BackgroundPeripheral and uterine natural killer cells (pNK and uNK cells) are key players in the establishment and maintenance of pregnancy and are disturbed in patients with recurrent miscarriage (RM). Different immunologic risk factors have been proposed between patients with primary RM (pRM, no previous live birth) and secondary RM (sRM, ≥ 1 previous live birth). However, so far, the study populations mainly consisted of small subgroups. Therefore, we aimed to analyse pNK and uNK cells in a large, well defined study population within a prospective study.MethodsIn total, n = 575 RM patients (n = 393 pRM, n = 182 sRM) were screened according to a standard protocol for established risk factors as well as pNK and uNK cells. Peripheral blood levels of CD45+CD3−CD56+CD16+ NK cells were determined by flow cytometry and uterine CD56+ NK cells by immunohistochemistry in mid-luteal non-pregnant RM patients. Exclusion of patients with ≥1 established risk factor revealed n = 248 idiopathic RM patients (iRM, n = 167 primary iRM (ipRM), n = 81 secondary iRM (isRM)).ResultsPatients with pRM and ipRM showed significant higher absolute numbers and percentages of pNK cells compared to sRM and isRM patients (pRM/ipRM vs sRM/isRM, mean ± SD /μl: 239.1 ± 118.7/244.9 ± 112.9 vs 205.1 ± 107.9/206.0 ± 105.6, p = 0.004/ p = 0.009; mean ± SD %: 12.4 ± 5.5/12.8 ± 5.4 vs 11.1 ± 4.6/11.1 ± 4.3, p = 0.001; p = 0.002). Only patients with isRM showed significantly higher uNK levels compared to patients with ipRM (mean ± SD /mm2 288.4 ± 239.3 vs 218.2 ± 184.5, p = 0.044).ConclusionsThe demonstrated differences in pNK and uNK cells in RM patients depending on previous live birth might indicate differences in NK cell recruitment and potentially different underlying immune disorders between pRM and sRM. As there is an overlap in the distribution of the NK cell results, further studies with focus on NK cell function are needed in order to clearly identify RM patients with distinct immune abnormalities. The clinical relevance of our findings should be interpreted cautiously until specificity and sensitivity are further evaluated.