4. Discussion
The present study is the first to describe the levels of PCSK9 circulating in women during early pregnancy (first trimester; 134 women), demonstrating how air pollutants impact PCSK9 and its association with birth weight and gestational age at birth. In this study population, mean PCSK9 levels were lower than those previously described in non-pregnant women (27, 28). We observed an association between exposure to PM (in particular PM10) and NO2 and increased levels of PCSK9. PCSK9 levels were also associated to a decrease of gestational age at delivery and to an increased probability of urgent c-sections at delivery.
A previous study evaluated PCSK9 levels in pregnancy characterized by intrauterine growth restriction (IUGR), with lipid/lipoprotein levels being available for 70 patients vs 102 controls during late gestation. This previous study confirmed that the immunological expression of PCSK9 could be detected in the placenta, as well as in fetal and maternal plasma. Plasma PCSK9 levels were at a higher range to those reported in the present study (29). PCSK9 levels were also reported in a study with a small sample size that compared normal pregnant women (n = 6), diabetics (n = 6), and overweight/obese women that were diabetic (n = 10) at term (30). The latter two groups had significantly reduced PCSK9 levels, with raised LDL receptor activity and reduced LDL-C levels. The authors attributed these findings to the maternal inflammatory status with raised placental cytokines, which is widely demonstrated in diabetic pregnancies (31). However, PM10 and PM2.5 exposure has been associated with acute placental inflammation; thus, these pollutants might contribute to adverse pregnancy outcomes (32, 33). A clear mediator of these effects has not been reported yet. Inflammation and dyslipidemia early in pregnancy are independently associated with an increased risk of pre-term birth. However, this risk might be elevated when both conditions are present before 21 weeks of gestation (34). Recently, the ABCD (Amsterdam Born Children and Their Development) study showed that atherogenic lipid profiles during the first trimester confer an increased risk of adverse pregnancy outcomes, including maternal morbidity, mortality, and preterm delivery. TG levels, but not total cholesterol levels, during the first third of pregnancy were independently and positively associated with adverse pregnancy outcomes for both mother and newborn (35). Thus, considering that none of our child-bearing women were either dyslipidemic or pathologically inflamed, the present report presents a highly novel finding on the association between plasma PCKS9 elevation and time of delivery. Overall, the time of delivery was estimated to advance by between 0.8 and 1.8 weeks for every 100 ng/mL increment in PCSK9 levels. Moreover, the observation that the risk of experiencing an urgent cesarean section is significantly higher (OR = 2.99) in women with higher levels of PCSK9 indicates a general condition that is not perceived during pregnancy, but is related to peripartum risk.
Relative to circulating PCSK9 levels, the present report appears to provide results that are in line with the physiological rise of estradiol (E2) levels during pregnancy. In other words, maternal E2 levels progressively rise from 367 pM (luteal phase) to between 11,000 and 37,000 pM at the end of pregnancy (36). At approximately nine weeks of gestation, a hormonal ovary-to-placenta shift occurs, resulting in direct E2 placental production (36). Retrospective studies reported that, compared to pre-menopausal women, post-menopausal women with low endogenous estrogens have higher PCSK9 levels (27, 37, 38). Conversely, after in vitro fertilization, the stimulation of endogenous estrogens markedly reduces PCSK9 levels (38).
The lack of an association between PCSK9 levels and newborn features (e.g., crown-rump length, nuchal translucency, fetal heart rate, and ductus venosus pulsatility index) might be of interest in the still questioned field of lipid management during pregnancy. This phenomenon is particularly relevant in the debate on whether implications exist for both the mother and unborn child. Also, knowledge remains limited on the safety of newer agents, such as PCSK9-inhibitory therapy with evolocumab or alirocumab. Of note, PCSK9 was initially identified as NARC-1 (neural apoptosis-regulated convertase-1), which was implicated in the differentiation of cortical neurons (39), and has been recently described as a possible modulator of brain cholesterol homeostasis (21). Studies evaluating the safety of the PCSK9 inhibitor evolocumab confirmed placental transfer to the infant (40); however, at present information remains unclear regarding its safety in pregnancy and lactation. This issue might be particularly pertinent for women with severe forms of familial hypercholesterolemia (FH), where this treatment would be required (41). Guidance on the care of familial hypercholesterolemia currently recommends that the use of statin/ezetimibe/niacin is stopped at least four weeks (preferably 12 weeks) before conception, and should not be used during pregnancy or lactation (42). The ongoing NCT 02957604 trial is enrolling 375 pregnant women with hypercholesterolemia, with and without atherosclerotic CV disease, who will be exposed to evolocumab for the whole pregnancy. These women will be followed to delivery or abortion and breastfeeding, if feasible. In addition, infants will be followed for five years post-partum.
CV risk factors in pregnancy have been evaluated by many investigators, indicating that in a large number of pregnancies are associated with preexisting CV risk factors, (34, 43-45). Thus, in addition to evaluating lipoprotein cholesterol concentrations, it was of interest to evaluate a key regulator of cholesterolemia (i.e., PCSK9 levels) in the current study. This protein fosters the catabolism of the LDL receptor, thus increasing cholesterolemia (35). The highly significant correlation between PCSK9 levels and LDL-cholesterolemia is of interest, since lipid levels during early pregnancy could be used to identify women at risk for hypertension and future CVD (34). A large meta-analysis by Sun et al. (46) reported a significant association between PM2.5 (not PM10) exposure and hypertension development, albeit to a modest degree. Ambient exposure appeared to be more strongly connected to the development of hypertension during the first and third trimesters. In our cohort, PCSK9 was not correlated with hypertension.
Moreover, complications occurring during pregnancy indicate a future increased risk for atherosclerotic disease (47). Epidemiologic data consistently showed the early onset cardiovascular disease in women who experienced pregnancy loss, preterm pregnancy, or pregnancy complicated by intrauterine growth restrictions. These phenomena possibly arise as a result of metabolic, endothelial, and inflammatory changes during complicated pregnancies. Based on these epidemiological observations, the American Heart Association recognizes pregnancy complications as independent risk factors for future CVD.
PM inhalation is an established trigger of CV events (48). Such events might occur within hours or days of exposure. Short-term exposure to PM pollution contributes to acute CV morbidity and mortality. In particular, long-term exposure to elevated PM levels is associated with a reduced life expectancy (49). A previous report on an obese population showed that 12- and six-month exposure to PM10 is associated with a significant rise in circulating PCSK9 levels, which are positively associated with the Framingham Risk Score (20). The results of the current study on pregnant women support the finding that ambient pollutants (i.e. PM10, PM2.5, and NO2) are associated with raised PCSK9 levels. Consequently, the current study provides a further insight into the potential association of CV risk variables with pollutants. A meta-analysis of 21.09 million participants showed that each 10 μg/m3 rise in PM2.5 corresponds to an increased relative risk (RR) of total CVD events (RR 1.12, 95% CI 1.05–1.19), CVD incidence (RR 1.12, 95% CI 1.05–1.19), and CVD mortality (RR 1.11, 95% CI 1.08–1.14). A more robust association was obtained with NO2, whereby every 10 μg/m3 increase in NO2 led to a higher risk of total CVD events (RR 1.36, 95% CI 1.09–1.64) and CVD mortality (RR 1.46, 95% CI 1.13–1.79) (50).