(685) Mid-pregnancy quantification of sub-clinical inflammation by blood-based and vaginal-derived cytokines and the prediction of PTB

PTB risk assessment is imprecise, relying largely on clinical factors. Whether early pregnancy inflammatory biomarkers improve PTB prediction in asymptomatic patients is unknown.


Study Design:

Primary analysis of a prospective, longitudinal cohort enriched for spontaneous PTB risk. Asymptomatic multiparous patients with singleton gestations were recruited < 22 wks. Up to 2 blood and 2 vag samples per patient were collected 12-27+6 wks. 12 cytokines per sample were quantified via Luminex. We evaluated blood and vag cytokines collected early (12-19+6), later (20-27+6), and overall (12-27+6); we analyzed the earliest sample collected per GA epoch for each patient. Outcomes: primary = PTB < 35 wks; Secondary = delivery GA, PTB < 32 wks. Cytokines were classified as ‘high’ or ‘low’ using cut points optimizing sensitivity & specificity for PTB < 35 wks. A pro-inflammatory cytokine score was calculated per sample: +1 point per high pro-inflammatory (IL-1𝛂, IL-1β, IL-2, IL-6, IL-8, GM-CSF, TNF-𝛂, VEGF) & +1 point per low anti-inflammatory (IL-1RA, IL-4, IL-10, G-CSF) cytokine. Logistic regression, Kaplan-Meier, and AUC evaluated clinical, blood, vag, and blood+vag cytokine models.


Results:

437 patients were included; 15% delivered < 35 wks, 8% < 32 wks; median delivery GA = 38.1 wks (IQR 36.7, 39.3). Blood (n=200 early, n=239 later) was collected at a median 19.3 (IQR 16.3, 21.3) wks overall; blood cytokine scores = higher in those with PTB < 35 (mean 3.7 ± 1.5) and < 32 (mean 3.9 ± 1.4) vs. ≥ 37 wks (mean 3.2 ± 1.1), p< 0.001. Vag samples (n=314 early, n=255 later) were collected at a median 16.7 (IQR 15.9-19.1; n=375) wks; vag cytokine scores did not differ by outcome, p >0.05. Blood & blood+vag models improved PTB prediction vs. clinical-only models (Table) and were associated with delivery GA (Figure).


Conclusion:

Subclinical systemic inflammation in maternal blood, particularly < 20 weeks, is associated with PTB. Adding blood-based cytokines to clinical data modestly improves PTB prediction and risk stratification.