- Email: [email protected]
Bone density changes in women who receive thromboprophylaxis in pregnancy
American Journal of Obstetrics and Gynecology (2006) 195, 1109–13
www.ajog.org
Bone density changes in women who receive thromboprophylaxis in pregnancy Holly Casele, MD,a,b Elaine I. Haney, BS, MT,b Andra James, MD,c Karen Rosene-Montella, MD,d Michael Carson, MDe Northwestern University Feinberg School of Medicine,a Chicago IL; Department of Obstetrics and Gynecology, Division of Maternal Fetal Medicine Evanston Northwestern Healthcare,b Evanston, IL; Department of Obstetrics and Gynecology, Division of Maternal Fetal Medicine, Duke University Medical School,c Durham, NC; Departments of Obstetric and Consultative Medicine and Medicine, Brown University Medical School, Women and Infants Hospital,d Providence, RI ; Department of Internal Medicine, St. Peters University Hospital,e New Brunswick, NJ Received for publication March 10, 2006; revised June 14, 2006; accepted June 29, 2006
KEY WORDS Bone mineral density Low molecular weight heparin therapy
Objective: The purpose of this study was to compare unfractionated heparin therapy to the low molecular weight heparin, enoxaparin sodium, and their effects on bone mineral density over the course of pregnancy. Study design: Pregnant patients whose condition required thromboprophylaxis were recruited in this prospective randomized controlled trial and were assigned to receive either unfractionated heparin therapy or low molecular weight heparin therapy. Bone mineral density was measured by dual energy x-ray absorptiometry at the proximal femur on enrollment and again shortly after delivery. Results: One hundred twenty women were enrolled, and 98 women completed the study. There was no difference in the change in bone mineral density at the femoral neck (P = .054) or total proximal femur (P = .584) between groups. Only 1 of 40 patients (2.5%) who received unfractionated heparin therapy and 1 of 49 patients (2.0%) who received low molecular weight heparin therapy (P = 1.0) experienced bone loss of R10% at the femoral neck. Conclusion: In this study, the incidence of clinically significant bone loss (R10%) in the femur in women who received thromboprophylaxis in pregnancy is approximately 2% to 2.5% and appears to be similar, regardless of whether the patient receives low molecular weight heparin therapy or unfractionated heparin therapy. Ó 2006 Mosby, Inc. All rights reserved.
Antithrombotic therapy is used during pregnancy for a variety of indications. Because warfarin sodium (Coumadin) can have adverse effects on the fetus, Presented at the 26th Annual Meeting of the Society for Maternal Fetal Medicine, January 30-February 4, 2006, Miami, FL. Reprints not available from the authors. 0002-9378/$ - see front matter Ó 2006 Mosby, Inc. All rights reserved. doi:10.1016/j.ajog.2006.06.080
unfractionated heparin (UH) and low molecular weight heparins (LMWH) are currently the preferred options. A recognized complication of chronic heparin therapy is osteoporosis. In 1965, Griffith et al1 identified 6 patients with spontaneous rib or vertebral fractures from a cohort of 10 nonpregnant women who had received large doses (15,000-30,000 units daily) of UH therapy for
1110 R6 months. Since then, there have been additional case reports of osteoporosis or fractures developing in patients who have received long-term heparin treatment.2-6 In 1993, Dahlman7 reported a symptomatic vertebral fracture rate of 2.2% in 184 pregnant women who received 13,000 to 40,000 units of UH daily (mean, 19,100 IU/d). Although 1 of 36 patients who received low-dose heparin therapy experienced a fracture (0.7%), there was a relationship between the incidence of fracture and the dose of heparin used.7 In another study that used single photon absorptiometry to measure bone mineral density (BMD) of the distal forearm at the start of treatment and within 1 week after delivery, Dahlman et al found a 4.9% reduction in trabecular bone in 39 pregnant women who had received a mean UH dosage of 17,300 U/d for a mean of 25 weeks and no such reduction in control subjects.8 Similarly, Barbour et al9 found a 5% decline in BMD at the proximal femur that was measured by dual energy x-ray absorptiometry (DEXA) in 14 pregnant women who received a mean UH dosage of 27,000 IU/d for 21 weeks; again, there was no such change in control subjects. In their study, 5 of 14 patients (36%) experienced R10% bone loss at the proximal femur.9 Several case series and animal studies had suggested that the bone loss with LMWH therapy is less than that seen with UH therapy.10-12 We performed a pilot study of 16 women who received 40 mg daily of enoxaparin for a mean of 25 weeks and found that no patient experienced bone loss of R10% that was measured at the femoral neck 6 weeks after delivery, and there was no significant change in BMD at the proximal femur overall.13 On the basis of these previous studies, we hypothesized that patients who receive thromboprophylaxis with enoxaparin sodium would experience less bone loss at the proximal femur than would patients who receive UH therapy.
Methods A multicenter prospective randomized controlled trial was performed at 9 centers from September 1998 until December 2005: Evanston Northwestern Healthcare, Duke University Medical School, Brown University Women and Infant’s Hospital, St. Peter’s University Hospital, Bridgeport Hospital, Prentice Women’s Hospital, University of Alabama, University of North Carolina, and Ohio State University. The Institutional Review Boards at all centers approved the study before patient recruitment at that site. Any patient who was a candidate for low-dose thromboprophylaxis for the duration of her pregnancy, who was at least 18 years old, and who could begin therapy at !24 weeks of gestation was eligible to participate. Patients were excluded only if they had a contraindication to
Casele et al anticoagulation. After informed consent was obtained, patients were assigned randomly to receive either UH or enoxaparin sodium therapy. Randomization was accomplished with a random number table, with each site stratified into blocks of 10. Baseline and serial platelet counts were obtained. All patients had a hypercoagulable evaluation. Patients in the UH arm self-administered 7500 units twice daily of subcutaneous heparin sodium from enrollment in the study until 28 weeks of gestation and then 10,000 units twice daily until delivery. Patients in the LMWH therapy arm self-administered 30 mg twice daily of subcutaneous enoxaparin sodium from enrollment until 28 weeks of gestation and then 40 mg twice daily until delivery. All patients received adjusted-dose Coumadin for 6 to 8 weeks after delivery. In addition to a prenatal vitamin, all patients were asked to take calcium supplementation of 500 mg daily from the time of enrollment until delivery. Bone density of the proximal femur was measured serially with DEXA. Measurements of the femoral neck and total hip bone density were made at baseline before the beginning of therapy or after 12 weeks of gestation. Bone density was measured again within 4 days following the patient’s delivery. Radiologists who were blinded to treatment arm interpreted all of the DEXA scans. Before the patients were enrolled, sample size calculations were performed. Our study design resembled that of Barbour et al,9 because we used serial DEXA measurements to determine the change in BMD at the femur as the primary outcome. We hypothesized that at least 30% of the UH group would experience a R10% decrease in BMD. To have 80% power to show a 50% reduction in the proportion of patients in the enoxaparin group (15%) who would experience the same degree of bone loss, we would need to recruit 240 patients. To show a 3-fold reduction to 10% in the enoxaparin group, we would need 124 patients; to show a 6-fold reduction to 5%, we would need 72 patients. We obtained approval for recruiting a total of 240 patients with a plan for a full interim analysis. All adverse events were reported to the Food and Drug Administration, Sanofi-Aventis, and the institutional review board at Evanston Northwestern Healthcare, where completed data collection sheets and copies of DEXA scans were sent and maintained by the research coordinator. The study was overseen by a data and safety monitoring group, which was comprised of 3 physicians at Evanston Northwestern Healthcare who did not have any financial or professional interest in the study. The data and safety monitoring group met twice per year to review reported adverse events and to assure that patient risks were being minimized. Premature termination of the study would be advised by the data and safety monitoring group if there was a statistically significant rate of unexplained pregnancy loss,
Casele et al recurrent thrombosis, bleeding, or other adverse events in 1 group compared with the other. Other criteria for premature termination of the study included directions from the manufacturer or an interim analysis that indicated that either fewer subjects were needed to show the hypothesized effect or that so many more subjects were needed that the study would be considered unfeasible. Statistical analysis was performed with SPSS software (version 13 for Windows; SPSS Inc, Chicago, IL). The Student t test, Fisher exact test, chi-square test, analysis of variance, and regression were used as appropriate. A probability value of !.05 was considered significant. Results are reported as means G SD.
1111 Table I
Characteristics of patients
Variable
LMWH therapy (n = 61)
UH (n = 59)
P value
29.6 G 4.8
30.7 G 4.6
.193 .438
72 13 12 3 26.6 27.7 G 4.9
73 21 5 2 35.1 26.3 G 5.6
.423 .198
8 (13.3%)
14 (24.6%)
.16
Maternal age (y)* Race (%) White Black Hispanic Other Primiparity (%) Treatment duration (wk)* Early treatment withdrawal (n)
* Data are given as mean G SD.
Results We requested consent from 147 women; 27 women declined enrollment, and 120 women gave consent and were enrolled. Twenty-two women did not complete the study: 11 women experienced a spontaneous abortion; 7 women were noncompliant; 2 women had an allergic reaction to the medication, and 2 women were switched to therapeutic anticoagulation at the discretion of their referring physicians. We enrolled 98 women in 7 years who completed the study. An interim analysis revealed that the incidence of bone loss of R10% in patients who received UH therapy was 2.5%, not 36% as reported by Barbour et al.9 A sample size that was calculated on the basis of 3% as the incidence of significant bone loss indicated that a total of 1628 women would be necessary to detect a 50% decrease in the incidence for the LMWH group; this was not deemed feasible, and the study was stopped. The patients in the UH group received a mean dose of 17,380 units for a mean of 26.3 weeks; the patients in the LMWH group received a mean of 68.4 mg enoxaparin for a mean of 27.7 weeks (P = .198) for the treatment duration. The groups were comparable with regard to maternal age, race, and parity (Table I). When the changes in BMD at the femoral neck were compared, the LMWH group gained a mean of increase of 0.9%, bone mass, and the UFH group lost a mean of 0.9% (P = .071). When BMD changes at the total proximal femur were compared, the LMWH lost 1.5%, and the UFH group lost 1.1% (P = .616, Table II). Bone loss at the femoral neck of R5% occurred in 4 of the 49 patients (8.2%) in the LMWH group and in 9 of 40 of the patients (22.5%) in the UFH group (P = .073). Bone loss at the total femur of R5% occurred in 6 of the 43 patients (13.9%) in the LMWH group and in 4 of the 40 patients (10%) in the UH group (P = .74). Clinically significant bone loss at the femoral neck (defined as a decrease of R10%) occurred in 1 of the 49 patients (2.0%) in the LMWH group and in 1 of the 40 patients (2.4%)
of the UH group (P = 1.0). Bone loss of R10% in the total proximal femur occurred in 1 of the 43 patients (2.3%) in the LMWH group and in none of the 40 patients in the UH group (P = 1.0; Table II). There was no association between age, parity, or institution and the degree of bone loss within groups or between groups. There was also no correlation between age and mean bone loss at the femoral neck (P = .719) or at the total femur (P = .838). Mean bone loss by age groups is summarized in Table III. Gestational age at delivery, infant birthweight, rate of early treatment withdrawal, and rate of complications (spontaneous abortion, bleeding, and recurrent thrombosis) were not significantly different in the 2 groups (Table IV).
Comment Fragility fractures occur in approximately 1 million women in the United States every year, which is more than breast cancer, gynecologic cancer, myocardial infarction, and stroke combined.14 The direct care cost of osteoporotic fractures is O$17 billion per year.14 Pregnancy is a particular time of risk because the growing fetus derives all of its calcium from maternal sources and because fewer than 6% of women of childbearing age in the United States consume the recommended amount of calcium daily.15 Although many clinical factors (such as age, family history, body mass index, use of steroids, previous fracture, and propensity for falling) contribute to a patient’s risk for fracture, BMD determined by DEXA scanning has been determined to be an important predictor of a patient’s risk for fragility fracture.16 Antithrombotic therapy is indicated frequently in pregnancy. LMWH therapies have become popular, despite their higher cost because of the ease of use with prefilled syringes, more predictable dose response, greater bioavailability, and a possibly lower risk for complications. Bleeding complications and heparin-induced
1112 Table II
Casele et al Mean bone density changes in patients who underwent thromboprophylaxis
Variable
UH
LMWH therapy
P value
BMD change at neck (g/cm2)* % change at neck BMD change total (g/cm2)* % change total Loss O 10% at neck Loss O 10% total
ÿ0.010 G 0.054 (n = 40) 0.9% (n = 40) ÿ0.012 G 0.029 (n = 40) 1.1% 1/40 0/40
C0.009 G 0.040 (n = 49) 0.9% (n = 49) ÿ0.016 G 0.036 (n = 43) 1.5% 1/49 1/43 (2.3%)
.054 .071 .584 .616 1 1
In LMWH grp 49 pts had 2 measurements at neck, 43 pts had 2 total measurements. * Data are given as mean G SD.
Table III
Mean bone density changes by age
Table IV
Pregnancy outcome
2
Mean change (g/cm ) Age group (y)
N
BMD at femoral neck
BMD at total femur
19-24.9 25-29.9 30-34.9 35-39.9 40-44.9
11 30 34 12 3
ÿ1.536 0.523 0.659 ÿ0.142 ÿ2.567
ÿ0.2 ÿ1.614 ÿ1.523 0.683 ÿ1.267
Analysis of variance between groups: P = .579; P = .742.
thrombocytopenia do appear to occur less frequently with LMWH use than with UH use.17 However, the evidence for less osteoporosis with LMWH use is less convincing. The results of our study indicate that prophylactic doses of UH and enoxaparin, when given to pregnant women for prolonged periods in conjunction with calcium supplementation, do not appear to result in clinically significant changes in BMD at the proximal femur. In addition, there was no significant difference between BMD changes in women who received UH or LMWH therapy. Our results differ from many previous reports. We hypothesize that the differences that were observed are related primarily to our study design. First, we included only patients who were receiving prophylactic doses of UH and LMWH with a mean daily dose of UH of 17,380 units and of LMWH of 68.4 mg, whereas previous studies included patients who received both prophylactic and therapeutic dosing and therefore a higher mean daily dosing. In the study by Dahlman et al,8 not only did the authors report an overall 4.9% loss in BMD, but they also observed a correlation with the degree of bone loss and the dose of UH. It is conceivable that we did not observe significant bone loss because all of our patients received only prophylactic doses. However, because patients whose conditions required therapeutic anticoagulation were excluded from this study, we are able to hypothesize only that this is a possible explanation.
Variable Spontaneous abortion (n) Recurrent thrombosis (n) Bleeding at delivery (n) Estimated gestational age at delivery (wk)* Infant birthweight (g)*
LMWH therapy (n = 60) UH (n = 57)
P value
4 (6.7%)
7 (12.3%)
.354
2 (3.3%)
4 (7%)
.431
4 (6.7%)
1 (1.8%)
.365
37.5 G 4.0
37.0 G 5.0
.583
3197 G 536
3125 G 633 .552
* Data are given as mean G SD.
In addition, with the exception of the study by Barbour et al9 and our pilot study,13 no previous authors have reported that they recommended calcium supplementation to their patients who received UH therapy. In fact, the study by Dahlman et al8 specifically states that the patients did not take calcium. We recommended that, in addition to prenatal vitamins, all of the patients in both arms of the study should take 500 mg of calcium. We did not specify the brand or type of calcium supplement so that many patients may have also been consuming varying amounts of vitamin D because some supplements contain 200 to 400 IU of vitamin D. The recommended daily level of calcium consumption is 1000 mg; however, the average consumption by women of childbearing age is only 700 mg daily.18-20 Prenatal vitamins typically contain 125 to 250 mg of calcium.19 The importance of calcium supplementation in pregnant women in general cannot be overemphasized. If adequate calcium is not part of the maternal diet, bone may be degraded to provide calcium for maternal needs.19,20 So bone loss can occur with normal pregnancy if calcium intake is inadequate. In addition, calcium supplementation in pregnancy and lactation has been demonstrated to diminish maternal bone resorption.20
Casele et al The precise mechanism of heparin-induced osteoporosis in unclear and may involve parathyroid hormonelike effects, direct osteoclastic activity, and derangements in mineralization.9 It is also not known whether calcium supplementation mitigates the effects of UH or LMWH therapy on bone. In our study, all of the patients were instructed and encouraged to take calcium supplementation, but we did not assess calcium intake in the diet nor compliance with calcium supplementation. Although we hypothesize that it may have been beneficial to our patients, further study is necessary to test this hypothesis and to clarify this issue. Because bone loss can occur in normal pregnancy in women who do not consume enough calcium and because calcium supplementation does appear to diminish bone loss in these women, we believe strongly that all pregnant women, especially those who require antithrombotic therapy, should be encouraged to take calcium supplementation of at least 500 mg daily. In conclusion, for women who require only prophylactic doses of antithrombotic therapy and take calcium supplementation during pregnancy, neither UH or LMWH therapy appear to have clinically significant impact on BMD.
References 1. Griffith GC, Nichols G Jr, Asher JD, Flanagan B. Heparin osteoporosis. JAMA 1965;193:91-4. 2. Griffiths HT, Liu DTY. Severe heparin osteoporosis in pregnancy. Postgrad Med J 1984;60:424-5. 3. Murphy MS, John PR, Mayer AD, Buckels JA, Kelly DA. Heparin therapy and bone fractures. Lancet 1992;340:1098. 4. Sackler JP, Liu L. Heparin-induced osteoporosis. Br J Radiol 1973;46:548-50. 5. Squires JW, Pinch LW. Heparin induced spinal fractures. JAMA 1979;241:2417-8. 6. Wise PH, Hall AJ. Heparin-induced osteopenia in pregnancy. BMJ 1980;281:110-1.
1113 7. Dahlman T. Osteoporotic fractures and the recurrence of thromboembolism during pregnancy and puerperium in 184 women undergoing thromboprophylaxis with heparin. Am J Obstet Gynecol 1993;168:1265-70. 8. Dahlman TC, Sjoberg HE, Ringertz H. Bone mineral density during long-term prophylaxis with heparin in pregnancy. Am J Obstet Gynecol 1994;170:1315-20. 9. Barbour LA, Kick SD, Steiner JF, LoVerde ME, Heddleston LN, Lear JL, et al. A prospective study of heparin-induced osteoporosis in pregnancy using bone densitometry. Am J Obstet Gynecol 1994; 170:862-9. 10. Matzsch T, Bergqvist D, Hedner U, Nilsson B, Ostergaard P. Effects of low molecular weight heparin and unfractionated heparin on induction of osteoporosis in rats. Thromb Haemost 1990;63: 505-9. 11. Monreal M, Vinas L, Monreal L, Lavin S, Lafoz E, Angeles AM. Heparin-related osteoporosis in rats: a comparative study between unfractionated heparin and low molecular weight heparin. Haemostasis 1990;20:204-7. 12. Ginsberg JS, Kowalchuk G, Hirsh J, Brill-Edwards P, Burrow R, Coates G, et al. Heparin effect on bone density. Thromb Haemost 1990;64:286-9. 13. Casele HL, Laifer SA. Prospective evaluation of bone density in pregnant women receiving the low molecular weight heparin enoxaparin sodium. J Matern Fetal Med 2000;9:122-5. 14. Miller PD, Barlas S, Brenneman SK, Abott TA, Chen YT, BarrettConnor E, et al. An approach to identifying osteopenic women at increased short-term risk of fracture. Arch Intern Med 2004;164: 1113-20. 15. Weisman SM. The calcium connection to bone health across a woman’s lifespan. J Reprod Med 2005;50:879-84. 16. Bates DW, Black DM, Cummings SR. Clinical use of bone densitometry: clinical applications. JAMA 2002;288:1898-900. 17. Hirsh J, Levine MN. Low molecular weight heparin. Blood 1992; 79:1-17. 18. Gold DT. Elevated calcium requirements for women and unique approaches to improving calcium adherence. J Reprod Med 2005; 50:891-5. 19. March of Dimes. Calcium. Available at: www.marchofdimes. com/pnhec/159_9472.asp. Accessed January 23, 2006. 20. Janakiraman V, Ettinger A, Mercado-Garcia A, Hu H, HernandezAvila M, et al. Calcium supplements and bone resorption in pregnancy: a randomized crossover trial. Am J Prev Med 2003;24:260-4.
www.ajog.org
Bone density changes in women who receive thromboprophylaxis in pregnancy Holly Casele, MD,a,b Elaine I. Haney, BS, MT,b Andra James, MD,c Karen Rosene-Montella, MD,d Michael Carson, MDe Northwestern University Feinberg School of Medicine,a Chicago IL; Department of Obstetrics and Gynecology, Division of Maternal Fetal Medicine Evanston Northwestern Healthcare,b Evanston, IL; Department of Obstetrics and Gynecology, Division of Maternal Fetal Medicine, Duke University Medical School,c Durham, NC; Departments of Obstetric and Consultative Medicine and Medicine, Brown University Medical School, Women and Infants Hospital,d Providence, RI ; Department of Internal Medicine, St. Peters University Hospital,e New Brunswick, NJ Received for publication March 10, 2006; revised June 14, 2006; accepted June 29, 2006
KEY WORDS Bone mineral density Low molecular weight heparin therapy
Objective: The purpose of this study was to compare unfractionated heparin therapy to the low molecular weight heparin, enoxaparin sodium, and their effects on bone mineral density over the course of pregnancy. Study design: Pregnant patients whose condition required thromboprophylaxis were recruited in this prospective randomized controlled trial and were assigned to receive either unfractionated heparin therapy or low molecular weight heparin therapy. Bone mineral density was measured by dual energy x-ray absorptiometry at the proximal femur on enrollment and again shortly after delivery. Results: One hundred twenty women were enrolled, and 98 women completed the study. There was no difference in the change in bone mineral density at the femoral neck (P = .054) or total proximal femur (P = .584) between groups. Only 1 of 40 patients (2.5%) who received unfractionated heparin therapy and 1 of 49 patients (2.0%) who received low molecular weight heparin therapy (P = 1.0) experienced bone loss of R10% at the femoral neck. Conclusion: In this study, the incidence of clinically significant bone loss (R10%) in the femur in women who received thromboprophylaxis in pregnancy is approximately 2% to 2.5% and appears to be similar, regardless of whether the patient receives low molecular weight heparin therapy or unfractionated heparin therapy. Ó 2006 Mosby, Inc. All rights reserved.
Antithrombotic therapy is used during pregnancy for a variety of indications. Because warfarin sodium (Coumadin) can have adverse effects on the fetus, Presented at the 26th Annual Meeting of the Society for Maternal Fetal Medicine, January 30-February 4, 2006, Miami, FL. Reprints not available from the authors. 0002-9378/$ - see front matter Ó 2006 Mosby, Inc. All rights reserved. doi:10.1016/j.ajog.2006.06.080
unfractionated heparin (UH) and low molecular weight heparins (LMWH) are currently the preferred options. A recognized complication of chronic heparin therapy is osteoporosis. In 1965, Griffith et al1 identified 6 patients with spontaneous rib or vertebral fractures from a cohort of 10 nonpregnant women who had received large doses (15,000-30,000 units daily) of UH therapy for
1110 R6 months. Since then, there have been additional case reports of osteoporosis or fractures developing in patients who have received long-term heparin treatment.2-6 In 1993, Dahlman7 reported a symptomatic vertebral fracture rate of 2.2% in 184 pregnant women who received 13,000 to 40,000 units of UH daily (mean, 19,100 IU/d). Although 1 of 36 patients who received low-dose heparin therapy experienced a fracture (0.7%), there was a relationship between the incidence of fracture and the dose of heparin used.7 In another study that used single photon absorptiometry to measure bone mineral density (BMD) of the distal forearm at the start of treatment and within 1 week after delivery, Dahlman et al found a 4.9% reduction in trabecular bone in 39 pregnant women who had received a mean UH dosage of 17,300 U/d for a mean of 25 weeks and no such reduction in control subjects.8 Similarly, Barbour et al9 found a 5% decline in BMD at the proximal femur that was measured by dual energy x-ray absorptiometry (DEXA) in 14 pregnant women who received a mean UH dosage of 27,000 IU/d for 21 weeks; again, there was no such change in control subjects. In their study, 5 of 14 patients (36%) experienced R10% bone loss at the proximal femur.9 Several case series and animal studies had suggested that the bone loss with LMWH therapy is less than that seen with UH therapy.10-12 We performed a pilot study of 16 women who received 40 mg daily of enoxaparin for a mean of 25 weeks and found that no patient experienced bone loss of R10% that was measured at the femoral neck 6 weeks after delivery, and there was no significant change in BMD at the proximal femur overall.13 On the basis of these previous studies, we hypothesized that patients who receive thromboprophylaxis with enoxaparin sodium would experience less bone loss at the proximal femur than would patients who receive UH therapy.
Methods A multicenter prospective randomized controlled trial was performed at 9 centers from September 1998 until December 2005: Evanston Northwestern Healthcare, Duke University Medical School, Brown University Women and Infant’s Hospital, St. Peter’s University Hospital, Bridgeport Hospital, Prentice Women’s Hospital, University of Alabama, University of North Carolina, and Ohio State University. The Institutional Review Boards at all centers approved the study before patient recruitment at that site. Any patient who was a candidate for low-dose thromboprophylaxis for the duration of her pregnancy, who was at least 18 years old, and who could begin therapy at !24 weeks of gestation was eligible to participate. Patients were excluded only if they had a contraindication to
Casele et al anticoagulation. After informed consent was obtained, patients were assigned randomly to receive either UH or enoxaparin sodium therapy. Randomization was accomplished with a random number table, with each site stratified into blocks of 10. Baseline and serial platelet counts were obtained. All patients had a hypercoagulable evaluation. Patients in the UH arm self-administered 7500 units twice daily of subcutaneous heparin sodium from enrollment in the study until 28 weeks of gestation and then 10,000 units twice daily until delivery. Patients in the LMWH therapy arm self-administered 30 mg twice daily of subcutaneous enoxaparin sodium from enrollment until 28 weeks of gestation and then 40 mg twice daily until delivery. All patients received adjusted-dose Coumadin for 6 to 8 weeks after delivery. In addition to a prenatal vitamin, all patients were asked to take calcium supplementation of 500 mg daily from the time of enrollment until delivery. Bone density of the proximal femur was measured serially with DEXA. Measurements of the femoral neck and total hip bone density were made at baseline before the beginning of therapy or after 12 weeks of gestation. Bone density was measured again within 4 days following the patient’s delivery. Radiologists who were blinded to treatment arm interpreted all of the DEXA scans. Before the patients were enrolled, sample size calculations were performed. Our study design resembled that of Barbour et al,9 because we used serial DEXA measurements to determine the change in BMD at the femur as the primary outcome. We hypothesized that at least 30% of the UH group would experience a R10% decrease in BMD. To have 80% power to show a 50% reduction in the proportion of patients in the enoxaparin group (15%) who would experience the same degree of bone loss, we would need to recruit 240 patients. To show a 3-fold reduction to 10% in the enoxaparin group, we would need 124 patients; to show a 6-fold reduction to 5%, we would need 72 patients. We obtained approval for recruiting a total of 240 patients with a plan for a full interim analysis. All adverse events were reported to the Food and Drug Administration, Sanofi-Aventis, and the institutional review board at Evanston Northwestern Healthcare, where completed data collection sheets and copies of DEXA scans were sent and maintained by the research coordinator. The study was overseen by a data and safety monitoring group, which was comprised of 3 physicians at Evanston Northwestern Healthcare who did not have any financial or professional interest in the study. The data and safety monitoring group met twice per year to review reported adverse events and to assure that patient risks were being minimized. Premature termination of the study would be advised by the data and safety monitoring group if there was a statistically significant rate of unexplained pregnancy loss,
Casele et al recurrent thrombosis, bleeding, or other adverse events in 1 group compared with the other. Other criteria for premature termination of the study included directions from the manufacturer or an interim analysis that indicated that either fewer subjects were needed to show the hypothesized effect or that so many more subjects were needed that the study would be considered unfeasible. Statistical analysis was performed with SPSS software (version 13 for Windows; SPSS Inc, Chicago, IL). The Student t test, Fisher exact test, chi-square test, analysis of variance, and regression were used as appropriate. A probability value of !.05 was considered significant. Results are reported as means G SD.
1111 Table I
Characteristics of patients
Variable
LMWH therapy (n = 61)
UH (n = 59)
P value
29.6 G 4.8
30.7 G 4.6
.193 .438
72 13 12 3 26.6 27.7 G 4.9
73 21 5 2 35.1 26.3 G 5.6
.423 .198
8 (13.3%)
14 (24.6%)
.16
Maternal age (y)* Race (%) White Black Hispanic Other Primiparity (%) Treatment duration (wk)* Early treatment withdrawal (n)
* Data are given as mean G SD.
Results We requested consent from 147 women; 27 women declined enrollment, and 120 women gave consent and were enrolled. Twenty-two women did not complete the study: 11 women experienced a spontaneous abortion; 7 women were noncompliant; 2 women had an allergic reaction to the medication, and 2 women were switched to therapeutic anticoagulation at the discretion of their referring physicians. We enrolled 98 women in 7 years who completed the study. An interim analysis revealed that the incidence of bone loss of R10% in patients who received UH therapy was 2.5%, not 36% as reported by Barbour et al.9 A sample size that was calculated on the basis of 3% as the incidence of significant bone loss indicated that a total of 1628 women would be necessary to detect a 50% decrease in the incidence for the LMWH group; this was not deemed feasible, and the study was stopped. The patients in the UH group received a mean dose of 17,380 units for a mean of 26.3 weeks; the patients in the LMWH group received a mean of 68.4 mg enoxaparin for a mean of 27.7 weeks (P = .198) for the treatment duration. The groups were comparable with regard to maternal age, race, and parity (Table I). When the changes in BMD at the femoral neck were compared, the LMWH group gained a mean of increase of 0.9%, bone mass, and the UFH group lost a mean of 0.9% (P = .071). When BMD changes at the total proximal femur were compared, the LMWH lost 1.5%, and the UFH group lost 1.1% (P = .616, Table II). Bone loss at the femoral neck of R5% occurred in 4 of the 49 patients (8.2%) in the LMWH group and in 9 of 40 of the patients (22.5%) in the UFH group (P = .073). Bone loss at the total femur of R5% occurred in 6 of the 43 patients (13.9%) in the LMWH group and in 4 of the 40 patients (10%) in the UH group (P = .74). Clinically significant bone loss at the femoral neck (defined as a decrease of R10%) occurred in 1 of the 49 patients (2.0%) in the LMWH group and in 1 of the 40 patients (2.4%)
of the UH group (P = 1.0). Bone loss of R10% in the total proximal femur occurred in 1 of the 43 patients (2.3%) in the LMWH group and in none of the 40 patients in the UH group (P = 1.0; Table II). There was no association between age, parity, or institution and the degree of bone loss within groups or between groups. There was also no correlation between age and mean bone loss at the femoral neck (P = .719) or at the total femur (P = .838). Mean bone loss by age groups is summarized in Table III. Gestational age at delivery, infant birthweight, rate of early treatment withdrawal, and rate of complications (spontaneous abortion, bleeding, and recurrent thrombosis) were not significantly different in the 2 groups (Table IV).
Comment Fragility fractures occur in approximately 1 million women in the United States every year, which is more than breast cancer, gynecologic cancer, myocardial infarction, and stroke combined.14 The direct care cost of osteoporotic fractures is O$17 billion per year.14 Pregnancy is a particular time of risk because the growing fetus derives all of its calcium from maternal sources and because fewer than 6% of women of childbearing age in the United States consume the recommended amount of calcium daily.15 Although many clinical factors (such as age, family history, body mass index, use of steroids, previous fracture, and propensity for falling) contribute to a patient’s risk for fracture, BMD determined by DEXA scanning has been determined to be an important predictor of a patient’s risk for fragility fracture.16 Antithrombotic therapy is indicated frequently in pregnancy. LMWH therapies have become popular, despite their higher cost because of the ease of use with prefilled syringes, more predictable dose response, greater bioavailability, and a possibly lower risk for complications. Bleeding complications and heparin-induced
1112 Table II
Casele et al Mean bone density changes in patients who underwent thromboprophylaxis
Variable
UH
LMWH therapy
P value
BMD change at neck (g/cm2)* % change at neck BMD change total (g/cm2)* % change total Loss O 10% at neck Loss O 10% total
ÿ0.010 G 0.054 (n = 40) 0.9% (n = 40) ÿ0.012 G 0.029 (n = 40) 1.1% 1/40 0/40
C0.009 G 0.040 (n = 49) 0.9% (n = 49) ÿ0.016 G 0.036 (n = 43) 1.5% 1/49 1/43 (2.3%)
.054 .071 .584 .616 1 1
In LMWH grp 49 pts had 2 measurements at neck, 43 pts had 2 total measurements. * Data are given as mean G SD.
Table III
Mean bone density changes by age
Table IV
Pregnancy outcome
2
Mean change (g/cm ) Age group (y)
N
BMD at femoral neck
BMD at total femur
19-24.9 25-29.9 30-34.9 35-39.9 40-44.9
11 30 34 12 3
ÿ1.536 0.523 0.659 ÿ0.142 ÿ2.567
ÿ0.2 ÿ1.614 ÿ1.523 0.683 ÿ1.267
Analysis of variance between groups: P = .579; P = .742.
thrombocytopenia do appear to occur less frequently with LMWH use than with UH use.17 However, the evidence for less osteoporosis with LMWH use is less convincing. The results of our study indicate that prophylactic doses of UH and enoxaparin, when given to pregnant women for prolonged periods in conjunction with calcium supplementation, do not appear to result in clinically significant changes in BMD at the proximal femur. In addition, there was no significant difference between BMD changes in women who received UH or LMWH therapy. Our results differ from many previous reports. We hypothesize that the differences that were observed are related primarily to our study design. First, we included only patients who were receiving prophylactic doses of UH and LMWH with a mean daily dose of UH of 17,380 units and of LMWH of 68.4 mg, whereas previous studies included patients who received both prophylactic and therapeutic dosing and therefore a higher mean daily dosing. In the study by Dahlman et al,8 not only did the authors report an overall 4.9% loss in BMD, but they also observed a correlation with the degree of bone loss and the dose of UH. It is conceivable that we did not observe significant bone loss because all of our patients received only prophylactic doses. However, because patients whose conditions required therapeutic anticoagulation were excluded from this study, we are able to hypothesize only that this is a possible explanation.
Variable Spontaneous abortion (n) Recurrent thrombosis (n) Bleeding at delivery (n) Estimated gestational age at delivery (wk)* Infant birthweight (g)*
LMWH therapy (n = 60) UH (n = 57)
P value
4 (6.7%)
7 (12.3%)
.354
2 (3.3%)
4 (7%)
.431
4 (6.7%)
1 (1.8%)
.365
37.5 G 4.0
37.0 G 5.0
.583
3197 G 536
3125 G 633 .552
* Data are given as mean G SD.
In addition, with the exception of the study by Barbour et al9 and our pilot study,13 no previous authors have reported that they recommended calcium supplementation to their patients who received UH therapy. In fact, the study by Dahlman et al8 specifically states that the patients did not take calcium. We recommended that, in addition to prenatal vitamins, all of the patients in both arms of the study should take 500 mg of calcium. We did not specify the brand or type of calcium supplement so that many patients may have also been consuming varying amounts of vitamin D because some supplements contain 200 to 400 IU of vitamin D. The recommended daily level of calcium consumption is 1000 mg; however, the average consumption by women of childbearing age is only 700 mg daily.18-20 Prenatal vitamins typically contain 125 to 250 mg of calcium.19 The importance of calcium supplementation in pregnant women in general cannot be overemphasized. If adequate calcium is not part of the maternal diet, bone may be degraded to provide calcium for maternal needs.19,20 So bone loss can occur with normal pregnancy if calcium intake is inadequate. In addition, calcium supplementation in pregnancy and lactation has been demonstrated to diminish maternal bone resorption.20
Casele et al The precise mechanism of heparin-induced osteoporosis in unclear and may involve parathyroid hormonelike effects, direct osteoclastic activity, and derangements in mineralization.9 It is also not known whether calcium supplementation mitigates the effects of UH or LMWH therapy on bone. In our study, all of the patients were instructed and encouraged to take calcium supplementation, but we did not assess calcium intake in the diet nor compliance with calcium supplementation. Although we hypothesize that it may have been beneficial to our patients, further study is necessary to test this hypothesis and to clarify this issue. Because bone loss can occur in normal pregnancy in women who do not consume enough calcium and because calcium supplementation does appear to diminish bone loss in these women, we believe strongly that all pregnant women, especially those who require antithrombotic therapy, should be encouraged to take calcium supplementation of at least 500 mg daily. In conclusion, for women who require only prophylactic doses of antithrombotic therapy and take calcium supplementation during pregnancy, neither UH or LMWH therapy appear to have clinically significant impact on BMD.
References 1. Griffith GC, Nichols G Jr, Asher JD, Flanagan B. Heparin osteoporosis. JAMA 1965;193:91-4. 2. Griffiths HT, Liu DTY. Severe heparin osteoporosis in pregnancy. Postgrad Med J 1984;60:424-5. 3. Murphy MS, John PR, Mayer AD, Buckels JA, Kelly DA. Heparin therapy and bone fractures. Lancet 1992;340:1098. 4. Sackler JP, Liu L. Heparin-induced osteoporosis. Br J Radiol 1973;46:548-50. 5. Squires JW, Pinch LW. Heparin induced spinal fractures. JAMA 1979;241:2417-8. 6. Wise PH, Hall AJ. Heparin-induced osteopenia in pregnancy. BMJ 1980;281:110-1.
1113 7. Dahlman T. Osteoporotic fractures and the recurrence of thromboembolism during pregnancy and puerperium in 184 women undergoing thromboprophylaxis with heparin. Am J Obstet Gynecol 1993;168:1265-70. 8. Dahlman TC, Sjoberg HE, Ringertz H. Bone mineral density during long-term prophylaxis with heparin in pregnancy. Am J Obstet Gynecol 1994;170:1315-20. 9. Barbour LA, Kick SD, Steiner JF, LoVerde ME, Heddleston LN, Lear JL, et al. A prospective study of heparin-induced osteoporosis in pregnancy using bone densitometry. Am J Obstet Gynecol 1994; 170:862-9. 10. Matzsch T, Bergqvist D, Hedner U, Nilsson B, Ostergaard P. Effects of low molecular weight heparin and unfractionated heparin on induction of osteoporosis in rats. Thromb Haemost 1990;63: 505-9. 11. Monreal M, Vinas L, Monreal L, Lavin S, Lafoz E, Angeles AM. Heparin-related osteoporosis in rats: a comparative study between unfractionated heparin and low molecular weight heparin. Haemostasis 1990;20:204-7. 12. Ginsberg JS, Kowalchuk G, Hirsh J, Brill-Edwards P, Burrow R, Coates G, et al. Heparin effect on bone density. Thromb Haemost 1990;64:286-9. 13. Casele HL, Laifer SA. Prospective evaluation of bone density in pregnant women receiving the low molecular weight heparin enoxaparin sodium. J Matern Fetal Med 2000;9:122-5. 14. Miller PD, Barlas S, Brenneman SK, Abott TA, Chen YT, BarrettConnor E, et al. An approach to identifying osteopenic women at increased short-term risk of fracture. Arch Intern Med 2004;164: 1113-20. 15. Weisman SM. The calcium connection to bone health across a woman’s lifespan. J Reprod Med 2005;50:879-84. 16. Bates DW, Black DM, Cummings SR. Clinical use of bone densitometry: clinical applications. JAMA 2002;288:1898-900. 17. Hirsh J, Levine MN. Low molecular weight heparin. Blood 1992; 79:1-17. 18. Gold DT. Elevated calcium requirements for women and unique approaches to improving calcium adherence. J Reprod Med 2005; 50:891-5. 19. March of Dimes. Calcium. Available at: www.marchofdimes. com/pnhec/159_9472.asp. Accessed January 23, 2006. 20. Janakiraman V, Ettinger A, Mercado-Garcia A, Hu H, HernandezAvila M, et al. Calcium supplements and bone resorption in pregnancy: a randomized crossover trial. Am J Prev Med 2003;24:260-4.