| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
Department of Epidemiology (M.S., J.G., M.J.), University of Michigan, Ann Arbor, Michigan 48109-2029; and University of Medicine and Dentistry of New Jersey (T.S., X.C.), Piscataway, New Jersey 08855-1179
Address all correspondence and requests for reprints to: MaryFran Sowers, Ph.D., Department of Epidemiology, University of Michigan, 109 South Observatory, Room 3073, SPH I, Ann Arbor, Michigan 48109-2029. E-Mail: mfsowers@umich.edu.
Abstract
Pre-eclampsia is a pregnancy disorder of uncertain etiology that affects 5–10% of all pregnancies, with symptoms typically presenting around or after 20 wk gestation. We hypothesized that IGF-I, osteocalcin, and bone loss would be different among women with pre-eclampsia compared with normotensive pregnant women.
There were 962 pregnant healthy women, aged 12–35, who were assessed at entry to care, at 28 wk, and at delivery for osteocalcin and IGF-I concentrations. Bone ultrasound was measured at entry to care and at 6 wk postpartum, whereas bone mineral density was measured by dual x-ray densitometry at delivery.
There were 64 women (6.7%) among the women being followed who developed pre-eclampsia. In women with pre-eclampsia, IGF-I concentrations were 74% greater in the third trimester compared with the first trimester, whereas there was little change in osteocalcin concentrations. In contrast, normotensive women had an average increase of 43% in IGF-I concentrations accompanied by a 63% decline in osteocalcin concentrations. In women with pre-eclampsia, IGF-I and osteocalcin concentrations were significantly correlated (r = 0.48 and 0.43) at both the first and third trimester time points, but only in the third trimester among normotensive women (r = 0.27). The bone change difference between the two groups was not statistically significant.
Women with pre-eclampsia appear to have an exaggerated IGF-I responsiveness compared with women who are normotensive; however, the strong correlation between IGF-I and osteocalcin in women with pre-eclampsia suggests that the IGF-I is able to retain its role as a local regulator of bone remodeling, as indicated by the osteocalcin concentrations.
THERE IS EVIDENCE of an important role of IGFs in the local regulation of bone formation (1, 2, 3). In rat calvarial explant cultures, IGF-I increased collagen synthesis, matrix apposition, and cell replication (4). The significant influence of IGF-I on bone metabolism is generally supported by findings in vivo among adults. Healthy postmenopausal women treated with recombinant human IGF-I had dose-dependent increments in biochemical markers for osteoblastic and osteoclastic activity (5). Likewise, individuals treated with recombinant human GH had rises in IGF-I concentrations (6) and increased bone formation and bone resorption, as reflected by the activity of biochemical markers (6, 7). Observational studies have been less consistent, with some studies concluding that there were associations between serum IGF-I and bone turnover or bone mass, whereas others have indicated no association between serum IGF-I and bone turnover or bone mass (7, 8, 9, 10, 11).
There are both theoretical and empirical reasons to believe that results relating IGF-I to osteocalcin concentrations and bone mineral density (BMD) could be extended to pregnancy. Pregnancy is a high bone turnover state because the maternal calcium homeostasis adapts to meet the increased calcium demands of the growing fetus. Wilson et al. (12) reported that IGF-I concentrations measured at a single time in the third trimester serum were significantly higher in pregnant women compared with women who were not pregnant, although the study could not determine whether the IGF-I concentrations were of placental or maternal origin. Naylor et al. (13) followed 16 women through pregnancy and reported increased serum IGF-I concentrations in the third trimester as compared with baseline (P < 0.05), increased bone turnover markers, and increased BMD at cortical bone sites but decreased BMD at trabecular bone sites measured in the postpartum.
The relationship of IGF-I to calcium metabolism in pregnancy may be different among those women with pre-eclampsia, a pregnancy disorder that affects 5–10% of all pregnancies and whose symptoms typically present after 20 wk gestation. However, there are few studies that address this area, and the number of patients studied is small. Guidice et al. (14) reported circulating IGF-I concentrations were significantly lower in the late second and early third trimesters of eight pregnant women with severe pre-eclampsia compared with the six controls (80.9 ng/ml compared with 179.4 ng/ml, respectively; P = 0.0001). Halhali et al. (15) reported that 24 women with pre-eclampsia, evaluated at delivery, had significantly lower circulating IGF-I concentrations compared with a normotensive control group (265 ng/ml compared with 383 ng/ml, respectively; P < 0.0001).
Studies have not integrated measures of IGF-I, bone status, and bone turnover during pregnancy. On the basis of the potentially important role of IGF-I in bone and calcium metabolism during pregnancy and the relative lack of information about IGF-I in pregnancy, we posed the following questions. Were the measured changes in IGF-I concentration during pregnancy associated with changes in bone turnover, assessed by osteocalcin concentrations? Were similar relationships observed in women with and without pre-eclampsia? Were the IGF concentrations associated with change in bone ultrasound and BMD?
Methods and Measures
Study population
The Camden Study of Calcium Metabolism in Pregnancy and Lactation asked 1117 women in 3 prenatal programs located in Camden, New Jersey, to participate in a research study through their pregnancy and postpartum. Of those contacted, 89 women refused participation, and 66 women who had a history of serious nonobstetric conditions (e.g. lupus, chronic hypertension before pregnancy, type I or type II diabetes, seizure disorders, malignancies, drug or alcohol abuse) were ineligible for participation. Of the 962 women recruited, 92% (n = 878) were followed in a research protocol commencing at time of entry to care and closing at delivery that included measures of IGF-I. Data from 869 women were used in these analyses, excluding the data from 9 women who developed hypertension during the course of pregnancy. The sample was 46% African-American, 38% Hispanic, and 16% Caucasian, consistent with the race and ethnic proportion of the community.
The participants were interviewed about sociodemographic characteristics and their medical history. Their body size and other physical characteristics were also measured. Selected aspects of data collection were designed to optimize participation and participant safety and to minimize participant burden. For that reason, bone characteristics of women during pregnancy were assessed by bone ultrasound measurements at entry to care and at the 6–7 wk postpartum check, whereas BMD was measured with dual energy x-ray densitometry at delivery as well as 3, 6, and 12 months postpartum. The schedule for phlebotomy during pregnancy and the postpartum period was at entry to care, 28 wk, delivery, 3, 6, and 12 months. All participants provided written informed consent documentation consistent with the policies of institutional review boards at the University of Medicine and Dentistry of New Jersey and the University of Michigan.
Measurement of osteocalcin concentrations and IGF-I levels
Blood was drawn by venipuncture at each examination. Serum samples were aliquoted and then stored at -80 C. A commercial RIA (INCSTAR Corp., Stillwater, MN) was used to assay for serum osteocalcin. The combined intra- and interassay variation was less than 12%. After an acid ethanol extraction, IGF-I concentrations were measured with a RIA (Diagnostics Systems Laboratories, Inc., Webster, TX). The level of detection was 5 ng/ml, and the combined intra- and interassay variation was 3.6–5.2%. Although IGF-I concentrations in the third trimester can reflect both maternal and placental contributions, assays do not distinguish the source.
Bone measurements
Bone ultrasound methodology (Sahara, Hologic, Inc., Bedford, MA), conducted at entry to care and at 6 wk postpartum, gave three measures of bone mass. The speed of sound (meters per second) was the speed of the signal transmission through the heel, whereas broadband ultrasound attenuation (dB/MHz) was the degree of the attenuation of the high-frequency sound waves. The quantitative ultrasound index (QUI) combined speed of sound and broadband ultrasound attenuation into a single measure that could range from 0 to 170, with the greater values being associated with greater bone mass. Machine calibration, using a quality control phantom, was completed daily. The coefficient of variation of this measure is approximately 3% (16). Bone ultrasound change during pregnancy was ascertained by subtracting the measure taken 6 wk after delivery from the baseline measure taken at entry to care.
BMD of the femoral neck and lumbar spine was measured by dual x-ray densitometry (DEXA) (Lunar Corp., Madison, WI; DPX-L, analysis software version 1.3y). Measurements were made within 2 d of delivery. Calibration was performed daily, and a lumbar spine phantom was scanned weekly. The coefficient of variation for DEXA was less than 1.0% for the femoral neck site.
This analysis used two covariates, including the participant’s age at initial entry to care and parity, a dichotomous variable that indicated whether or not the individual was nulliparous or parous.
Data analysis
Data analysis was performed separately for study participants with pre-eclampsia and normotensive participants or identified with the use of a dummy variable with normal blood pressure as the referent group. Pre-eclampsia was classified on the basis of two criteria: diagnosis abstracted from the medical records and substantiated with blood pressure cut points of 140 mm Hg systolic (and >30 mm Hg change) and 90 mm Hg diastolic with greater than 15 mm Hg change.
Univariate statistics were calculated for continuous variables, and frequencies with statistics were calculated for categorical variables. Variables (IGF-I, osteocalcin, and bone ultrasound) were evaluated for the normality of their distribution and for outliers, and this resulted in the use of logarithmic transformations.
Initially, a cross-sectional approach was chosen to examine the relationships of IGF-I concentrations during the course of pregnancy and osteocalcin concentrations and bone ultrasound change. These cross-sectional analyses were approached systematically with an evaluation of the bivariate associations, followed by a multiple variable analysis that adjusted for the covariates. Model building was undertaken systematically using a two-step process. First, the unadjusted associations were examined. Second, covariates were evaluated for their correlation with IGF-I concentration, and variables that were statistically significant at the 0.10 level were considered candidates for the multivariate analysis.
Longitudinal analysis with mixed models was also used to investigate the pattern of IGF-I level change during pregnancy as well as to evaluate those IGF-I changes to change in osteocalcin concentrations or change in ultrasound bone status. This approach capitalizes upon the information provided by repeated observations over time of both IGF-I and osteocalcin concentrations. These analyses assume that errors were correlated within an individual, but that observations were independent across individuals. In addition, missing observations were assumed to be missing at random. The ß coefficients and SE in the longitudinal models remain in their transformed state because of the complexity of the SE measure. All statistical procedures were performed using SAS v.6.12 (SAS Institute, Inc., Cary, NC).
Results
Table 1
summarizes select
characteristics of study participants subcategorized according to the
presence or absence of pre-eclampsia. The profiles were classic. The
age of women with pre-eclampsia was, on average, 17 months younger than
the average age of normotensive women (P < 0.02).
Approximately three fourths of participants with pre-eclampsia were
nulliparous, whereas only half of women with normal pregnancies were
nulliparous (P < 0.01). There were no statistically
significant differences in the race and ethnic composition of the two
groups (P < 0.46).
|
), the mean IGF-I
concentrations were similar in the two groups (P <
0.75), with concentrations being approximately 200 ng/ml. However, the
average IGF-I concentrations increased differentially in the second and
third trimesters, depending upon whether or not women developed
pre-eclampsia. Normotensive women had an IGF-I increase of
approximately 43% in the third trimester, compared with the first
trimester. Among women who would be defined with pre-eclampsia, the
IGF-I concentrations were 74% greater in the third trimester, compared
with the first trimester. The difference was statistically significant
(352 vs. 294 ng/dl; P < 0.005).
As shown in Table 1, average osteocalcin concentrations among women
that would be diagnosed with pre-eclampsia were relatively consistent
throughout pregnancy and delivery (0.294, 0.250, 0.268, and 0.229
nmol/liter, respectively). The pattern was different among women with
normotensive pregnancies whose osteocalcin concentrations dropped by
57%, remaining suppressed through delivery, and these values were
significantly different from those values observed in women with
pre-eclampsia.
In normotensive women, IGF-I concentrations were positively correlated with osteocalcin concentrations in the third trimester (r = 0.27; P < 0.0001). IGF-I values and osteocalcin concentrations were correlated in women with pre-eclampsia at entry to care and the third trimester [correlations of 0.48 (P < 0.01) and 0.43 (P < 0.001), respectively].
As seen in Table 2, the association of
IGF-I concentrations, measured across pregnancy, was significantly
different in women with pre-eclampsia compared with normotensive women.
Fig. 1 graphically depicts the
longitudinal model of the relationship between the patterns of
osteocalcin concentrations and IGF-I concentrations during pregnancy in
the two groups. Normotensive women have a relatively constant
relationship between osteocalcin and IGF-I, irrespective of IGF-I
concentrations at entry to care, whereas osteocalcin concentrations are
more responsive to the higher IGF-I concentrations at entry to care in
women who will develop pre-eclampsia. There was an interaction of a
positive relationship between IGF-I and osteocalcin concentrations
across pregnancy with a pattern changing slightly more positively with
each succeeding trimester, after adjusting for age. In contract, women
who would develop pre-eclampsia had almost a linear rise in the ratio
of IGF-I to osteocalcin, instead of the more curvilinear relationship.
This would indicate that there is a greater unit rise in osteocalcin
per unit change of IGF-I among women with pre-eclampsia at IGF-I
concentrations greater than 150 ng/ml.
|
|
). Although changes in IGF-I were associated with bone
ultrasound change among women with pre-eclampsia, the amount of bone
change was not statistically significant between the two groups. There
were no associations of IGF-I concentrations measured in pregnancy and
DEXA measures of BMD among normotensive or women with pre-eclampsia
women. Discussion
We prospectively investigated the relationship between serum IGF-I levels and a measure of bone turnover (osteocalcin and its change) as well as bone (bone ultrasound and delivery DEXA BMD) and bone change by ultrasound in pregnant women. The analysis yields important contributions to the literature. First, there are few studies that have described IGF-I concentrations across the course of pregnancy. In this study, the first trimester concentrations of IGF-I were only slightly lower than those reported by Naylor et al. (13). Concentrations in normotensive women in our study rose in the third trimester to a value above the first trimester value, also observed by Naylor et al. (13). Other cross-sectional studies have also reported higher IGF-I concentrations in the third trimester (12), probably reflecting placental contribution.
In this population with 64 women characterized as having pre-eclampsia, there were no significant mean differences in IGF-I concentrations between normotensive women and pre-eclampsia at entry to care. The average IGF-I concentrations, by trimester, measured in Camden study women with pre-eclampsia were significantly greater than concentrations observed during the second and third trimesters of pregnancy in normotensive pregnant women. These findings were inconsistent with the data of Guidace et al. (14) who compared 16 patients measured at a single point-in-time between 20 and 34 weeks with 29 controls, as well as the data by Halhali et al. (15) who compared IGF-I concentrations in 24 women at delivery with 24 control women. However, the Camden study data were internally consistent, biologically, based on the role of IGFs and bone turnover, as measured by osteocalcin concentrations. In this study population, women with pre-eclampsia had higher mean osteocalcin concentrations and higher IGF-I concentrations compared with normotensive women. Although there was greater bone ultrasound change among women with pre-eclampsia, the amount of change was not statistically significant.
Collectively, this is consistent with the hypothesis of an environment in which there is an increasing demand for calcium mobilization from maternal bone stores; however, studies have not identified those signal(s) that are mediating the need for this increased bone mobilization. It has been reported that IGF-I concentrations are lower in persons with osteoporosis (17, 18), but it is uncertain whether the environment of an aging woman with osteoporosis is informative for or consistent with the environment of a pregnant woman, especially with pre-eclampsia. Nor have these osteoporosis studies investigated the role of IGF-I longitudinally with respect to bone change. However, one could speculate that IGF-I may be a part of this process.
There are a number of explanations for the differential IGF-I to osteocalcin patterns according to pregnancy status. One explanation for the higher IGF-I concentrations among women with pre-eclampsia could be an increased proteolysis of IGF binding proteins (IGFBPs) (19, 20). During pregnancy, partial proteolysis occurs in proportion to circulating IGFBP-3, one of the two major binding proteins associated with IGF-I, and resulting in increased bioavailable IGF-I (19, 20). However, the concentrations of IGFBPs, including IGFBP-3, were not measured in this study. Furthermore, a signal that might initiate a differentially accelerated proteolysis with pre-eclampsia has not been identified. Alternatively, Halhali et al. (15) suggested that women with pre-eclampsia may reflect an IGF-I resistance state similar to that reported by Berkowitz (21) with respect to insulin.
Any interpretation of the findings must be mindful of the differential volume expansion that occurs in normal pregnancy compared with the pre-eclamptic pregnancy. To assure ourselves that the higher IGF-I levels in pre-eclamptic women over those of normotensive women were not an artifact of this volume expansion, we evaluated the data using two different approaches to adjustment for volume expansion. First, weight and height, measured at examination in each trimester, were used to calculate plasma volume for each enrollee using regression equations from Whittaker and Lind (22) derived from their administration of Evans blue dye during pregnancy. Second, we used the difference in the hematocrit measures at entry to care and the 28-wk evaluation, indexed to the woman’s surface area that was derived from measures of height and weight (23). Because neither approach altered the core findings, we presented the unadjusted data.
The finding of a negative correlation between age and IGF-I levels in all three trimesters is consistent with Nicolas et al. (11), who concluded that there was an age-related decline in IGF-I for both men and women. It is also widely recognized that younger age is related to increased risk of pre-eclampsia. It is difficult to consider age as a confounder from a statistical point of view in this study because age may be a surrogate for an integral part of the metabolic pathways associated with pre-eclampsia. If age reflects the increased demands of a metabolic environment to be sustained by the youthful mother and her fetus, it is logical that age will be inversely associated with IGF-I concentrations as well as osteocalcin concentrations.
Information in this area is limited; relevant studies of the effects in pregnancy are limited in number and sample size, frequently speculate on change in biology based on a single point-in-time observations, and are generally inconclusive as to the dynamics of the relationship during pregnancy. The Camden study permitted us to differentiate women with pre-eclampsia from women with normotensive pregnancies in a population that did not have the selection factors associated with obstetrical research among referrals to tertiary care centers. The ability to define these subgroups within the context of the study population was crucial because this and other studies have indicated that IGF-I concentrations in women with pre-eclampsia are significantly different from IGF-I concentrations in normotensive women, although the direction is not consistent (14, 15). The Camden study also has relatively larger samples of both women with pre-eclampsia and normotensive pregnant women that improve statistical power, resulting in the better identification of differences if they indeed exist. The diversity of the study participants enables consideration of the effects of ethnicity and race, if any, on the relationship between IGF-I and BMD and bone turnover.
It is also important to identify limitations in the study. First, osteocalcin was the only measure of bone turnover. It might have been very useful to include a marker of bone resorption to determine the association of IGF-I concentrations with respect to the bone coupling process. However, there was concern that the more commonly used markers of bone resorption, which are urinary assays, would be compromised during pregnancy and at delivery when there are marked renal and urinary volume changes. Furthermore, there was also no measure of IGF-I binding proteins so that one could assess the role of a more bioavailable IGF-I.
In summary, this study indicated that there were different relationships between IGF-I and osteocalcin concentrations during pregnancy in women who had pre-eclampsia compared with normotensive pregnant women. Furthermore, there was a greater ultrasound bone loss among the women (albeit not statistically significant) with pre-eclampsia that may be related, in part, to the differential relationship between IGF-I and osteocalcin. This prospective observational study suggests that in the first trimester among women who will subsequently be classified as having pre-eclampsia, there are metabolic changes usually associated with increasing demand for calcium mobilization as exemplified by greater IGF-I and osteocalcin concentrations and loss from maternal bone stores; however, this study does not identify those signals that are mediating the need for this increased mobilization.
Acknowledgments
We acknowledge Becton Dickinson and Co. (Franklin Lakes, NJ) for providing special blood collection containers for collection of cord blood samples; and Our Lady of Lourdes Osborn Hospital, St. John’s Hospital, and Kennedy Malcarney Hospital in Camden, New Jersey, for access to their facilities. Most of all, we thank the young women of Camden who participated in this study.
Footnotes
This work was supported by NIH Grants ES-07437 (to M.F.S. and T.S.) and HD18269 (to T.S.).
Abbreviations: BMD, Bone mineral density; DEXA, dual x-ray densitometry; QUI, quantitative ultrasound index.
Received June 12, 2001.
Accepted September 5, 2001.
References
This article has been cited by other articles:
 |
|
 |
|
  K. O O'Brien, C. M Donangelo, C. L V. Zapata, S. A Abrams, E M. Spencer, and J. C King Bone calcium turnover during pregnancy and lactation in women with low calcium diets is associated with calcium intake and circulating insulin-like growth factor 1 concentrations Am. J. Clinical Nutrition, February 1, 2006; 83(2): 317 - 323. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Fuglsang, F. Lauszus, A. Flyvbjerg, and P. Ovesen Erratum J. Clin. Endocrinol. Metab., October 1, 2003; 88(10): 5042 - 5042. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Fuglsang, F. Lauszus, A. Flyvbjerg, and P. Ovesen Human Placental Growth Hormone, Insulin-Like Growth Factor I and -II, and Insulin Requirements during Pregnancy in Type 1 Diabetes J. Clin. Endocrinol. Metab., September 1, 2003; 88(9): 4355 - 4361. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
