Osteoporosis International

, Volume 17, Issue 7, pp 1008–1012

Bisphosphonates in pregnancy and lactation-associated osteoporosis

Authors

    • Department of MedicineUniversity of Auckland
  • A. B. Grey
    • Department of MedicineUniversity of Auckland
  • R. Singh
    • Department of MedicineUniversity of Auckland
  • I. R. Reid
    • Department of MedicineUniversity of Auckland
Original Article

DOI: 10.1007/s00198-006-0112-3

Cite this article as:
O’Sullivan, S.M., Grey, A.B., Singh, R. et al. Osteoporos Int (2006) 17: 1008. doi:10.1007/s00198-006-0112-3

Abstract

Introduction

Pregnancy and lactation-associated osteoporosis (PLO) is an uncommon condition characterized by the occurrence of fracture(s) during late pregnancy or the puerperium. The aetiology is uncertain, and its management and natural history poorly defined.

Methods

We report a series of 11 women with PLO seen at our institution over the past 20 years, with follow-up ranging from 1 to 19 years.

Results

Ten women presented with painful low-trauma vertebral fractures, at a median of 1 month postpartum. In nine cases the fractures were multiple (median: 3, range: 2–5). At least one recognised risk factor for osteoporosis (low body weight, smoking history, family history of osteoporosis/fracture, vitamin D insufficiency) was present in nine patients. Bone density was in the osteoporotic range at the spine (mean T score: −2.8), with less marked reduction at the proximal femur (mean T score: −1.9). Nine patients received bisphosphonate treatment, for a median duration of 24 months. In the five women who received a bisphosphonate within 1 year of presentation, spinal bone density increased by 23% over baseline values after 2 years of treatment (p=0.0014). Of the 5 women who had subsequent pregnancies, one, who had declined bisphosphonate therapy after the initial presentation, sustained a fracture in the postpartum period. Two patients (both of whom were followed for at least 10 years) sustained fractures outside of pregnancy.

Conclusions

PLO is therefore associated with significant morbidity, a high prevalence of recognized risk factors for osteoporosis and a risk of recurrence in subsequent pregnancies. Bisphosphonate therapy administered soon after presentation substantially increases spinal bone density in patients with PLO.

Keywords

BisphosphonatesLactationOsteoporosisPregnancy

Introduction

Pregnancy and lactation-associated osteoporosis (PLO) is an uncommon condition characterized by the occurrence of fragility fracture(s), most commonly vertebral, in late pregnancy or the early postpartum period. The aetiology of PLO remains unclear, although a pathogenic role for breast-derived parathyroid-related peptide (PTHrP) has been suggested by animal studies [1] and individual human cases [2, 3]. Despite its relative rarity, PLO can be a devastating condition that causes severe, often prolonged, back pain and height loss in affected women. While the majority of cases have been reported in primagravid women, a substantial number have occurred in women who have had a previously uneventful pregnancy(s) [4]. To date, approximately 100 cases have been reported [2, 531], mostly as individual case reports, with the largest series comprising 29 cases drawn from a national database and contacted by post [17].

Currently, data on the effects of antiresorptive therapy and long-term outcome in PLO are limited. In most cases, women receive calcium supplements and are advised to wean [4]. Overall, the limited available data suggest that bone mineral density (BMD) improves with weaning and calcium supplements, although it does not usually return to normal [14, 18, 24, 26, 32]. Bisphosphonate use has been reported in a few individual cases [2, 13, 17, 18, 24, 26, 33].

In this series we present data on 11 patients with PLO, managed in a single metabolic bone clinic, 9 of whom received bisphosphonate therapy, and in some of whom long-term follow-up data were available.

Experimental subjects

We performed a retrospective chart review of cases of PLO seen at our osteoporosis clinic between 1986 and 2004. PLO was defined as the occurrence of low-trauma fractures occurring during late pregnancy or puerperium (up to 18 months postpartum). Clinical notes were reviewed for data relating to recognised risk factors for osteoporosis [cigarette smoking, low body weight, oestrogen deficiency (using the surrogate of oligo- or amenorrhoea), personal history of low-trauma fracture and a family history of low trauma fracture or osteoporosis], treatment type and duration, and relevant pregnancy-related data (number of pregnancies, duration of pregnancy and subsequent pregnancies).

Materials and methods

Bone density and laboratory measurements

Bone mineral density was measured using a Lunar DPX densitometer (Lunar, Madison, WI, USA). The same densitometer was used for all bone mineral density measurements. Laboratory data were variably available, reflecting the individual practices of the attending physician.

Statistics

Descriptive statistics are presented as mean±SD. Baseline and follow-up values of bone density were analysed using Student’s t-test for paired samples.

The current report represents an audit of clinical practice in our unit, for which our Institutional Ethics Committee does not require that written informed consent be obtained. Patients consented to the treatment and investigations described as part of their normal medical care.

Results

Demographic data on the 11 women are shown in Table 1. The majority of women were Caucasian (10/11), the other was Fijian Indian. Most of the women were non-smokers (7/11); 2 were ex-smokers and 2 were current smokers. The mean body weight at the time of fracture was 57.4 kg (range: 47.6–72.2 kg), mean body mass index was 21.9 kg/m2 (19.2–27.9 kg/m2); 5 of 11 (45%) patients had a BMI below 20, and 7 (64%) had body weight below 55 kg. The mean body weight in comparably aged New Zealand women is 68 kg [34]. More than half of the women (7/11, 65%) had sustained prepartum fractures. In all but one case, the previous fracture was either traumatic or occurred during childhood. Four (36%) patients had relatives with a history of osteoporosis or insufficiency fractures. Baseline calcium intake was available in 9 of 11 patients. The mean daily calcium intake was 680 mg (range: 400–1500). No patients were taking vitamin D supplements prior to fracture.
Table 1

Demographic data

Patient

Age

Ethnicity

Smoking history

Affected first-degree relative

Prepartum osteoporosis

Prepartum fracture

Weight/BMI (kg/kg/m2)

Parity

Duration of lactation (months)

1

22

Caucasian

N

Y

N

N

47.6/21.5

G1P1

 

2

33

Caucasian

N

N

N

Y (arm, childhood)

53.6/19.4

G1P1

18

3

31

Caucasian

Y

N

N

N

51.0/19.2

G1P1

5

4

29

Caucasian

N

N

N

Y (wrist, traumatic)

65.0/23.3

G1P1

7

5a

28

Caucasian

Y

Y

N

Y (clavicle, childhood)

55.2/21.0

G1P1

3.5

6

33

Fijian Indian

N

N

N

N

50.0/19.5

G2P2

0.75

7

35

Caucasian

N

N

N

N

72.2/26.2

G1P1

6

8

33

Caucasian

N

Y

N

Y (L wrist, childhood; R humerus, childhood)

53.0/19.9

G2P1

5

9

25

Caucasian

Y

N

Y

Y (R wrist, traumatic)

69.8/27.9

G2P2

0

10

33

Caucasian

Y

Y

Y

Y (L wrist, childhood; R ankle, traumatic; thoracic spine and clavicle, traumatic)

62.0/23.6

G1P1

3.5

11

37

Caucasian

N

N

Y

Y (L2,4 1999, amenorrhoeic)

52.0/19.6

G1P1

6

Mean

30

     

57.4/21.9

 

5.5

aPreviously reported

Table 2 contains clinical data relating to PLO. The median time to fracture was 1 month postpartum (range: 0–16 months). Nine women (82%) sustained multiple vertebral fractures (median: 3, range: 2–5), one patient sustained a single vertebral fracture and one patient suffered multiple non-vertebral fractures. Most of the women (8/11, 73%) presented during or after their first pregnancy. Among the three patients who presented following their second pregnancy, no fractures had occurred during or following the first pregnancy. All but 1 of the women had breast-fed, although most (8/11) had weaned before or at 6 months—the median duration of lactation was 5.5 months (0–18 months).
Table 2

Fracture data, treatment and follow-up

Patient

Fracture time postpartum (months)

Fracture site

Treatment

Duration of follow-up (duration of treatment with bisphosphonate) (months)

Subsequent pregnancies (fracture)

1

1.0

T6,9,12, L2

Calcium, fluoride, pamidronate

60 (24)

Y (N)

2

14

Bilat. wrists, L hip

Calcium

12 (0)

N

3

1.25

T7–9, T11–12

Calcium, pamidronate, alendronate

192 (59)

Y (N)

4

0.5

L1,2, T11,12

Calcium, vitamin Da, alendronate

26 (26)

N

5

3.5

L1–4

Calcium, vitamin D, pamidronate, alendronate

228 (61)

N

6

0.75

L1

Calcium, vitamin D, pamidronate, zoledronate

42 (28)

N

7

0.25

T6, 9

Calcium, vitamin D, alendronate, zoledronate

12 (12)

N

8

1

T7,9,11, L1

Calcium

12 (0)

N

9

0

T7–9

Calcium, fluoride, alendronate

132 (24)

Y (N)

10

3.5

L2–4

Calcium, vitamin D, alendronate

42 (21)

Y (N)

11

16

L2,4

Calcium, alendronate

39 (27)

Y(Y)

Median

1.0

  

42 (24)

 

aVitamin D given at replacement doses

One patient was hypercalcaemic, as has been detailed elsewhere [2]. Bone turnover data were available in nine women, but data were obtained using three different assays [urine hydroxyproline, urine N-telopeptide, serum NH2-terminal propeptide of type I procollagen (P1NP)]. Three women had vitamin D insufficiency [serum 25(OH)D<50 nmol/L].

BMD in one patient (number 5) was initially measured using single photon absorptiometry, so her BMD data were not further analysed. Baseline spinal BMD was not available in a second patient (number 11). In the other nine patients, the average initial BMD T-score at the spine was −2.8 (range: −0.7 to −3.8); that at the femoral neck was −2.0 (range: −0.1 to −2.8). All but two patients received at least one bisphosphonate (one pamidronate only, three pamidronate then alendronate, one pamidronate then zoledronate, three alendronate only, one alendronate then zoledronate), the exceptions being two women who declined advice to wean (numbers 2 and 8). Of these two women, one was reviewed at 1 year, and it was decided not to treat her with a bisphosphonate; the other had repeat bone density measurement but no clinical review at 1 year. Five women commenced bisphosphonate therapy within 1 year of presentation, while bisphosphonate therapy was not commenced until a median of 72 months (range: 24–168) after presentation in the other four. The median duration of bisphosphonate treatment was 24 months. All the women received calcium supplementation, and five also received vitamin D supplementation.

In those patients treated with a bisphosphonate within 2 years of presentation, spinal BMD increased by 17% at 1 year (n=4, p=0.006), and by 23% at 2 years (n=5, p=0.001). In patients not treated with a bisphosphonate within 2 years of presentation, spinal BMD increased by 2% at 1 year (n=4, p=0.6), and by 11% at 2 years (n=2, p=0.4) (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs00198-006-0112-3/MediaObjects/198_2006_112_Fig1_HTML.gif
Fig. 1

BMD of lumbar spine in individual patients with PLO during early follow-up

Only seven patients had paired proximal femur BMD measurements performed. In patients treated with a bisphosphonate within 2 years of presentation, femoral neck BMD did not change at 2 years (mean increase 0.7%, n=3, p=0.6). Only one patient who was not treated with a bisphosphonate within 2 years of presentation had follow-up femoral neck BMD at 2 years.

Four women were followed for ≥6 years (range: 6–18 years). Although each woman received some form of treatment for the majority of this time, on average they received a bisphosphonate for only 3.5 years (range: 2–5). In this group, the mean increase in spinal BMD over the entire period of follow-up was 20% (n=3, p=0.10) (Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00198-006-0112-3/MediaObjects/198_2006_112_Fig2_HTML.gif
Fig. 2

BMD of lumbar spine in four patients with PLO who were followed for ≥6 years. Numbers refer to individual patients, as per Table 1. aBaseline BMD measurements by single photon absorptiometry were excluded

Of the five women who had another pregnancy following the index event, one sustained further vertebral fractures following the subsequent pregnancy. Two patients experienced fractures during long-term follow-up, outside of pregnancy. One sustained a rib fracture following coughing, the other a left knee and right shoulder fracture following a fall.

Discussion

In this paper, we describe our experience of PLO in a single centre over the last 20 years. Like others [14, 17, 18, 24, 26] we found that this disorder causes considerable morbidity in the form of pain and disability, the majority of patients presenting with multiple vertebral fractures in the early puerperium. Our patients exhibited preferential bone loss at the spine, and there was a substantial (23%) increase in vertebral BMD in response to treatment with weaning, bisphosphonate therapy, correction of vitamin D deficiency and calcium supplementation. In those who did not receive early bisphosphonate therapy, the increase in spinal BMD appeared to be significantly smaller (11% at 2 years).

Among the patients in our series, there was a high prevalence of established risk factors for osteoporosis and/or fragility fracture, in that at least one risk factor was present in 82% of the affected individuals. Thus, low body weight was present in 64%, a family history of fragility fracture/osteoporosis in 36%, vitamin D insufficiency in 27% and smoking in 36%. Our review therefore highlights the importance of “traditional” osteoporosis risk factors in the pathogenesis of PLO. Other series have not consistently identified risk factors in patients with PLO. However, in one series, each of three women presenting with non-traumatic vertebral fractures after delivery and lactation had body weight below 60 kg, and in two body mass index was below 20 kg/m2 [24]. Dunne reported that first-degree relatives of women with PLO have an increased rate of fracture [17]. In some cases, vitamin D deficiency has been documented, but this has not been a universal finding [2, 14, 19].

We found both spine and hip BMD in the women in our series to be significantly below the normal value, but the decrement was greater at the spine, where the mean value was in the osteoporotic range at presentation. Spinal BMD improved substantially in our patients during follow-up, particularly in those who received early bisphosphonate therapy. Although the number of patients was small, there seemed to be a more substantial improvement in spinal bone density in those patients who received bisphosphonate therapy than in those who did not. Phillips et al. reported that spinal BMD in eight women with PLO treated with weaning, calcium and vitamin D increased by 6% at 8–18 months and 9.5% at 2–4 years [26]. Another study reported an increase in the spinal T-score of 0.4SD and in the femoral neck T-score of 0.1 in women treated with weaning, calcium and vitamin D [14]. In a small series of women treated with calcium, vitamin D and alendronate, there was a 5–15% improvement in spinal BMD [24]. We observed an increase of 23% in spinal BMD in the women treated early with bisphosphonates, suggesting that the addition of potent antiresorptive therapy produces greater improvements in BMD than weaning and supplementation with vitamin D and calcium alone. That the increases in BMD we observed are greater at the spine than the proximal femur is attributable to the higher bone turnover at that site and is a well-recognized pattern of response to potent antiresorptive agents. The increase in BMD seen in these patients is considerably greater than that seen in patients with post-menopausal osteoporosis who are being treated with bisphosphonates [35, 36]. This may be due to a contribution from weaning and may also reflect the high pre-treatment bone turnover in these patients. Our study does not address the optimal duration of bisphosphonate therapy in PLO, an issue which remains unresolved in post-menopausal osteoporosis. However, treatment for up to 5 years, followed by a clinical and densitometric review, seems a reasonable approach [37].

An important issue in the management of PLO is whether the condition recurs in subsequent pregnancies. Of the five patients in our cohort who had a subsequent pregnancy, one experienced further fractures. It is perhaps noteworthy that this patient did not receive bisphosphonate therapy after the first pregnancy. Two other patients, both of whom were followed for at least 10 years, experienced fractures in the long term, outside of pregnancy. Previous reports have indicated variable rates of fracture recurrence, as high as 33% [14, 26]. The available data, therefore, suggest that women with PLO should be warned that subsequent pregnancy carries a modest risk of further fracture.

Previous investigators have raised the possibility that PLO occurs in patients with pre-existing osteopenia [1012]. Investigations tend to be negative for secondary causes of osteoporosis, although Smith et al. reported that 4 of 24 women with PLO had a secondary cause of osteoporosis [14] and in another series, 6 women were excluded because of secondary osteoporosis [17]. In our series, three patients were known to have low BMD prior to pregnancy; all had lower BMD following pregnancy, and all sustained new fractures. It seems likely that their pre-existing disease predisposed them to postpartum fracture; however, it is clear that pregnancy itself provided an additional stressor which contributed to the fractures. It seems that a variable combination of low pre-pregnancy BMD and exaggerated bone loss during pregnancy leads to fracture.

In summary, women with PLO suffer considerable morbidity from vertebral fractures which are often multiple. Acknowledged risk factors for low BMD and/or fracture are often present. Bisphosphonate therapy is well tolerated and effective in management of this condition and appears to confer greater increases in BMD than are observed in patients managed by weaning and supplementation with vitamin D and calcium. The risk of fracture recurrence either in subsequent pregnancies or in the absence of further pregnancy appears to be low, but not negligible.

Acknowledgements

This work was supported by grants from the Australian and New Zealand Bone and Mineral Society (Dr. O’Sullivan) and by the Health Research Council of New Zealand.

Copyright information

© International Osteoporosis Foundation and National Osteoporosis Foundation 2006