1. Introduction
Pregnant women are at high risk of vitamin D deficiency. It is estimated that it affects 46–98% of women worldwide [
1,
2,
3]. There is currently very little separate data on women with twin pregnancies, which does not allow to draw general conclusions. However, the experts indicate that hypovitaminosis D may be even more pronounced in this population [
4,
5]. The concentration of vitamin D metabolite – calcidiol [25(OH)D] in the umbilical cord blood of newborns is strongly correlated with maternal levels. According to estimates, it can reach an average of 80% of the maternal value [
6,
7,
8], although some studies have shown that the concentration in umbilical cord blood was higher than in mothers [
9,
10,
11]. The concentration of the active metabolite of vitamin D - calcitriol [1,25(OH)
2D] has not been fully elucidated, although it should be noted that the placenta is able to convert calcidiol to calcitriol due to the production of 1 α-hydroxylase [
12,
13].
The most well-known and documented role of vitamin D during pregnancy is its participation in fetal skeletal growth and mineralization [
6,
12]. The knowledge on the impact of this nutrient on pregnancy outcomes is much more limited. Some studies indicate that low vitamin D concentrations increase the risk of preeclampsia, gestational diabetes, premature birth and affect newborn size [
14,
15,
16,
17]. However, some authors have expressed doubts [
18] or even denied [
19] this relationship.
In the light of scientific literature the reference value of vitamin D concentration in pregnant women is an equally debatable issue. Some experts believe that the normal 25(OH)D concentration during pregnancy is above 30 ng/mL [
20,
21,
22,
23], others suggest >20 ng/mL [
24,
25,
26,
27] while some researchers diagnose vitamin D deficiency only with concentrations below 12 ng/mL [
28]. There is also no consensus regarding the concentration of vitamin D in the umbilical cord blood of the newborn. A deficit is sometimes considered to be <20 ng/mL [
29,
30] or <12 ng/mL [
31], while some experts believe that the normal level of vitamin D in umbilical cord blood is ≥30 ng/mL [
23]. Such significant discrepancies have so far prevented a discussion on whether the criteria for saturation of the body with vitamin D and the guidelines for its supplementation should be the same for women in multiple and in singleton pregnancies.
So far, studies on the impact of vitamin D concentrations during pregnancy on neonatal anthropometric parameters have been conducted almost exclusively among women in singleton pregnancies. Our study included women with twin pregnancies and investigated whether vitamin D concentrations in maternal and cord blood (expressed as 25(OH)D) were associated with neonatal birthweight, length, head and chest circumference.
3. Results
3.1. Vitamin D Status in Mothers and Newborns
Vitamin D deficiency (≤20 ng/mL) was found in 6% of women and 13% of newborns (
Figure 1). Although the newborns had markedly lower concentrations of 25(OH)D than the mothers (24.4 ± 5.9 vs. 39.7 ± 10.7 ng/mL), there was a strong correlation between the level of 25(OH)D in maternal and cord blood (the Pearson linear correlation coefficient is 0.653; p < 0.001). In the linear regression model, an increase in maternal 25(OH)D concentration by one unit, corresponds to an increase in newborn’s level by 0.34 ± 0.04 units (p < 0.001).
The percentage of vitamin D deficient newborns decreased with the higher vitamin D blood saturation in the mothers. All neonates born to women with deficiency were also vitamin D deficient. Among women with suboptimal vitamin D concentration, 29% gave birth to children with a deficiency in comparison to 9% of women with optimal levels. None of the neonates born to mothers with high vitamin D concentrations had a deficit of this nutrient. Vitamin D supplements were taken by 98% of the mothers, with the median dose of 2000 IU (50 µg).
3.2. Neonatal Weight
No statistically significant relationship was found between the concentration of 25(OH)D both in maternal and cord blood and neonatal birth weight, despite a wide range of concentrations being observed: 15.1–64.4 ng/mL in mothers and 10.6 - 39.5 ng/mL in newborns (
Figure 2). There was also no statistically significant difference in the birth weight of children with diagnosed deficiency of vitamin D in umbilical cord blood and those with the normal concentrations.
Neonatal birth weight was related to several other analyzed factors i.e. pregnancy type, gestational weight gain and maternal height. The average birth weight of neonates from dichorionic pregnancies was 202 g higher compared to neonates from monochorionic pregnancies (2651 g vs. 2449 g) (p < 0.001). The average birth weight of neonates born to mothers with excessive weight gain was 151 g higher compared to children born to mothers with normal weight gain (2727 g vs. 2576 g) (p = 0.035). The minimum and maximum birth weight of newborns delivered by women with excessive weight gain was 2350 g and 3030 g, respectively, compared to 1935 g and 3270 g in children born to women with normal weight gain. The average birth weight of children born to mothers ≤ 160 cm was 206 g lower than children born to taller mothers (2368 g vs. 2574 g) (p = 0.006). Also, all neonates born to short-statured women had low birth weight (<2500 g). The birth weight differences among newborns of mothers with and without hypertension during pregnancy were of borderline statistical significance (p = 0.071) (2346 g vs. 2571 g). The other analyzed factors (maternal age, education, place of residence, number of pregnancies, pregestational BMI, season of delivery, vitamin D intake from diet and supplements, calcium intake, caffeine intake, use of multivitamin preparations, gestational diabetes, anemia and sex of the newborn) had no statistically significant relationship with neonatal birth weight.
3.3. Neonatal Length
Similarly to body weight, neonatal length had no statistically significant relationship with 25(OH)D concentration in maternal and cord blood. There were also no significant differences in body lengths among newborns with and without vitamin D deficiency. Statistically significant factors included only maternal height, the type of pregnancy and additionally women’s age. The average (median) length of the neonates from dichorionic twin pregnancies was 1 cm greater compared to the neonates from monochorionic pregnancies (51 cm vs. 50 cm) (p = 0.006). Children born to mothers ≤ 160 cm tall were on average (median) 3.5 cm shorter compared to children of taller mothers (47.5 cm vs. 51 cm) (p = 0.003). Children of mothers aged ≥27 years were on average (median) 1.5 cm longer than children of younger mothers (p = 0.038). As in the case of neonatal birth weight, the length difference in the neonates born to women with and without gestational hypertension was of borderline statistical significance (p = 0.072) (49.5 cm vs. 51 cm).
3.4. Head Circumference
The distribution, but not the average value of head circumference was statistically significantly related to pregestational body weight of the mother and the sex of the neonate. Although the average head circumference of infants born to mothers with normal and excessive body weigh was the same (median 33 cm), the distribution of results was different. The most prevalent head circumference (dominant value) in children born to mothers with normal pregestational BMI was 33 cm vs. 34 cm in children born to mothers with excessive BMI (p = 0.012). The same was observed for the sex of the newborns. The average head circumference was the same for boys and girls (median 33 cm), but the range of head circumference variability was greater in boys (30–35 cm) compared to girls (31–34 cm) (p = 0.020).
3.5. Neonatal Chest Circumference
Neonatal chest circumference was statistically significantly related to several factors including pregnancy type, maternal height, gestational weight gain, gestational hypertension and the sex of the neonate. Infants from dichorionic twin pregnancies had on average (median) 1 cm larger chest circumference compared to infants from monochorionic pregnancies (31 cm vs. 30 cm) (p < 0.05). In infants of women ≤160 cm tall, this parameter was on average (median) 2 cm lower than in infants of women over 160 cm (29 cm vs. 31 cm) (p = 0.005). The chest circumference of infants born to mothers with excessive weight gain was on average (median) 1 cm larger than that of infants born to mothers with normal weight gain (32 cm vs. 31 cm) (p = 0.007). Children born to mothers with hypertension had on average (median) 2 cm smaller chest circumference than children of mothers without hypertension (29 cm vs. 31 cm) (p = 0.046), and boys had on average (median) 1 cm larger chest circumference than girls (31 cm vs. 30 cm) (p = 0.034).
4. Discussion
To the best of our knowledge, the relationship between vitamin D status in women with twin pregnancies and the birth weight of their newborns has been investigated only in two published studies conducted in China [
36] and Italy [
37]. Both studies found no association between maternal 25(OH)D concentrations and neonatal birth weight, although in the first study almost 80% of children had vitamin D deficiency at birth, whereas in the second study 38% of mothers had vitamin D levels below 30 ng/mL in the third trimester of pregnancy. Our study also did not show any relationship between maternal or cord blood 25(OH)D levels and any of the anthropometric parameters of the newborns routinely measured at delivery, although with the same deficiency criterion as in the Chinese study (vitamin D concentration <20 ng/mL), only 13% of our newborns had vitamin D concentration below guideline values. The significant difference in the percentage of deficient newborns is probably related to the fact that Polish women generally take much higher doses of vitamin D than women in China (2,000 IU vs. 500 IU).
Numerous studies have been conducted among women in singleton pregnancies. In an observational study in the UK, no differences in neonatal body parameters were found between groups of mothers with vitamin D levels <12 ng/mL vs. >30 ng/mL [
38]. In an Australian study, maternal vitamin D deficiency (<20 ng/mL) did not correlate with lower neonatal weight and length, but such low levels were only found in <7% of the mothers, probably due to the widespread supplementation of this nutrient (declared by 98% of the patients) [
1]. No association between maternal and cord blood vitamin D concentrations and any of the neonatal anthropometric parameter [
39] or fetal femoral length [
40] was found in Polish studies, despite the fact that 50% of the mothers and almost 30% of the newborns had vitamin D concentrations below 20 ng/mL. On the other hand, a Chinese and a Turkish study found a correlation between neonatal parameters and vitamin D status of the mother. In the former study, infants born to deficient women had a slightly (by 65 g) but significantly lower birth weight than newborns of women with normal vitamin D status [
41]. In the latter study, infants born to women supplementing vitamin D during pregnancy (58% of the group) were longer and had greater head and chest circumferences than infants born to women who did not use vitamin D supplementation [
42]. Contradictory results regarding the relationship between maternal and umbilical cord blood vitamin D concentrations and anthropometric parameters of the newborns were obtained in a study conducted in Finland [
43].
The results of a randomized, placebo-controlled
trials (RCTs) which have so far only been conducted in women in singleton pregnancies are equally inconsistent. The above can be illustrated by several Iranian studies. Vaziri et al. [
44] and Mohammad-Alizadeh-Charandabi et al. [
45] showed that maternal vitamin D supplementation was not associated with newborns’ size, whereas Naghshineh and Sheikhaliyan [
46] concluded that infants born to mothers taking vitamin D during pregnancy were heavier by an average of 230 g. In an Indian study, vitamin D supplementation in dosages depending upon blood concentrations was found to increase neonatal birth weight by an average of 200 g [
47]. Roth et al. stated that infants born to mothers from the supplementation group did not differ in size from the newborns of mothers from the control group [
48]. Finally the conclusions from meta-analyses of RCTs are also inconsistent [
2,
49,
50,
51,
52].
In our study, the anthropometric parameters of the newborns were associated with the type of pregnancy. Larger neonates were born to women in dichorionic pregnancies compared to monochorionic pregnancies, which is consistent with the current state of knowledge. In a large Dutch study, the mean birth weight of newborns from dichorionic pregnancies was 221 g greater compared to newborns from monochorionic pregnancies [
53], which is very similar to the results of our study (202 g difference). It is worth adding that the median length of dichorionic pregnancy in both studies was one week longer than for monochorionic pregnancy (37 vs. 36 weeks). Higher birth weights of infants from dichorionic pregnancies compared to monochorionic pregnancies were also reported in Finland [
54] and Thailand [
55].
In twin pregnancies; as in singleton pregnancies; the growth of the fetus and the birth weight of the newborns are associated with gestational weight gain [
35,
39,
56,
57], which was also proven in our study. Children born to mothers with excessive weight gain were heavier than infants born to mothers with normal weight gain. It is worth noting that in twin pregnancies; which are associated with a higher risk of low birth weight; greater gestational weight gain may in some cases even promote better fetal growth of the newborns. In a large American study by Bodnar et al.; the risk of delivering small-for-gestational-age neonates decreased with increasing gestational weight gain in women in twin pregnancies; while the risk of large-for-gestational-age neonates was observed to increase [
56]. The same conclusion emerges from a meta-analysis of observational studies conducted among women in singleton pregnancy [
57].
The relationship between maternal height and neonatal size determined in our study is also not surprising, as it has been quite well documented in women in singleton pregnancies. Maternal and paternal height reflects the genetic growth potential of the fetus, with taller mothers generally giving birth to larger children [
58,
59,
60]. In addition to genetic factors, anatomical factors may also play an important role. Such factors can include a smaller pelvis or smaller uterus size in short-statured women, which can affect uterine expansion during pregnancy, the growth of the placenta and the fetus [
60]. In our study, neonates born to mothers up to 160 cm tall were over 200 g lighter and markedly shorter (3.5 cm on average) compared to neonates born to mothers of taller stature (above 160 cm). In the Finnish study, birth weight, length and head circumference increased with maternal height, regardless of the sex of the neonate. For example, boys born to mothers ≤ 158 cm tall weighed on average 265 g less and were 1.3 cm shorter than boys born to mothers ≥ 173 cm tall [
59]. These neonatal anthropometric parameters were also found to correlate with maternal height in an Austrian study [
58], while in the American study, the mother’s height correlated with the birth weight of the newborns [
61]. In the light of a meta-analysis of studies conducted in singleton pregnancies, short-statured women have a greater risk of giving birth to newborns with low birth weight [
60].
As to the sex of the neonate, it is often related to birth size. Girls are often born smaller than boys [
58,
59], but there are also studies that do not show such a difference [
39]. The newborn girls included in our study had smaller head and chest circumferences.
When interpreting our results, the limitations of the study should be taken into account. Firstly, the measurements of maternal 25(OH)D concentrations were performed in the perinatal period, which does not mean that they were typical throughout or during most of the pregnancy course. Therefore, it is not known whether newborns had a relatively constant supply of vitamin D in utero. Another limitation of our study is the relatively small sample size mostly due to the small overall population of women pregnant with twins. It is also worth noting that the study was carried out in two separate tertiary centres in Warsaw, which also run separate outpatient clinics for multiple gestations. It can therefore be assumed that women had better access to antenatal education, including guidelines on vitamin supplementation. Therefore, the study group may not adequately reflect vitamin D status of patients from smaller centres.