestational diabetes) and originating from samples taken during hospital admission; the reason for women being inpatients may have influenced serum bile acid concentrations and therefore confounded the findings. To assess the impact of excluding outliers, we used the block D/R (Zellner et al. arXiv preprint; 1907.09637.) procedure to identify outliers in our dataset (from outpatient samples routinely taken and limited to uncomplicated pregnancies), and reanalysis revealed a slight reduction in the upper limit of the reference interval and narrowing of the confidence interval (Table 1).
Reference intervals in clinical pathology are typically calculated using an ‘indirect resource’, such as stored laboratory samples. Results are partitioned into biologically relevant groups (typically age and sex), outliers are excluded to remove anomalous results due to disease, and the central 95% of the population is calculated for each group. Two large studies, that of Huri et al. and our own (Mitchell et al. BJOG 2021;128:1635–44), have recently calculated reference intervals for non-fasting TSBA concentrations in the third trimester of pregnancy, finding strikingly similar results for the upper limit (20.2 and 18.3 μmol/l, respectively). Both studies excluded samples with cholestatic pathology before analysis, treating the cohorts as a ‘direct resource’ and obviating exclusion of outliers.
However, the results demonstrate outliers within each dataset. In the study by Huri et al. this may have been the result of the selected participants having other gestational diseases (for example, gestational diabetes) and originating from samples taken during hospital admission; the reason for women being inpatients may have influenced serum bile acid concentrations and therefore confounded the findings. To assess the impact of excluding outliers, we used the block D/R (Zellner et al. arXiv preprint; 1907.09637.) procedure to identify outliers in our dataset (from outpatient samples routinely taken and limited to uncomplicated pregnancies), and reanalysis revealed a slight reduction in the upper limit of the reference interval and narrowing of the confidence interval (Table 1).
| Original language | English |
|---|---|
| Pages (from-to) | 1897-1898 |
| Number of pages | 2 |
| Journal | BJOG: An International Journal of Obstetrics and Gynaecology |
| Volume | 129 |
| Issue number | 11 |
| Early online date | 15 Apr 2022 |
| DOIs | |
| Publication status | Published - Oct 2022 |