Bone status of children born from mothers with autoimmune diseases treated during pregnancy with prednisone and/or low molecular weight heparin

Pregnancy is characterized by the combination of a procoagulant state, meant to avoid massive maternal bleeding during delivery, and venous stasis caused by venous dilation and compression by the gravid uterus [1]. Thus, pregnant women are at higher risk of developing thromboembolic complications; this risk is further increased among women suffering from prothrombotic conditions such as systemic autoimmune disorders [2]. Among patients affected with connective tissue diseases, thrombosis is a major issue in women with systemic lupus erythematosus (SLE), and in those with antiphospholipid syndrome (APL). Thus thromboprophylaxis with low-molecular-weight heparins (LMWH) is advised, although the potential fetal and maternal toxicity of antithrombotic drugs and their impact on delivery still needs to be fully elucidated. Concomitant treatment with low dose aspirin is generally also recommended [3]. Moreover, in patients with connective tissue diseases, corticosteroids such as prednisone and prednisolone are usually administered in order to control maternal disease flares.

Bone health is influenced by genetic and environmental factors, some of which concern early childhood or even prenatal life [4‐6]. In this regard, chronic maternal disease and maternal treatments could potentially have an impact on bone status of their offspring.

Both long-term LMWH therapy and corticosteroids may be associated with osteopenia in mothers, although calcium and vitamin D supplementation seems to reduce the risk of drug-induced osteoporosis during pregnancy [7‐11]. Although it has been shown that LMWHs do not seem to cross the placenta [12] and that they seem to be safe for the fetus, and although it is known that prednisone and prednisolone are mostly inactivated by placental hydroxilase, their effects on bone status in children born from mothers treated with these drugs during pregnancy are currently unknown.

The aim of our study was to evaluate bone status in a cohort of children born from mothers affected by systemic autoimmune disorders and treated during pregnancy with agents (prednisone and heparin) known to reduce bone mass.

We enrolled 27 girls and 14 boys (mean age at clinic visit 5 years and 10 months, range 9 months- 12 years), born from 31 mothers with systemic autoimmune diseases (there were 9 enrolled mothers who had two pregnancies during which prednisone and/or LMWH were administered).

All mothers had been continuously treated during all pregnancies with daily LMWH in 10 cases, prednisone in 15 cases, or both in 15 cases. There were 11 preterm deliveries (gestational age < 37 weeks), in 7 women affected by SLE, 3 by primary antiphospholipid syndrome (PAPS) and one by granulomatosis with polyangitis; fetal distress was reported in 4 cases. Median birth weight was 2935 g, range 520–3790. Eight newborns had neonatal complications: respiratory distress (n = 3), jaundice (n = 3), transient hypocalcemia and hypoxic-ischemic syndrome (n = 1 each). Breastfeeding for at least 6 months was reported in 12 cases; 37 children had received vitamin D supplementation (400 IU/day; 6/37 for 6 months, 31/37 for the first year of life), and growth and development were within normal limits. No history of fractures in mothers or children was recorded.

In all children clinical examination was within normal limits for age. Of note, 2 patients had alterations on primary teeth, with cavities and enamel abnormalities, but without any damage on permanent teeth; one of these babies was born from a mother with SLE treated with LMWH and prednisone and the other one, with neonatal transient hypocalcemia, was born from a woman with granulomatosis and polyangitis who had received prednisone.

Table 1 shows main clinical, epidemiological, laboratory, and instrumental results collected from both mothers and their children.

Routine laboratory tests were in the normal range in all cases. Despite previous routine supplementation, vitamin D serum levels were in the normal range (>30 ng/ml) in only 15/41 patients, while they were decreased in 26 children; these had been exposed to LMWH (n = 6), prednisone (n = 12), and LMWH and prednisone (n = 8). Bone formation and resorption markers were in all cases in the normal range provided by our laboratory. Autoantibodies (ANA, aCL IgM and IgG), at the moment of evaluation, were present in 13 cases (9 children were ANA-positive, and 4 aCL-positive), all born from mothers with the same autoantibodies.

Bone ultrasound recorded QUS values ≤3 percentile for age and gender in 10 children, 10 resulted between >3° and ≤25° and all the other 21 showed a QUS percentile value higher than 25°.

In univariate analysis, QUS percentile values showed a significant correlation with age at examination (r_s: 0.41, p < 0.006), and osteocalcin values (r_s: -0.32, p < 0.04). No significant correlations were detected with the other entered variables, including heparin and steroid use during pregnancy. In multi-stepwise regression analysis, weighted for BMI values at the time of bone ultrasound determination, age at examination resulted the single predictor of QUS values (multiple R = 0.44, multiple adjusted R²: 0.18, p < 0.03) (Figure 1). In children with low (≤3 percentile) QUS values, age at examination resulted lower than in children with normal (>25 percentile) QUS values (median, range: 27 months, 9–129 vs 79 months, 11–135, p < 0.006) (Figure 2).

Tandem mass spectroscopy performed on DBS failed to determine traces of heparin in newborn blood.

Many factors influence the accumulation of bone mineral during childhood and adolescence, including heredity, gender, diet, physical activity and endocrine status [17, 18]. Measures for maximizing bone mineral acquisition, particularly through physical activity and adequate dietary calcium intake, are likely to affect the risk of fracture in later life. In addition to these modifiable factors during childhood, evidence has also shown that the risk of fracture might be programmed during intrauterine life. Maternal smoking, diet (particularly vitamin D deficiency), physical activity, diseases and drugs also appear to modulate bone mineral acquisition in intrauterine life; furthermore, birthweight, weight in infancy, poor childhood growth seem also to be linked to adult bone mass [4‐6, 19, 20].

Systemic autoimmune disorders largely affect women during their reproductive age. Pregnancy in connective tissue diseases has long been associated with poor obstetric outcomes, and can potentially affect disease activity as well. For example, in SLE the risk of spontaneous abortion, intrauterine fetal death, preeclampsia, intrauterine growth restriction and preterm birth is common, such as an active lupus nephritis in mothers [21]. So, a close follow-up is mandatory, including repeated laboratory test and serial ultrasonography in mothers, fetal surveillance tests, and fetal echocardiography for mothers with SS-A (Ro) or SS-B (La) antibodies [22]. In women with primary antiphospholipid syndrome (PAPS) the risk of pregnancy complications increases, including spontaneous miscarriages that typically occur after 10 weeks of gestation, midpregnancy fetal distress, intrauterine growth restriction, prematurity and preeclampsia. In others autoimmune conditions, such as rheumatoid arthritis (RA), systemic sclerosis and Sjögren syndrome the risk of adverse pregnancy outcome is variable.

Moreover, women affected by autoimmune diseases are at high risk of developing osteopenia, osteoporosis and fractures through both disease-specific (chronic arthritis, reduced physical activity, induction of cytokines promoting bone resorption, renal impairment, endocrine factors) and nondisease-specific mechanisms (chronic anticoagulant, glucocorticoid and immunosuppressive treatments, vitamin D deficiency). Regarding thromboprophylaxis, subcutaneous heparin is crucial against the risk of recurrent thromboembolism or pregnancy loss [23]. Heparin and its low molecular weight derivatives can be associated with bone loss and an increased risk of fractures; however, LMWHs seem to be less deleterious to bone than unfractionated heparin [24, 25]. In addition, pregnancy by itself and lactation may also predispose to osteopenia [26, 27].

Unfractionated heparin is a large molecule, so it does not cross the placenta. It seems not to be teratogenic and not cause toxic fetal effects. Several studies have confirmed the safety and efficacy of LMWH during pregnancy as well [28‐34]. LMWHs also does not seem to cross the placenta [30, 31, 33], thus with a fetal safety profile equivalent to that of unfractionated heparin. However, the possibility that unfragmented heparin or LMWH pass the placental barrier to the fetus has been studied only indirectly. Mätzsch et al. studied 21 women who had an abortion by hysterotomy. Laboratory assessments included assays of factor Xa inhibitory activity (XaI), antithrombin III (ATIII) and a measurement of heparin-like substances in plasma with a competitive binding assay. The authors did not demonstrate any evidence for the passage of heparin or LMWH across the placental barrier, furthermore no differences were detected whether unfragmented heparin or LMWH had been given to the mothers [35]. In another study [12], 14 women who were going to have an early termination of pregnancy due to major fetal malformations were given 40 mg of enoxaparin. Maternal blood samples were drawn before and after the injection of enoxaparin, while fetal blood samples were taken only after the drug administration. Anti-IIa and anti-Xa activities were measured. A statistically significant increase of anti-Xa activity in the mothers studied was shown, while there was no detection of anti-IIa and anti-Xa activities in fetus. It was therefore concluded that enoxaparin does not cross the placenta and appeared to be safe for the fetuses. However, such studies and others [8, 35] measured only heparin activity in fetal plasma, and not direct drug presence. We were able to retrieve dried blood spots taken at birth from children whose mothers were administered with LMWH during pregnancy, and could investigate the possible presence of heparin with a new liquid chromatography tandem mass spectrometry method. Tandem mass spectroscopy performed on DBS failed to determine traces of heparin in newborn blood, thus confirming its fetal safety. With regard to corticosteroids, although a fetal effect of high-dose prednisone cannot be excluded, it is known that the fetus is protected from the effect of prednisone because of the placental enzyme 11-β-dehydrogenase, which oxidizes it into an inactive form.

In order to assess bone status, we performed phalangeal QUS, since it is a noninvasive method that provides a surrogate estimation not only of bone density, but also of bone elasticity and architecture, with specific parameters derived from sound waves passing through both cortical and trabecular bone [36‐38]. The measurement site chosen was the hand phalanges, which are composed of predominantly cortical bone and are easily accessibile [39]. Although it is known that dual X-ray absorptiometry (DXA) is gold standard in order to asses bone density, its use in the pediatric age is hampered by the need to sedate young children and by radiation dose, albeit minimal. Moreover, we and others have shown a good correlation between ultrasound and DXA in pediatric rheumatic diseases [40‐43].

Despite the fact that phalangeal QUS measurements have shown the ability to reveal changes due to skeletal growth [44], it is known that bone density values can artificially be modified by bone or body size. Hence, we have transformed QUS results by correcting centile values by BMI, which takes into account both height and weight. In our statistical analysis we have considered all factors that could be confounders for reduced bone density, such as birth weight [20], gender [18], intrauterine complications [5], prematurity, vitamin D supplementation and levels. In the multivariate analysis only age at examination resulted the single predictor of QUS values, in that older children had mostly normal values while younger infants did not. In particular, not only drugs did not seem to pass the placenta but also they were not linked to a low bone mass.

We hypothesize that mothers with autoimmune systemic diseases are at risk for delivering babies with low bone mass, by multiple mechanisms; however, the prognosis for these children with regard to their bone health seems to be good, since when tested during late childhood or preadolescence bone ultrasound values were in the normal range. We acknowledge the limit of the lack of a control group; however, QUS values for healthy children have already been obtained by our laboratory and did correlate with the reference values from the manufacturer that we have used. In addition, we did not have bone marker values from age-matched healthy controls, but most of the results in our series were already in the normal range provided by our laboratory, hence ruling out the possibility of false positives due to high bone turnover in a growing skeleton.

In conclusion, our study suggests that children born from mothers with autoimmune diseases are at risk for low bone mass during the first years of life, and shows that this risk is not linked to exposure to potentially osteotoxic drugs. Moreover, we have confirmed for the first time with a direct assay that LMWHs do not cross the placental barrier.