Evaluation of salivary ELISA for Oestradiol and Progesterone monitoring in IVF patients

Abstract Study question How do salivary E2 and P correlate with serum determinations across ovarian stimulation (OS) for IVF and artificial endometrial preparation cycles for embryo transfer? Summary answer Saliva-blood correlations were higher for E2 than for P. E2 saliva-blood ratios were patient-dependent. High serum P cases on trigger day could be identified. What is known already Steroid hormone concentrations in saliva are lower than in serum, even if the hormone is predominantly free. Therefore, quantitative assays need to be as sensitive as possible. Saliva-based assessments are a reliable and proven method for quantifying female hormone levels in premenopausal women during the natural cycle while being very precise and painless for patients. However, the cumulative experience in ovarian stimulation is relatively low. A recent study including 652 pairs of saliva-serum E2 samples reported correlations between 0.68 and 0.91, while 237 pairs of samples for P showed correlations ranging from -0.02 to 0.22 (Sakkas et al, 2020). Study design, size, duration Prospective cohort study including 300 saliva-blood pair samples from 120 patients undergoing OS for IVF (n = 90) or artificial endometrial preparation for embryo transfer (n = 30) between February and October 2022. Data were collected at each of 3 points of OS: early (days 1-5), mid (days 6-9), and trigger. Patients following endometrial preparation collected just one sample on the day of embryo transfer Participants/materials, setting, methods Thirty expected poor responders, 30 normal, and 30 expected high responders according to ovarian reserve evaluation were included. Two saliva samples per blood sample were collected: immediately after waking up, and at the clinic just before or after the blood sample was taken. Saliva samples were stored at -20ºC and shipped to Mint Diagnostics (Kent, UK) for analysis of salivary E2 and P4 Main results and the role of chance Saliva-blood E2 correlations (Spearman) were 0.68 (fasting) and 0.69 (clinic). 72% of individuals showed saliva-serum correlations > 0.8. Correlations improved when controlling for subgroup (low, normal, high responders). P4 saliva-blood correlations (Spearman) were 0.44 (fasting) and 0.29 (clinic). 41% of individuals showed correlations >0.8 on fasting samples, and 33% on clinic samples. There was a high variance of individual correlations (0-1), which were dependent on ovarian response subgroups (p < 0.05). A multivariable approach was developed to identify patients with a serum P4>1.5 ng/mL the day of trigger (n = 11) with clinic samples showing 81% sensitivity and 91% specificity and fasting samples showing 83% and 91% respectively demonstrating the possibility of shifting hormone analysis to the home rather than the clinic. Eighteen of 30 (60%) saliva samples collected on the day of ET on patients under hormonal replacement therapy showed values out of the assay range (>5000 pg/mL). Therefore, no correlations could be estimated. Limitations, reasons for caution Detection of patients with high P4 on the day of trigger was limited due to a short number of cases. No conclusions could be drawn in HRT patients for embryo transfer due to high salivary P4 values. Wider implications of the findings The current study shows that salivary monitoring of E2 and P4 is feasible. This more friendly non-invasive method could allow frequent hormone monitoring without the need for appointments at the clinic and for phlebotomy, thus resulting in more convenience for both the clinics and the patients. Trial registration number NCT05184777