Prospective observational study of bone mineral density during pregnancy: low molecular weight heparin versus control

A.J. Carlin1, R.G. Farquharson1,3, S.M. Quenby1, J. Topping1 and W.D. Fraser2

1 Department of Obstetrics and Gynaecology, Liverpool Women’s Hospital, Crown Street, Liverpool L8 7SS and 2 Department of Clinical Biochemistry, Royal Liverpool University Hospital, Prescot Street, Liverpool L7 8XP, UK

3 To whom correspondence should be addressed. e-mail: rgfarquharson{at}lwh-tr.nwest.nhs.uk


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
BACKGROUND: The aim of this study was to assess the effect of long-term low molecular weight heparin (LMWH) on bone mineral density (BMD) during pregnancy. METHODS: Fifty-five patients with recurrent miscarriage and known thrombophilia (antiphospholipid syndrome) were followed through pregnancy in an ethically approved prospective observational study. All women had dual energy X-ray absorptiometry (DEXA) scans at the lumbar spine (L1–L4) performed within 6 months prior to conception and in the immediate post-natal period, within 6 weeks of delivery. LMWH (5000 U/day) plus low-dose aspirin was commenced after a positive urine pregnancy test and continued throughout pregnancy and after delivery until 6 weeks post-partum. A group of 20 volunteers with recurrent miscarriage, not requiring any treatment intervention, acted as controls and were monitored in an identical fashion. RESULTS: Characteristics were not significantly different between treated patients and controls. Both groups showed a similar loss in lumbar spine (L1–L4) BMD by the end of the pregnancy [LMWH 4.17% or 0.045 g/cm3, 95% confidence interval (CI) 0.036–0.062 versus control 3.56% or 0.043 g/cm3, 95% CI 0.027–0.059]. However, the difference in bone loss between the groups was not statistically significant (0.002 g/cm3, CI –0.0124 to 0.00865; P = 0.88). No patient suffered vertebral fracture. CONCLUSIONS: Bone loss associated with the use of long-term LMWH is not significantly different from physiological losses during pregnancy.

Key words: bone/heparin/thrombophilia


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
There has been a steady rise in the use of heparin in obstetrics. The main indications are thromboprophylaxis in cases of previous venous thromboembolism (Greer and de Swiet, 1993Go; Barbour and Pickard, 1995Go) or prothrombotic states, e.g. the antiphospholipid syndrome (Granger and Farquharson, 1997Go; Rai et al., 1997Go). The most recent Confidential Enquiry into Maternal Deaths has revealed that following the introduction of the Royal College of Obstetricians and Gynaecologists guidelines on thromboprophylaxis there has been a dramatic fall in the number of deaths related to venous thromboembolism (Lewis, 2002Go).

One concern is the reported association of chronic heparin use in pregnancy and osteoporosis. Several studies have suggested that heparin can lead to a reduction in bone mineral density (BMD) (de Swiet et al., 1983Go; Dahlman et al., 1990Go; Barbour, 1994Go; Douketis et al., 1996Go) and, rarely, osteoporotic fractures (Squires and Pinch, 1979Go; Dahlman, 1993Go; Haram et al., 1993Go). These earlier studies were uncontrolled, retrospective or measured BMD after several months of lactation. Inaccurate and poorly reproducible methods of assessment, such as plain X-rays, have also been used (Dahlman et al., 1990Go).

We designed a prospective, observational study of BMD commencing preconceptually, comparing the effects of long-term low molecular weight heparin (LMWH) in pregnancy with an untreated control group.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Patients
Volunteers were recruited following preconceptual assessment in the Recurrent Miscarriage Clinic at Liverpool Women’s Hospital. The patients gave verbal consent and study and ethical approval was granted by the Liverpool Research Ethics Committee.

Inclusion criteria
Women were included if they consented to the study and were able to return to the hospital for the post-natal BMD measurement. All patients had a confirmed history of three or more consecutive miscarriages. Those in the LMWH group had antiphosphopholipid syndrome (APS), prolonged dilute Russells viper venom test and/or raised anti-cardiolipin IgG/IgM antibody. Women with idiopathic recurrent miscarriage not requiring LMWH provided the control group, as they had similar obstetric histories to the group with APS. Only women delivering at gestations ≥34 weeks were included.

Exclusion criteria
Women were excluded if they had a pre-existing metabolic/endocrine disease or were taking medications known to affect bone metabolism. Women delivering prior to 34 weeks’ gestation were also excluded.

LMWH was started when urinary pregnancy test was positive, at <8 weeks gestation in all cases, and continued until 6 weeks post-partum. The treated group received LMWH (Dalteparin) at a prescription dose of 5000 U/day by subcutaneous injection and oral low-dose aspirin (75 mg day). The control group received no treatment and were monitored in exactly the same way.

No patient had pre-existing metabolic or endocrine disease.

BMD
BMD of the lumbar spine (L1–L4) was measured in g/cm3 using dual-energy X-ray absorptiometry (DEXA; QDR 4500 Hologic; Hologic Inc., Bedford, MA, USA) within 6 months of conception at the time of regular menstruation with gonadal shielding. This assessment was repeated post-natally within 6 weeks of delivery.

Successive analyses involved the same bone as the original scan, allowing direct comparison between successive scans, with an accuracy of <0.5% for the lumbar spine (Laskey et al., 1991Go). Daily quality control is performed for the DEXA scanner using a phantom bone block.

Data analysis
The Shapiro–Francia test (Altman, 1991Go) was used to test for normality and found suitable for the Student’s t-test for paired data within the groups. Differences between the groups were estimated by Student’s t-test for unpaired data.

P < 0.05 was considered to be statistically significant.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Fifty-five patients using LMWH and 20 control patients were studied. The LMWH patients had a mean age of 32.5 years, compared with 33.8 years for the control group. There was no significant difference between the two groups in preconceptual BMD, and both groups had a comparable number of previous miscarriages and livebirths (for demographics see Table I).


View this table:
[in this window]
[in a new window]
 
Table I. Mean (range) demographic and clinical characteristics of the LMWH and control groups
 
There was a significant decrease in BMD at the lumbar spine in both the LMWH [mean decrease 0.045 g/cm3 or 4.17%; 95% confidence interval (CI) 0.036–0.062; P < 0.001] and control group (mean decrease 0.043 g/cm3 or 3.56%; 95% CI 0.027–0.059; P < 0.001) (Table II).


View this table:
[in this window]
[in a new window]
 
Table II. BMD at baseline and immediately post-partum in LMWH and control groups
 
No significant difference in BMD changes was detected between the LMWH and control groups (difference 0.002 g/cm3; 95% CI –0.0085 to 0.123; P = 0.88) (Table III).


View this table:
[in this window]
[in a new window]
 
Table III. BMD changes during pregnancy in women who have received LMWH versus control
 
All patients conceived within 6 months of the first DEXA scan and the second scan was performed within 6 weeks of delivery; therefore, all had completed their pregnancy between successive scans. Only those patients delivering at ≥34 weeks were included in the final analysis. This meant that the LMWH group received between 34 and 42 weeks of heparin prophylaxis.

All women completing the study delivered live infants and none had pregnancies complicated by venous thromboembolism, thrombocytopenia or any other significant side effects.


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Our results suggest that long-term LMWH treatment throughout pregnancy and the puerperium does not significantly affect BMD compared with normal physiological change.

We have completed the first prospective, observational study of BMD commencing preconceptually, comparing the effects of long-term LMWH in pregnancy with an untreated control group. This facilitates standardization of errors and avoids the problem of baseline measurements being taken after first trimester (Barbour et al., 1994Go; Dahlman et al., 1994Go; Backos et al., 1999Go), when physiological BMD losses may already have occurred (Purdie et al., 1988Go; Black et al., 1996Go; 2000Go). Accuracy was optimized by concentrating on the lumbar spine, where the co-efficient of variation (short- and long-term precision) of DEXA is <0.5%, compared with <1.5% at the femoral neck (Laskey et al., 1991Go). These levels of precision assume great significance when dealing with the small changes in BMD observed during and after pregnancy. The lumbar spine is also composed of trabecular bone, which appears to be more sensitive to the effects of both heparin and pregnancy (Dahlman et al., 1994Go). This overall approach has improved reproducibility, reduced radiation exposure and eliminated additional risk to the developing fetus by avoiding antenatal assessments.

We found a significant decrease in BMD at the lumbar spine in both the LMWH (4.17%) and the untreated (3.56%) group. However, the difference in BMD loss between the two groups was not significant, confirming that LMWH does not exert stresses on the pregnant skeleton greater than those caused by normal pregnancy. Our results are similar to studies that have used DEXA to prospectively assess BMD changes at the lumbar spine in LMWH treated pregnancies. A small prospective study of 17 LMWH patients and eight controls during pregnancy showed a lumbar spine loss of 5.2% in the heparin thromboprophylaxis group compared with a 3.1% reduction in the control group (Shefras and Farquharson, 1996Go). More recently, a larger prospective, longitudinal study found a 3.7% decrease in BMD at the lumbar spine (Backos et al., 1999Go). These authors compared bone density changes in pregnant women receiving aspirin and either unfractionated heparin or LMWH, and found no significant difference in BMD change between the two groups at the three sites studied (forearm, femoral neck and spine); however, the study was not controlled for untreated patients. The authors decided to set 12 weeks’ gestation as baseline for their study, which means that they may have missed the previously reported bone loss that occurs early in the first trimester (Black et al., 1996Go; 2000Go), and thus may have underestimated the true bone loss. This issue was partially addressed by measuring BMD in a small cohort of nine treated women prepregnancy (aspirin and heparin) and comparing this with BMD at 12 weeks’ gestation. Surprisingly, their results showed that the mean BMD increased by 1.8% (P = 0.01) at the lumbar spine and by 1.1% (not significant) at the femoral neck, contradicting the earlier studies (Black et al., 1996Go; 2000Go).

Idiopathic recurrent miscarriage patients were selected for controls because they have similar obstetric histories. In our study there was a large discrepancy in total numbers between the LMWH (n = 55) and the control (n = 20) group. At the time of recruitment, fewer control group patients were eligible for entry into the study, and this is reflected in the differing sizes of the two groups. However, once the recruitment period had passed and the disparity in numbers was clear, it would have been difficult to balance the groups without incorporating selection bias. Demographic data are included in this study (Table I), although previous multiple regression analysis has demonstrated that age, BMI, ethnic origin, smoking status, previous alcohol intake, previous number of miscarriages and live births, time elapsed since last pregnancy, baseline BMD, and length of heparin treatment are not predictive of BMD changes during pregnancy (Backos et al., 1999Go). We do not have full breast-feeding data because of the difficulties involved in collecting this information from such a geographically widespread population, predominantly referred for tertiary level care. However, as all post-partum BMD assessments were made within 6 weeks of delivery, and in the majority of cases within several days, the effect of such a short period of lactation is presumed to be negligible.

Long-term heparin therapy during pregnancy has been associated with the development of osteoporosis and vertebral fractures. However, retrospective studies lack preconceptual BMD data and may include those who are already osteopenic (de Swiet et al., 1983Go; Dahlman et al., 1990Go; Douketis et al., 1996Go), while other workers have taken post-partum assessments after long periods of lactation (de Swiet et al., 1983Go; Nelson-Piercy et al., 1997Go), which is known to cause bone loss of up to 10% within 6 months (Sowers et al., 1993Go; Kalkwarf and Specker, 1995Go). The earlier studies also relied upon less accurate and poorly reproducible methods of asssessment, such as plain X-rays (de Swiet et al., 1983Go; Dahlman et al., 1990Go; Dahlman, 1993Go), and all were uncontrolled.

A contemporaneous, but non-sequential, study of bone biopsy analysis in the first and third trimester of pregnancy showed clear evidence that the bone response to pregnancy is biphasic, with early bone resorption and later bone formation (Purdie et al., 1988Go). A prospective observational study starting preconceptually yielded further discriminatory changes in maternal bone metabolism (Black et al., 1996Go; 2000Go). These authors showed further proof of the biphasic uncoupling of bone resorption (starting early first trimester) with later bone formation (starting in third trimester), with subsequent reduction in BMD of the maternal spine in normal untreated pregnancy. These studies highlight the importance of preconceptual assessment, as baseline measurements made only in the first trimester may underestimate true bone loss.

Some individuals may be at increased risk of developing osteopenia or osteoporosis, and they may either enter pregnancy having failed to achieve peak bone mass or suffer an exaggerated or inappropriate physiological response to pregnancy. In the latter group there may be greater resorption in the first and second trimester than would normally be expected, and/or the anticipated bone formation associated with the third trimester is either delayed or deficient.

The overall controlling factors involved in bone homeostasis during pregnancy are as yet unknown, but as more information becomes available it may soon be possible to identify those individuals at most risk of severe demineralization.

The overall risk of developing osteoporosis in pregnancy is probably low, but heparin and pregnancy may both create an abnormal stress load in a susceptible individual. This theory is supported by an epidemiological study which showed that mothers of patients with idiopathic osteoporosis of pregnancy had a higher prevalence of fractures compared with controls (Dunne et al., 1993Go). Further studies are required to assess the long-term implications of bone loss during both normal and heparin-treated pregnancies.

Although there are no proven differences between LMWH and unfractionated heparin with regard to bone density in pregnancy, the advantages of simpler dosing regimes, less bleeding complications and a lower incidence of thrombocytopenia (Warkentin et al., 1995Go) make LMWHs the treatment of choice for both thromboprophylaxis and the management of venous thromboembolism.

In conclusion, our study reinforces the view that LMWH does not reduce lumbar spine BMD significantly below that seen in normal pregnancy


    Acknowledgements
 
We would like to thank all the women who agreed to participate in this study and the Radiology Department at the Liverpool Women’s Hospital for performing the DEXA scans. We also thank Craig Cowan for his assistance with data analysis. The authors declare no conflicts of interest.


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Altman DG (1991) Practical Statistics for Medical Research. Chapman and Hall, London, UK, pp. 291–292.

Backos M, Rai R and Thomas E (1999) Bone density changes in pregnant women treated with heparin: a prospective, longitudinal study. Hum Reprod 14,2876–2880.[Abstract/Free Full Text]

Barbour LA and Pickard J (1995) Controversies in thromboembolic disease during pregnancy: a critical review. Obstet Gynaecol 86,621–633.[Abstract/Free Full Text]

Barbour LA, Kick SD, Steiner JF, LoVerde ME, Heddleston LN, Lear JL, Baron AE and Barton PL (1994) A prospective study of heparin induced osteoporosis in pregnancy using bone densitometry. Am J Obstet Gynecol 170,862–869.[Medline]

Black AJ, Topping J, Farquharson, RG and Fraser WD (1996) Bone metabolism in pregnancy. Contemp Rev Obstet Gynaecol 8,192–196.

Black AJ, Topping J, Durham B, Farquharson RG and Fraser WD (2000) A detailed assessment of alterations in bone turnover, calcium homeostasis and bone density in normal pregnancy. J Bone Miner Res 15,557–563.[Medline]

Lewis G (ed.) (2002) Confidential Enquiries into Maternal Deaths in the United Kingdom—Fifth Report. RCOG Press, London, UK.

Dahlman TC (1993) Osteoporotic fractures and the recurrence of thromboembolism during pregnancy and puerperium in 184 women undergoing thromboprohylaxis with heparin. Am J Obstet Gynecol 168,1265–1270.[Medline]

Dahlman TC, Lindvall N and Hellgren M (1990) Osteopenia in pregnancy during long-term heparin treatment: a radiological study. Br J Obstet Gynaecol 97,221–228.[Medline]

Dahlman TC, Sjoberg HE and Ringertz H (1994) Bone mineral density during long-term prophylaxis with heparin in pregnancy. Am J Obstet Gynecol 170,1315–1320.[Medline]

deSwiet M, Ward PD, Fidler J, Horsman A, Katz D, Letsky E, Peacock M and Wise PH (1983) Prolonged heparin therapy in pregnancy causes bone demineralisation. Br J Obstet Gynaecol 90,1129–1134.[Medline]

Douketis JD, Ginsberg JS, Burrows RF, Duku EK, Webber CE and Brill-Edwards P (1996) The effects of long-term heparin therapy during pregnancy on bone density. A prospective matched cohort study. Thromb Haemost 75,254–257.[Medline]

Dunne F, Walters B, Marshall T and Heath D (1993) Pregnancy-associated osteoporosis. Clin Endocrinol 39,487–490.[Medline]

Granger K and Farquharson RG (1997) Obstetric outpatient in Antiphospholipid syndrome. Lupus 6,509–513.[Medline]

Greer I and de Swiet M (1993) Thrombosis prophylaxis in obstetrics and gynaecology. Br J Obstet Gynaecol 100,37–40.[Medline]

Haram K, Hervig T, Thordarson H and Aknes L (1993) Osteopenia caused by heparin treatment in pregnancy in pregnancy. Acta Obstet Gynecol Scand 72,674–675.[Medline]

Kalkwarf H and Specker B (1995) Bone mineral density loss during lactation and recovery after weaning. Obstet Gynecol 86,26–32.[Abstract/Free Full Text]

Laskey MA, Flaxman ME, Barber RW, Trafford S, Hayball MP, Lyttle KD, Crisp AJ and Compston JE (1991) Comparative performance in vitro and in vivo of Lunar DPX and Hologic QDR 1000 Dual Energy Xray Absorptiometry. Br J Radiol 64,1023–1029.[Abstract]

Nelson-Piercy C, Letsky E and de Swiet M (1997) Low molecular weight heparin for obstetric thromboprophylaxis: experience of sixty-nine pregnancies in sixty one women at high risk. Am J Obstet Gynecol 176,1062–106.[Medline]

Purdie DW, Aaron JE and Selby PL (1988) Bone histology and mineral homeostasis in human pregnancy. Br J Obstet Gynaecol 95,849–854.[Medline]

Rai R, Cohen H, Dave M and Regan L (1997) Randomised controlled trail of aspirin and aspirin plus heparin in pregnant women with recurrent miscarriage associated with phospholipid antibodies (or antiphospholipid antibodies). BMJ 314,253–257.[Abstract/Free Full Text]

Shefras J and Farquharson RG (1996) Bone density studies in pregnant women receiving heparin. Eur J Obstet Gynecol 65,171–174.[CrossRef][Medline]

Sowers M, Corton G, Shapiro B, Jannausch ML, Crutchfield M, Smith ML, Randolph JF and Hollis B (1993) Changes in bone density with lactation. JAMA 269,3130–3135.[Abstract]

Squires JW and Pinch LW (1979) Heparin-induced spinal fractures. JAMA 241,2417–2418.[Abstract]

Warkentin TE, Levine MN, Hirsh J, Horsewood P, Roberts RS, Gent M and Kelton JG (1995) Heparin-induced thrombocytopenia in patients treated with low molecular weight heparin or unfractionated heparin. N Engl J Med 332,1330–1335.[Abstract/Free Full Text]

Submitted on February 4, 2003; resubmitted on November 3, 2003; accepted on November 11, 2003.





This Article
Abstract
Full Text (PDF )
All Versions of this Article:
19/5/1211    most recent
deh115v1
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (2)
Request Permissions
Google Scholar
Articles by Carlin, A.J.
Articles by Fraser, W.D.
PubMed
PubMed Citation
Articles by Carlin, A.J.
Articles by Fraser, W.D.