Introduction

Vitamin D deficiency is a common problem in reproductive aged women in industrialized countries and its prevalence may be increasing [1]. The etiology for this increase is likely multifactorial, but due in part to low dietary intake of vitamin D and limited exposure to sunlight. Certain high risk groups for vitamin D deficiency in pregnancy have been identified, including vegetarians, women with limited sun exposure (e.g., those who live in cold climates, northern latitudes, or wear sun and winter protective clothing), and ethnic minorities, especially those with darker skin [2]. Vitamin D deficiency is also more common among heavier women than leaner individuals [3]. During pregnancy, severe maternal vitamin D deficiency, as defined as serum 25-hydroxyvitamin D (25-OH vitamin D) concentrations less than 5 ng/mL, has been associated with disordered skeletal homeostasis, congenital rickets, and fractures of the newborn [4]. However, the effect of less severe vitamin D deficiency and insufficiency on maternal and fetal outcomes during pregnancy is less clear. Recent observational and randomized control trials have attempted to investigate this issue, but there remains limited guidance on the management of vitamin D deficiency during pregnancy.

Vitamin D deficiency and insufficiency have been associated with a variety of adverse maternal and fetal outcomes, ranging from preeclampsia, gestational diabetes, preterm delivery, intrauterine growth restriction, spontaneous abortion, and cesarean section [5–10]. However, other studies have demonstrated no association between vitamin D status and adverse pregnancy outcomes [11–12]. The majority of randomized control trials and observational studies measure vitamin D late in the pregnancy, when prenatal vitamins containing 400 IU of vitamin D have been prescribed [13]. Thus, we sought to determine if vitamin D deficiency in the first trimester of nulliparous patients is associated with clinical outcomes most associated with significant maternal and fetal morbidity and mortality: preeclampsia, gestational diabetes, preterm delivery, spontaneous abortion, and growth restriction. Nulliparous women were studied to decrease the effect of prior pregnancy complications on study outcomes. The objective of this study was to analyze the relationship between hypovitaminosis D in the first trimester and subsequent clinical outcomes in a population of nulliparous women.

Study Design

Results

Of the 315 patients approached for study participation, a total of 310 patients agreed to participate and gave informed consent according to the Institutional Review Board approval of the study protocol (response rate: 98.7%). A total of 235 patients were included in the final analysis (mean age 24.3 years, range 18–40 years). Of the original 310 patients recruited, 75 were not included in the study because they withdrew from the study (n = 8), did not deliver at our institution (n = 44), underwent elective pregnancy termination (n = 3), were lost to follow-up (n = 18) or were diagnosed with twin gestations after enrollment (n = 2) (attrition rate of 24.2%). Of the 235 patients included in the study, 41 of them had a prior pregnancy which had resulted in a spontaneous first trimester miscarriage or elective termination. Mean gestational age of recruitment and vitamin D plasma measurement was 62.2 ± 29.7 days (8.9 ± 4.2 weeks).A pregnancy was considered complicated if the patient was diagnosed with any of the following adverse pregnancy outcomes: preeclampsia, gestational diabetes, spontaneous preterm delivery, intrauterine growth restriction, or spontaneous abortion (Table 1). For the purposes of this study, patients who did not experience one of these complications were categorized as uncomplicated. Approximately 25% of patients had complicated pregnancies. Frequencies of pregnancy complications are tabulated in Table 1 and compared with predicted US and institutional frequencies for preeclampsia [20–26], preterm delivery [27–28],intrauterine growth restriction [29–30], gestational diabetes mellitus [31–36], and spontaneous abortion [37–38]. As shown in Table 1, three patients categorized in the spontaneous preterm delivery category had comorbidities including intrauterine growth restriction (IUGR) and gestational diabetes. These patients were not included in the IUGR or gestational diabetes categories for subsequent statistical analysis. Similarly, the 5 patients with a medically indicated preterm delivery subsequent to preeclampsia were not included in the preterm delivery categories for subsequent analysis. The predicted adverse event frequencies were comparable to national and international epidemiological predictions (Table 1). The patient population comprised active duty military or dependent spouses (Table 2). Information on patient time in the area prior to pregnancy, dietary intake of vitamin D, and sun exposure was not available. All patients were prescribed prenatal vitamins (approximately 400 IU/day) at their initial intake appointment, but compliance rates on use were unknown. When actual plasma vitamin D concentrations were determined for the 235 patients included in the final analysis, technical reproducibility among replicates was assessed by calculating the coefficient of variation (CV) among calibrators, controls, and samples. CVs were 15.8% for calibrators, 8.5% for assay controls, and 7.8% for patient plasma samples. Ten percent of our study population was vitamin D deficient with a serum concentration less than 20 ng/mL (50 nmol/L); 60% were vitamin D insufficient (serum concentrations between 20–29 ng/mL or 50–74 nmol/L); and 30% of our study population were vitamin D sufficient with serum concentrations above 30 ng/mL (75 nmol/L) (Table 2). None of our patients were classifiable as severely vitamin D deficient (<10 ng/ml or 25 nmol/L). The mean serum 25-OH vitamin D concentration for our study population was 27.6 ng/mL (range 13–71.6 ng/mL).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Frequency of clinical outcomes compared to national and global epidemiology.

https://doi.org/10.1371/journal.pone.0123763.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Demographics of the study population by vitamin D classification.

https://doi.org/10.1371/journal.pone.0123763.t002

Forty-seven percent of our study population was of normal weight with a BMI <25 (Table 2). Thirty-five percent were overweight with a BMI at intake between 25–30, and about 18% of our patients were obese with a BMI >30. Most of our study population was Caucasian (78%), with 7% African American, 3% Asian, and 12% other, unspecified races. Thirty-eight percent of the blood draws occurred in the winter, 19% in the spring, 22% in the summer, and 21% in the fall. About 6% of our study population used tobacco during pregnancy. Statistical associations between vitamin D categorization and demographic variables are depicted in Table 2. Race was the only statistically different variable among the vitamin D groups, with significantly more African American, Asian, and other unspecified racial groups in the vitamin D deficient category compared to the insufficient and sufficient groups (p = 0.03).

The distribution of clinical outcomes among the vitamin D categories is depicted in Table 3. Study population percentages and with medians and ranges of vitamin D measured in plasma were calculated. Preeclampsia was sub-stratified into severe, early onset preeclampsia and mild, late-onset preeclampsia.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Plasma vitamin D concentrations in women according to clinical outcome.

https://doi.org/10.1371/journal.pone.0123763.t003

Plasma vitamin D concentrations were subdivided into 3 categories based on the Institution of Medicine and Endocrine Society criteria for vitamin D categorization: sufficient (>30ng/ml [75nmol/L]), insufficient (21-29ng/ml [50-75nmol/L], or deficient [<50nmol/L). None of our patients were categorized as severely deficient (<10ng/ml [25nmol/L]) (Table 4). Because there were too few patient in the category of vitamin D deficient (<20ng/ml [50nmol/L]) for meaningful statistical analysis, patients categorized as either vitamin D insufficient or deficient were combined into one category (insufficient/deficient, <30ng/ml [75nmol/L]) for subsequent statistical analysis (Table 5).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Distribution of clinical outcomes in women categorized as Vitamin D deficient, insufficient, or sufficient.

https://doi.org/10.1371/journal.pone.0123763.t004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Logistic regression analysis of clinical outcomes in women categorized as vitamin D sufficient compared to vitamin D insufficient or deficient women.

https://doi.org/10.1371/journal.pone.0123763.t005

There was no association between vitamin D deficiency in the first trimester and adverse pregnancy outcomes with an adjusted odds ratio of 1.01 (p value 0.738, 95% CI 0.947–1.164). Logistic regression was performed on combined outcomes and individual outcomes separately. There was no association between first trimester vitamin D levels and subsequent development of combined complications, or complications sub-stratified into preeclampsia, gestational diabetes, preterm delivery, intrauterine growth restriction (IUGR), or spontaneous abortion when each outcome was evaluated individually (Table 5). Odds ratios were adjusted for BMI, ethnicity, tobacco use, and season. There was no association between first trimester vitamin D levels and subsequent development of composite adverse pregnancy complications (including preeclampsia [severe, early onset or mild, late onset], gestational diabetes, preterm delivery, growth restriction, or spontaneous abortion (combined outcomes; p > 0.05 in Table 5).

Discussion

This study investigated the prevalence of hypovitaminosis D in young (24.3 ± 4.4 years), nulliparous women at a tertiary military medical facility in the Pacific Northwestern United States at an average gestational age of 8.9 ± 4.2 weeks. Our study is distinctive as we examined a healthy population of women who should be at relatively lower risk for pregnancy complications. These women did not have preexisting hypertension, diabetes, or other medical problems that would increase their risk for pregnancy complications or vitamin D deficiency. Although we recruited only young, nulliparous, previously healthy women, the majority of our patients were either vitamin D deficient (10%, <20 ng/ml) or vitamin D insufficient (60%, 20–29 ng/ml) according to the criteria published by the Institute for Medicine.¹⁸ Of the 235 patients included in the study, 41 of them had a prior pregnancy which had resulted in a spontaneous first trimester miscarriage or elective termination. As these prior losses were not associated with a diagnosis of infertility, recurrent pregnancy loss or other medical conditions, we felt there would be minimal confounding by including them in our study. We found no association between hypovitaminosis D (plasma levels of <30 ng/ml) and the combined clinical outcomes of preeclampsia, preterm delivery, intrauterine growth restriction, gestational diabetes, or spontaneous abortion in this patient population. This study was not powered to detect associations between vitamin D status and these individual pregnancy complications. Fig 1 compares our study population with recently published cohort studies and randomized controlled trials (reviewed in Christesen [2012] and Wei [2013]) [13, 39].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Comparison of study parameters in recent vitamin D observational and intervention studies.

https://doi.org/10.1371/journal.pone.0123763.g001

Risk factors for vitamin D deficiency include inadequate sun or UV light exposure, physiologic factors (such as dark skin pigmentation, obesity, pregnancy, certain medical problems), and low vitamin D intake [2–3]. The prevalence of vitamin D deficiency in reproductive aged women appears to be increasing worldwide over the past two decades.^1,3,13 Recent studies indicate the prevalence of vitamin D insufficiency and deficiency in nulliparous women, a result corroborated by our study [40]. The etiology of the decline of vitamin D status in women is likely multifactorial, but is due in part to increasing BMI and increased sun protection, coupled with decreased dietary intake of vitamin D, such as milk and other dairy products. Unfortunately, the increasing use of multivitamins and mineral dietary supplements in younger to older adults does not appear to be associated with a corresponding increase in serum 25-OH vitamin D concentrations [3, 19, 41]. Recent clinical guidelines suggest that doses of vitamin D greater than 600 IU/day may be needed to maintain adequate plasma levels greater than 30 ng/ml (75 nmol/L), but the appropriate standard is still debated [19, 42–43]. In recent randomized controlled studies, loading doses as high as 1000–4000 IU/day were needed before adequate vitamin D plasma concentrations were measured [18–19, 44].

Other risk factors for vitamin D deficiency include ethnicity, as darker skinned individuals have decreased vitamin D3 synthesis in response to ultraviolet light. Vitamin D deficiency has been proposed as a possible biologic explanation for racial disparities noted in adverse pregnancy outcomes including preeclampsia, spontaneous preterm birth, gestational diabetes and growth restriction [5]. Because vitamin D deficiency is so widespread, determining whether vitamin D deficiency is associated with adverse pregnancy outcomes is an area of major public health concern. It is unclear whether vitamin D deficiency is causal in the disparities in adverse pregnancy outcomes between Caucasians and African Americans. Out of 235 patients in our study, only about 7% were African American and thus we were not powered enough to examine this significance. However, even after adjusting for ethnicity in our analysis, there was still no significant difference between hypovitaminosis D and clinical outcome.

There are two forms of vitamin D in the body: vitamin D2 and vitamin D3. Measures of blood levels reflect both dietary intake as well as synthesis from exposure to the sun. The liver converts both forms of vitamin D to 25-OH vitamin D, which is the specific metabolite measured to determine vitamin D status. These levels are considered to be an accurate representation of the vitamin D status of an individual [1, 45]. A serum level of at least 20 ng/mL (50 nmol/L) is needed to avoid bone problems, and the optimal level for promoting health was recently published as 30 ng/mL (75 nmol/L) [2, 5, 18–19]. However, it is important to note that these thresholds were derived from a population of nonpregnant individuals and an optimal serum level during pregnancy has not been established.

Season and latitude dramatically alter vitamin D3 production [46–47]. In wintertime, the angle of the sun is so oblique that latitudes greater than 35 degrees receive almost no ultraviolet rays capable of stimulating vitamin D3 synthesis. As our study center is around 47 degrees latitude and 38% of our first trimester blood draws occurred in the winter, this certainly could contribute to the prevalence of vitamin D deficiency in our population. In addition, the Pacific Northwest experiences fewer sunny days than the rest of the country, with only 46% of daylight hours that are predominantly sunny or partly cloudy. For example, the greater Seattle area experiences an average of 8 and a half hours of daylight hours during December but only roughly 2 hours during the day are considered “sunny”. Compare this to South Carolina, which experiences over 6 hours of sunshine out of almost 10 hours of daylight during the same month. Unfortunately, data on individual daily sun exposure, travel and dietary intake are not available [48–49]. Other limitations and sources of possible bias for this study include low generalizability (small cohort of patients in the Pacific Northwestern United States), small sample size, self-selection of study subjects, and the high loss-to-follow-up rate in our study population. Information on the job description of active duty women was unavailable for this study, as was data on physical activity and location of duty station influencing sun exposure (i.e., indoors vs. outdoors). Body mass index in this population may not necessarily be indicative of overweight or obese given that some active duty women may have more muscle mass contributing to BMI measurement.

Several recent randomized controlled trials and observational cohort studies investigate a possible association between hypovitaminosis D and adverse clinical outcome in pregnancy (reviewed in Christesen [2012]) [13]. The role vitamin D plays in adverse clinical outcomes such as preeclampsia, gestational diabetes, preterm labor, growth restriction, and spontaneous abortion is speculative, but may involve vitamin D’s regulation of placental trophoblast invasion and angiogenesis, in addition to its anti-inflammatory properties [5]. For example, there is some evidence vitamin D affects transcription and function of genes responsible for trophoblast invasion and angiogenesis, two factors critical for placenta development. Vitamin D deficiency may predispose individuals to abnormal trophoblast invasion, reduced placental perfusion, and the subsequent cascade of events resulting in preeclampsia. Maternal vitamin D deficiency may also predispose women to an increased inflammatory response that characterizes preeclampsia, especially severe, early onset disease (Reviewed in Christesen [2012] and Wei [2013]) [13, 39]. In our cohort of young, nulliparous patients, we had a preeclampsia rate of 8%. However, vitamin D deficiency did not increase the odds of preeclampsia, even when stratifying by disease severity. This discrepancy may be explained in part by differences in our study population relative to other published cohort studies, including difference in parity, maternal age at intake, and/or gestational age of specimen collection (Fig 1) [6, 10–11, 40, 50–70]. During pregnancy, 25-OH vitamin D diffuses across the placenta and the fetus relies entirely on the vitamin D stores of the mother: if the mother is deficient, so is the fetus [5]. Thus, in combination with its immunomodulatory and anti-inflammatory properties, vitamin D may play a potential role in the prevention of preterm birth and small-for-gestational-age neonates. Vitamin D deficiency also has been shown to increase insulin resistance and reduce insulin secretion, which has shown to be a risk factor for gestational diabetes [5, 71–72]. However, whether vitamin D deficiency is a risk factor for gestational diabetes in itself or if vitamin D supplementation can prevent GDM is yet unknown [72].

Although vitamin D has a biologically possible role in all the aforementioned disorders, this was not demonstrated in our population as vitamin D deficiency did not increase the odds for adverse pregnancy outcome in our study. Study-specific parameters such as maternal demographics and parity may account for some of these differences in our cohort relative to published studies (Table 2 and Fig 1). For example, we had a lower incidence of gestational diabetes in our study population compared to what is reported in the literature [31–35, 73–75].Risk factors for gestational diabetes include women older than 25, obesity (BMI >30), history of a large baby or previously affected pregnancy, and women who are African American, American Indian, Asian American, Hispanic, Latina, or Pacific Islander [14]. We feel our low incidence of GDM is likely due to the young age (average age 24 years old), nulliparity, relatively low rate of obesity (17.9%), and predominantly white ethnicity (77%) of our study population. Obesity is also a known risk factor for vitamin D deficiency[73]. We reviewed the BMI of our study participants to see if there was any association between vitamin D deficiency classification and BMI. Out of the 235 patients analyzed, 42 (17.9%) had a prepregnancy BMI >30 and were considered obese. However, there was no significant difference in the distribution of these patients among the various vitamin D categories, with 10/70 (14.3%) women who were vitamin D sufficient, 29/141 (20.5%) vitamin D insufficient, and 3/24 (12.5%) vitamin D deficient (p = 0.22) (Table 2).

Our nulliparous patient population is also at different risk for certain pregnancy complications, such as preeclampsia and preterm delivery, compared to their multiparous counterparts. The prevalence of preeclampsia in the United States is about 3.4 percent, but 1.5-fold to 2-fold higher in first pregnancies [22]. In fact, most cases of preeclampsia occur in nulliparous women without significant medical history or risks factors [15]. Our nulliparous population may also be at lower risk for preterm delivery as one of the strongest clinical risk factors for preterm birth is a prior preterm birth [76]. Behavioral risk factors for preterm birth include low maternal prepregnancy weight, smoking, substance abuse, and short interpregnancy interval. Only three of our patients had a BMI less than 19; all three of these women had full term deliveries. Additionally, our maternal tobacco use rate was only 6.4%, compared to national averages closer to 13% [77]. All of these factors may have contributed to our low rate of preterm delivery compared to nationally reported averages.

Conclusion

Vitamin D deficiency was not associated with the composite clinical outcomes of preeclampsia, preterm delivery, intrauterine growth restriction, gestational diabetes, or spontaneous abortion in this study population of nulliparous military members and their beneficiaries. This study was not powered to detect associations between vitamin D status and these individual complications of pregnancy. However, the high percentage of affected individuals highlights the prevalence of vitamin D deficiency and insufficiency in young, healthy reproductive-aged women in our study population. Vitamin D deficiency is a growing topic of interest around the world. Although our study did not demonstrate an association of vitamin D plasma concentration with select clinical outcomes, the prevalence of sub-optimal vitamin D concentrations in patient plasma emphasizes the need for continued research on hypovitaminosis D in pregnancy and reproductive health.

Acknowledgments

We would like to thank Diazyme Laboratories for their technical assistance with the 25-OH Vitamin D enzyme immunoassay as well as for their review of our manuscript.

Disclaimer: The views expressed in this presentation are those of the authors and do not reflect the official policy of the Department of the Army, the Department of Defense or the U.S. Government.