Results

A total of 612 women participated in the study. The ethnic distribution was 525 Caucasian, 50 South-Asian, 27 Hispanic, and 10 Sub-Saharan African women. We collected 528 fasting samples and 377 postprandial samples. In 293 patients both samples were taken, exceeding the minimum suggested 120 samples for determining reference intervals and confidence intervals according to the Clinical Laboratory and Standards Institute31.
In our analysis, we found that TSBA concentrations for both groups of samples were not normally distributed, but rather had a positive skewness distribution, in accordance with other studies26,28 (figure 1-2). We found a median of 7.6 µmol/L for the fasting TSBA with an upper reference limit (97.5th percentile) of 14.1 µmol/L (95% CI 12.7-15.5 µmol/L). We found a median of 9.1 µmol/L in the postprandial samples with an upper reference limit of 20.2 µmol/L (CI 95% 17.3-32.3 µmol/L). We established reference intervals of 4.4-14.1 µmol/L for fasting TSBA, and 4.7-20.2 µmol/L for postprandial TSBA levels (table 1).
When applying the currently used thresholds to our normal asymptomatic pregnant population, we found that TSBA levels exceeded 10 µmol/L, the most commonly used threshold, in 15.7% of the fasting and in 38.5 % of postprandial measurements. The fasting threshold ≥6 µmol/L was present in 83.3% of the pregnant population, and postprandial ≥15 µmol/L was present in 6.9% of the measurements.
When both measurements were performed (293 patients), the postprandial values were significantly higher than the fasting measurement, with a mean increase of 1.77 µmol/L (22%). On correlation analysis, fasting TSBA levels were moderately correlated with postprandial TSBA levels (Pearson’s coefficient 0.44, P value < 0.0001) (table 2-3).
We observed higher fasting TSBA levels in pregnancies obtained with the use of assisted reproductive technology (ART) compared to spontaneous conception (8.59 µmol/L and 7.98 µmol/L, respectively, P value 0.02). The reasons for the association between ART and higher TSBA levels are unclear, but they could be related to some metabolic disturbances regarding infertility itself33, or due to hormonal maintenance therapy used for ART. We also noticed higher fasting TSBA levels in nulliparous compared to multiparous (mean 8.6 µmol/L and 7.75 µmol/L, respectively, P value 0.04), although the reason for this association is unclear.
Interestingly, we also found a correlation between fasting TSBA and fetal gender, with the male gender being associated with higher levels of TSBA (8.47 µmol/L versus 7.6 µmol/L, P value <0.001). To our knowledge, this is the first time a correlation with fetal gender has been made and can be explained by differences in hormonal metabolism by the fetal liver. The magnitude of the observed differences, however, may not be clinically relevant.
Fasting TSBA levels were not correlated with maternal age, BMI, weight gain and neonatal birth weight, and no association was found with gestational diabetes mellitus, hyperthyroidism, hypertensive disorders, Intrauterine growth restriction (IUGR) or progesterone therapy (table 2-3).
We found that postprandial TSBA values were significantly lower in pregnancies complicated by IUGR (median 8.59 µmol/L versus 9.98 µmol/L in non-IUGR fetuses, P value 0.03), although this difference is difficult to explain and may not be clinically relevant. Postprandial values were not associated with any of the other variables examined (table 2-3).
We noticed a seasonal pattern for both fasting and postprandial TSBA levels, with highest values in the winter season, a decline during spring and summer, and minimum values in the autumn (p values <0.01 and 0.02, respectively) (figure 3-4).
ICP is known to be more common in South America and in Northern Europe5. Nevertheless, we did not find a strong correlation with ethnicity, probably due to the fact that the majority of patients in our study was of Caucasian origin.