Bone mineral density during long-term prophylaxis with heparin in pregnancy

Bone mineral density during long-term prophylaxis with heparin in pregnancy Thomas C. Dahlman, MD, PhD, * Hans E. Sjdberg, MD, PhD, ' and Hans Ringertz, MD, PhD' Stockholm, Sweden OBJECTIVE: We studied the effect of long-term heparin treatment on bone mass during pregnancy. STUDY DESIGN: Thromboprophylaxis with heparin was given to 39 women during pregnancy for a mean of 28 weeks and for an average of 6 weeks post parturn. Bone mineral density measured with single-photon absorptiometry of the distal and ultradistal parts of the forearm was determined at the time of the start of heparin treatment (mean, twelfth week of gestation), immediately post parturn, and on average 7 weeks post parturn. The mean dosage of heparin was 17,300 IU/day. A control group of 34 normal pregnant women was studied for comparison. RESULTS: In women treated with heparin, there was almost a 5% reduction in trabecular bone during pregnancy (p < 0.01) and an insignificant recovery post partum. There were no significant changes in bone mass during pregnancy or in the puerperium in the control group. CONCLUSION: Long-term treatment with heparin during pregnancy is associated with bone loss, but indications of reversible changes are observed. (Am J OBSTETGYNECOL 1994; 170: 1315-20. )

Key words: Bone mineral, heparin, pregnancy

Clinical signs of osteoporosis, predominantly vertebral fractures, are reported in patients after long-term heparin treatment. ` ' The most frequent indication for this treatment is prophylaxis during pregnancy for women with a high risk of thromboembolic complications.' Most casereports on multiple spinal fractures in association with heparin are in connection with pregnancy. Oral anticoagulants, such as warfarin, are usually avoided during pregnancy because they pass through the placenta and have adverse affects on the fetus.' However, thromboprophylaxis with heparin does not affect the fetus, and heparin is therefore usually considered the drug of choice during pregnancy.' However, adverse effects on the mother with increased risk of hemorrhage after accidents, miscarriage, and other conditions should not be neglected. Whether normal Pregnancy can contribute to bone loss is unclear. A lossof trabecular bone, as determined From the Departmentof Obstetrics ' the Department and Gynecology, -h' and the Departmentof RaXolog-. of Endocnnolog. ý' Karolinska Hospital. Supportedby the SwedishMedical ResearchCouncil (ProjectNo. 5992) and by KarolinskaInstitute ResearchFunds. Receivedforpublicationjuly 8,1993; revisedNovember30,1993; December10,1993. accepted ThomasC. Dahlman,MD, Departmentof Obstetrics Reprintrequests: KarohnskoHospital, P.O. Box 140, S-171 76 and Gynecology, Stockholm,Sweden. Book,Inc. CopyrightC 1994 by Mosby--Year 0002-9378194$3.60 +0 611153592

by x-ray spectrophotometry of the wrist, has been reported, ' and even some case reports on idiopathic osteoporotic fractures of the hip during normal pregnancy have been found." On the other hand, there have been studies that did not show any loss of radial bone' or femoral bone' during pregnancy. The purpose of this study was to obtain information about bone mineral density by means of single-photon absorptiometry in pregnant patients treated with subcutaneous unfractionated heparin compared with untreated women during normal pregnancy and the puerperium. Patients and methods Patients and control subjects. Fifty women, considered to have an indication for prophylactic treatment with heparin during pregnancy, were consecutively recruited to the current study. A history of a thromboembolic complication was the predominant indication, and most of the women had undergone long-term treatment with heparin during preceding pregnancies. Ten heparin for 20 had been > given previously women weeks (mean 37.2 weeks). Forty-one healthy pregnant women, not deviating from the patient group for age, weight, height, parity, and smoking habits, served as controls (Table 1). They but be intended to matched controls, were originally because of dropouts in this group and mainly miscarriages in the patient group, the protocol had to be interrupted in this respect (Table 11). In all, 39 and 34

1315

May 1994 Am j Obstet Gynecol

1316 Dahlman, Sj6berg, and Ringertz

Table 1. Pregestationalcharacteristicsfor heparin and control groups completing study Clinical data

flepatin group

Controlgroup

Significance

No. Age (yr, mean and SD) Weight (kg, mean and SD) Height (cm, mean and SD) Smokers(%)

39 29.8 (5.7) 61.8 (7.5) 166.7 (4.9) 18

34 28.5 (5.5) 59.4 (6.1) 166.0 (5.0) 20

NS NS NS NS

36

40

NS

NuIliparous (%)

NS, No significant difference.

Table 11. Reasons for exclusion from study

Entering study* Miscarriage Dropout Warfarin (postpartum) Completing study

Heparingroup (No.)

Controlgroup (No.)

50 6t 17 40 39

41 0 34

*Mean twelfth week of gestation for heparin group and sixteenth week of gestation for control group. tFour of the women were treated with warfarin in early pregnancy and later were switched to hepann.

women in the heparin and control groups, respectively, completed the study. Bone mineral density of the forearm was measured on three separate occasions in both groups. The first measurement for the heparin group was performed when heparin was instituted (mean, twelfth week of gestation), and another measurement was taken at a mean of 4 weeks later for the control group. The second bone mineral density measurement was performed the first week after delivery in both groups. The third examination was carried out an average of 7 weeks post partum in the treatment group, just after heparin was discontinued, and in the control group somewhat later (mean, 12 weeks post partum). All of the patients treated with heparin and all of the control subjects had normal pregnancies, uneventful deliveries, and a normal puerperium. No calcium supplementation or any medication, except heparin, known to interfere with calcium homeostasis was given. In the puerperium no women used oral contraceptive pills, and all were breast-feeding. Heparin treatment. Unfractionated standard sodium heparin (Heparin, Kabi-Pharmacia AB, Stockholm) was injected subcutaneously twice daily by the women. The mean time for initiating the heparin treatment was the twelfth week of gestation, and the mean duration of treatment during pregnancy was 28 weeks (range 14 to 34 weeks). The dose of heparin was adjusted by monitoring the effect with an anti-factor Xa activity method 3 hours after injection, which was performed every 4

frelast thereafter the trimester more and weeks until quently. The method indirectly measures the concentration of heparin and its inhibitory effect on factor Xa and is far more sensitive than activated partial thromboplastin time. ' The mean dose of heparin during pregnancy was 17,300 IU/day (range 15,000 to 30,000 IU/day). Heparin was continued after delivery in a dose of 15,000 IU/day and was discontinued 6 weeks post partum. Bone mineral measurement. Bone mineral density was measured in each forearm with single-photon absorptiometry by means of an iodine 125 source (Nuclear Data 1100, Schaumburg, Ill. ) previously described. " Bone mineral density of the ultradistal end of the forearm, which contains 55% trabecular bone, " was calculated as the mean of four scans. Bone mineral density, one third from the distal end of the forearm, which contains 87% cortical bone, " was calculated as the mean of six scans. The mean bone mineral density of both forearms was calculated. Bone mineral density was expressed as the quotient between bone mineral content (grams per centimeter) and the radioulnar bone width in centiis 2% to in The (reproducibility) precision meters. vivo 4% and the accuracy is 3% to 4%. " The local radiation is 10 mrem to the mother for every examination of 10 scans, and the dose to the fetus is negligible. The total dose to each mother was thus 30 mrem compared with the maximum "permitted" radiation dose for industrial workers in the United States, which is 500 mrem. Statistical analyses. Differences within the groups were estimated by Student's t test for paired data and differences between the groups by Student's i test for be data. A 0.05 to < unpaired value ofp was considered statistically significant. All of the women gave their informed consent to the study, which was approved by the Ethics Committee at the Karolinska Hospital. Results As shown in Table III, bone density of the mineral two parts of the forearm examined was initially not significantly different in the heparin group from that of the control group. Furthermore, those 10 women who had previously undergone long-term treatment with

Volume 170, Number 5, Part I Am j Obstet Gynecol

Dahlman, Sj6berg, and Ringertz 1317

Table 111. Bone mineral density during pregnancy and puerperium control subjects

Distal one third of forearm (predominantly cortical bone) (g,,iilcm2)

Ultradistal forearm (predominantly tiabecular bone) (gmlcm')

I Penod Baseline Delivery Puerperium

in women treated with heparin and

Heparingroup (n =ýI

Controlgroup (n = 34)

Heparingroup (n = 39)

1.025 (0.11) 0.975 (0.10)* 0.998 (0.10) NS

1.018 (0.12) 0.995 (0,12) NS 1.011 (0.12) NS

1.398 (0.11) 1.393 (0.12) NS 1.392 (0.13) NS

L

Controlgroup (n = 34) 1.411 (0.12) 1.403 (0.13) NS 1.407 (0.13) NIS

Values are mean and SD. No significant differences were found between the groups at any period (unpaired t test). NS, No significant difference versus baseline. *p < 0.01 versus baseline (paired t test).

heparin (mean 37.2 weeks) were not significantly different from the control subjects regarding initial mean bone mineral density (1.01 [SD 0.12] vs 1.02 [SD 0.121 gm/cM2 in the ultradistal part and 1.36 [SD 0.131 vs 1.41 [SD 0.12 ] gM/CM2 in the one third distal part of the forearm). The second measurement taken at delivery showed a significant reduction of 4.9% from the baseline of bone mineral density in the ultradistal part of the forearm in the heparin group (p < 0.01) (Fig. 1) and a nonsignifi2.3% from baseline in the control decrease the of cant between difference The the groups was, howgroup. ever, not significant. In the puerperiurn there was a nonsignificant increase of bone mineral density in both groups. Bone mineral density of the predominantly cortical bone-containing part of the forearm (one third from the distal end) was not significantly different between the two groups during the study period and it did not change significantly in any group (Table III). No thromboembolic complications occurred. One of the women in the heparin group had spinal crush fractures.

Comment This study is the first prospective one to follow the effect of long-term heparin treatment on bone density, An almost 5% reduction of trabecular bone mass was observed in the ultradistal part of the forearms during pregnancy, which is a loss of bone comparable with that of post-menopausal women during 2 years. " Some degree of bone loss may occur during normal in indicated which was our control subjects pregnancy, and which was also observed by Lamke et al. ' using an x-ray spectrophotometric method of measuring bone mineral density. With the same method (single-photon absorptiometry) used in our study but with a smaller number of women (n = 13), no significant bone loss in the distal forearm was found by Christiansen et al. ' In a

recent study in which dual-photon absorptiometry of the femur was used, no significant mean bone loss was found during normal pregnancy.' Furthermore, it was shown that bone density increasedduring pregnancy in smaller women compared with that of control subjects. During lactation, gain of bone mineral was observed by Lamke et al.,' which coincides with our findings in both women treated with heparin and control subjects. The single-photon absorptiometry method is preferable in prospective studies of pregnant women in which each woman servesas her own control. The radiation dose is almost negligible. The distal and ultradistal parts of the forearm contain mainly cortical and trabecular bone, respectively. Trabecular bone has a higher turnover and is more susceptible to bone loss than cortical bone," which is in accordancewith the results in our study. Furthermore, most fractures reported after long-term heparin treatment are located in the vertebral bodies,which consist predominantly of trabecular bone. Techniques for measuring spinal bone mass such as dual-photon absorptiometry or computed tomography may thus be preferable. During spinal examination, the radiation to the fetus is not negligible. Nevertheless,bone mineral density of the forearm correlates well with that of the total skeleton."A single normal radial bone mineral density value does not, however, exclude spinal osteoporosis." In a previous radiologic study we found significant differences in the grade of osteopenia of the spine post partum between patients treated with heparin and untreated normal patients." Assessmentof osteoporosis from x-ray films is, however, subjective, and considerable loss of bone has to be present before osteopenia is evident. Whether the osteoporotic effect of heparin is reversible is uncertain. In the current study there was a bone indication (7 regain of of mass short-term weeks) post partum while the patients were still on a regimen bone density heparin. Follow-up mineral of of studies of these patients after discontinuation of heparin have

1318 Dahlman, Sj6berg, and Ringertz

Difference from baseline values t%) +2-t

May 1994 Am j Obstet Gynecol

Heparin group Control group

0

-2

-4

04

AJ

Baseline

BMD g/cm

Delivery

Puerperium

2

Heparin group Control group

1.04-

1.02-

l. OJ

B

ýx IIIII

12 16

PartU3

7

12

Weeks of gestatlon and postpartum

Fig. 1. A and B, Bone mineral density in ultradistal part of forearm during pregnancy and puerperium in women treated with heparin and in control subjects. Two asterisks,p<0.01 from baseline.n.s., No significant difference from baseline. No significant differenceswere found between the two groups at any period. Values are mean and SEM.

unfortunately not yet been performed. However, in another study by us (unpublished observations), 25 other women on a regimen of long-term heparin treatment during pregnancy were examined on the average

3 to 4 years later by means of single-photon absorptiometry of the distal forearm and compared with a difThere matched control group. were no significant ferences in bone mineral density between the groups.

Dahlman, Sj6berg, and Ringertz 1319

Volume 170, Number 5. Part I Am j Obstet. Gynecol

In a previous study on postpartum osteopenia" 85% of the women who had osteopenia of the spine were considered to have less osteopenia 2 years later. In a retrospective report by Ginsberg et al., " 61 women who had been on a regimen of long-term treatment with heparin (56 of them during pregnancy) were compared on average 2 years later with a control group by means dualthe wrist and of single-photon absorptiometry of photon absorptiometry of the spine. There were no in between differences the mean groups significant spinal or mean radial bone density. On radiographs of the spine only one of 46 patients had osteopenia. Furthermore, in the current study the mean trabecular bone mineral density of the heparin group was initially not different from that of the control group, even though most of the patients in the prophylaxis group had previously undergone long-term treatment with heparin. At entry into the current study 10 women with a history of a considerable duration of treatment with heparin (mean 37.2 weeks) in earlier pregnancies had a forearm density both bone the of sites at mineral mean that was not statistically different from the mean bone mineral density of the control sub ects. These findings indicate that if bone loss occurs during heparin treatbe induced. to processes seem reversible ment, Whether the dose of heparin or the duration of treatment is of importance for the development of osteoporosis is unclear. According to the first report on the connection between heparin and osteoporosis by Griffith et al., ' a dose relationship was found. Most case reports on osteoporotic fractures are among patients receiving high-dose heparin (>20,000 IU/day), but a case report in which only 10,000 IU/day induced osteoporotic fractures was also published. " The duration of heparin treatment might be of importance in inducing osteoporosis, as indicated by de Swiet et al. " No relation between osteopenia and the dose of heparin or between osteopenia and the duration of heparin treatment was, however, detected in our previous study" or by Ginsberg et al. " In the current study the mean dose of heparin was not high (17,300 IU/day), and yet there was a significant decrease in the mean trabecular bone mineral density. The mean duration of heparin treatment was 28 weeks. The pathogenesis of heparin-induced osteoporosis is also unclear. The effect of heparin and related substances on bone was reviewed by Avioli, ' who suggested direct follows: (1) a as effect of hepmechanisms several decreased bone cells osteoblastic or inwith on arin creased osteoclastic activity; (2) binding of calcium ions by heparin acting as a chelating agent, thus leading to a decrease in ionized calcium and inducing secondary hybone with enhanced resorption; (3) perparathyroidism defective ossification caused by disturbance of the matrix mucopolysaccharides of bone. Furthermore, a lack of vitamin D is suggested, and a decrease in the synthesis of bone collagen is indicated. " A further connection be-

tween heparin and the lossof bone is seenin mastocytosis,where an increasein heparin-containing mast cells is often accompaniedby osteopenia. Today, effective prevention of heparin-induced bone loss has not been established.Fractionated low-molecular-weight heparin has been shown to induce reduction in bone mineral content to the samedegree as conventional heparin, when given to rats in similar doses of heparin has " Xa. Hitherto, type this of not antifactor been used extensively in pregnant women and its effect on bone in man is unknown. When heparin prophylaxis is prescribed to pregnant women, the lowest possible dose to prevent thrombosis heparin be The of should concentration given. should preferably be monitored by antifactor Xa activity, which measures the anticoagulatory effect more sensitively than activated partial thromboplastin time. ' In the is thromboprophylaxis recomperiod postpartum mended for at least 6 weeks when coagulation and fibrinolysis are normalized." As indicated by our findings, an increase of bone mineral density in the puerperium during breast-feeding in both control subjects and patients treated with heparin and thus lactation does not appear to enhance loss of bone. Warfarin can be given to breast-feeding mothers" and may be preferred to heparin. As shown in this study and several case reports, long-term heparin treatment is associated with bone loss. The benefits of heparin compared to oral anticoagulants during pregnancy in high-risk patients must be weighed against the risk of osteoporotic fractures. Nevertheless,heparins is an effective tool in preventing thromboembolism and must not be shunned when obvious indications during pregnancy are given. We are grateful to Shirley Kal6n for qualified work with measurements of bone mineral and to Astrid HAggblad for preparing the manuscript.

1. 2. 3.

4. 5.

6.

7.

8.

9.

REFERENCES Griffith GC, Nichols G jr, Asher JD, Flanagan B. Heparin osteoporosis. JAMA 1965; 193: 91-4. Avioli LV. Heparin-induced osteopenia: An appraisal. Adv Exp Med Biol 1975; 52: 375-87. Bonnarj. Haemostasis and coagulation disorders in pregnancy. In: Bloom A], Thomas DP, eds. Haemostasis and thrombosis. London: Churchill Livingstone, 1987: 570-84. Ginsberg JS, Hirsh J. Anticoagulants during pregnancy. Annu Rev Med 1989; 40: 79-86. Lamke B, Brundin J, Moberg P. Changes in bone mineral Acta Obstet Gylactation. during and pregnancy content necol Scand 1977; 56: 217-9. Smith R. Stevenson JC, Winearls CG, Woods CG, Woods1: Lancet 1985; Osteoporosis BP. pregnancy. of worth 1178-80. Christiansen C, Rodbro P, Heinild D. Unchanged total body calcium in normal human pregnancy. Acta Obstet Gynecol Scand 1976; 5: 141-3. Sowers M, Crutchfield M, jannausch M, Updike S, Corton in bone mineral change G. A prospective evaluation of 1991; 77: 841-5. Gynecol Obstet pregnancy. M. Thrombosis Blomb5ck M, proT, Hellgren Dahlman heparin subcutaneous in of pregnancy with use phylaxis

1320 Dahlman, Sj6berg, and Ringertz

adjusted by monitoring heparin concentration in plasma. GYNECOL1989; 161: 420-5. Am J Oa= 10. Christiansen C, Rodbro P. Long-term reproducibility of bone mineral content measurements. Scand J Clin Invest 1977; 37: 321-3. 11. Nilas L, Norgaard H, Podenphant J, Gotfredsen A, Christiansen C. Bone composition in the distal forearm. Scand J Clin Lab Invest 1987; 47: 41-6. 12. Chesnut CH. Non-invasive methods of measuring bone mass. In: Avioli LV, ed. The osteoporotic syndrome. Detection, prevention and treatment. Orlando: Grune & Stratton, 1987: 31-43. 13. Christiansen C, Riis BJ, Five years with continuous combined estrogen progestogen therapy. BrJ Obstet Gynaecol 1990; 97: 1087-92. 14. Lam SY, Baker HWG, Seeman E, Pepperell RJ. Gynaecological disorders and risk factors in pre-menopausal women predisposing to osteoporosis. Br J Obstet Gynaecol 1988; 95: 963-72. 15. Cummings SR, Kelsey JL, Nevitt MC, O'Dowd KJ. Epidemiology of osteoporosis and osteoporotic fractures. Epidemiol Rev 1985; 7: 178-208. 16. %Vahner HW, Dunn WL, Riggs BL. Assessment of bone mineral. Part 2. J Nucl Med 1984; 25: 241-53.

May 1994 Am j Obstet Gynecol

17. Dahlman T, Lindvall N, Hellgren M. Osteopenia in pregnancy during long-term heparin treatment. Br j Obstet Gynaecol 1990;97:221-8. 18. Ginsberg JS, Kowalchuk G, Hirsh 1, et al. Heparin effect on bone density. Tbromb Haemost 1990;64:286-9. 19. Griffith HT, Liu DTY. Severe heparin osteoporosis in pregnancy. Postgrad Medj 1984;60:424-5. 20. de Swiet M, Darrington Ward P, Fidler J, et al. Prolonged heparin therapy in pregnancy causesbone demineralization. Br j Obstet Gynaecol 1983;90: 1129-34. 21. Hurley M, Kream B, Raisz L. Structural determinants of the capacity of heparin to inhibit collagen synthesis in 2 I-day fetal rat calvariaej Bone Miner Res 1990;11:112733. 22. MatzschT, Bergqvist D, Hedner U, Nilsson B, Ostergaard P. Effects of low molecular weight heparin and unfragmented heparin on induction of osteoporosis in rats. Thromb Haemost 1990;63:505-9. 23. Dahlman T, Hellgren M, Blomback M. Changesin blood coagulation and fibrinolysis in the normal puerperium. Gynecol Obstet Invest 1985;20:37-44. 24. Orme L'E, Lewis Pj, de Swiet M, et al. May mothers given warfarin breast-feedtheir infants? BMJ 1977;1:1564-5.