Female reproductive history and the skeleton—a review

Authors


Dr M. Karlsson, Department of Orthopaedics, Malmo University Hospital, SE-205 02 Malmo, Sweden.

Introduction

Bone mineral density (BMD) is generally regarded as the best predictor of bone strength, and a decrease of one standard deviation (SD) in BMD (around 10% of initial density) is known to double the fracture risk.1 Osteopenia is defined when the BMD reaches a value between 1 and 2.5 SD lower than young individuals (age 30) of the same gender, and osteoporosis is defined when the BMD is more than 2.5 SD lower than the reference value.2 Osteoporosis is known to carry an increased risk of fractures, proportional to the degree of bone loss. The purpose of this review is to evaluate the potential of pregnancy and lactation to influence BMD. In addition, we examine evidence for the theory that multiple pregnancies and an extended period of lactation are risk factors for osteoporosis. The search for relevant publications was determined with a Medline search from 1966 onwards, using the following search terms: BMD, bone mass, BMD, pregnancy, lactation, multiparity, multiple pregnancy and fractures. Only articles published in English were included. For each identified article, the search was extended using the Internet link ‘related manuscript’ repeatedly until no new relevant articles were found.

Pregnancy and BMD

A variety of prospective studies (both controlled or uncontrolled) suggest that pregnancy is associated with a loss in BMD (Table 1). Percentage bone loss reported varied from 2.1%3 to 9.4%4 in the spine, 0.9%5 to 3.9%6 in the hip and between 2.2%7 and 3.9%3 in the radius. However, some of these papers included fewer than 10 women, and the two most cited studies that oppose the view of a higher BMD loss in pregnant women included 32 women5 and 10 women,8 respectively. Furthermore, the follow up interval differed in pregnant and non-pregnant women at 475 and 669 days, respectively, in one of these studies.5 Only two studies performed the follow up measurements within two weeks of delivery.5,6 In the other studies, the baseline measurements were conducted up to 12 months before conception and the follow up measurements up to 12 months after delivery, thus including the influence of lactation in the differences identified. Only three studies reported in Table 1 were controlled, and none of the prospective studies adjusted the BMD values for changes in soft tissue composition. Overall, the published data, and a recently published large review,9 suggest that pregnancy is associated with a maternal loss of BMD around 5%(Table 1). Hypothetically, this loss would increase the fracture risk by 50%, if BMD remained at this level after delivery.

Table 1.  Reported change in BMD during pregnancy.
Region of body measuredReferenceNumber of womenRelative percentage of BMD changes
  1. All changes significant at 5% level compared with baseline except those indicated by an asterisk (*).

Spine76−2.4
414−9.4
235−3.0
2416−4.6*
2510−3.5
673−7.6
338−2.1
Hip/pelvis532−0.9*
2416−3.2
2510−3.6
673−3.9
Distal radius76−2.2
338−3.9

Pregnancy-related osteoporosis and transient osteoporosis of the hip

Most women lose a small percentage of BMD during pregnancy, but pregnancy-related osteoporosis is a rare complication with unknown incidence.10 Pregnancy-related transient osteoporosis of the hip is a similar condition localised only in the hip.11,12 Published data include mostly case reports, but two published cases series exist, including 2410 and 35 patients,13 respectively. The condition usually affects slightly built, primigravid, lactating women and does not usually recur. It is often diagnosed during the third trimester, but little is known about the timing of bone loss or recovery.10,13–16 Transient osteoporosis is usually regarded as benign because virtually all patients return to normal BMD without treatment 6–12 months after weaning without long term deficits, except where complicated by a vertebral, sacrum or hip fracture.10,13–17

Can supplementation or any other intervention influence BMD during pregnancy?

Nutrition is one of the more important regulatory factors in both the development and the maintenance of BMD. Approximately, 90% of BMD consists of calcium and phosphorus, but other dietary components, such as protein, magnesium, zinc, copper, iron, fluoride, vitamins D, A, C and K are also included and required for normal bone metabolism. The calcium content is vital for skeletal development not only during growth,18 but also for maintenance of bone mass in both premenopausal1 and postmenopausal women.19 Dietary calcium intake is limited by an absorption threshold above which increased supplementation will not improve the BMD.20 Pregnancy confers an extra demand for calcium as the fetus requires 50 mg/day at 20 weeks of gestation and 330 mg/day at 35 weeks of gestation.21 The concentration of 1,25-dihydroxyvitamin D is increased while the concentration of parathyroid hormone and calcitonin stays almost unchanged.21,22 As a result, calcium absorption increases in both the gut and the kidney. So, can dietary calcium supplementation influence BMD during pregnancy? Most data suggest that calcium supplements in pregnant women with normal or high calcium intake have little or no effect on BMD,8 but there are weak data suggesting that pregnant women with low calcium intake may benefit from calcium supplementation.26 Smoking, caffeine and alcohol exert a negative influence on BMD but it is unclear whether reduced use of these substances could affect BMD,27 or whether increased physical activity could reduce the pregnancy-related BMD loss.28

Lactation and BMD

A large number of prospective controlled or non-controlled studies suggest that lactation is associated with a loss in BMD (Table 2). BMD loss during six months of lactation has been reported at between 0.4% and 7.5% in the lumbar spine, between 2% and 5% in the femoral neck and between 0.2% and 7% in the distal radius. These data are consistent irrespective of study design, suggesting a 3–6% loss in BMD, most pronounced in the axial skeleton, with three to six months of lactation (Table 2). There appears to be a dose–response relationship, so that a longer period of lactation is associated with larger BMD loss and shorter periods of lactation with lower BMD loss.3,4,29 Only one study provided BMD values adjusted for changes in soft tissue composition during lactation and reported data supporting the other studies (BMD loss of 4.1% in the lumbar spine and 2.0% in the femoral neck).6

Table 2.  Reported change in BMD during lactation.
Region of body measuredReferenceNumber of womenRelative percentage of BMD changes
  1. All changes significant at 5% level or lower compared with baseline except those indicated by an asterisk (*).

Spine3298−5.0
3436−7.5
3125−0.4*
3059−5.2
2959−4.8
673−3.3
Hip/pelvis710−3.1
3298−4.8
3125−3.0
3059−5.3
2959−4.2
673−2.0
Distal radius4140−7.1
710−0.2*
3311−5.2
3436−5.0
4226−4.0
3059−0.5*
2959−2.0

A recently published review supports the suggestion that a reduction in BMD occurs with lactation, but that this reduction is reversible after weaning9(Fig. 1). This is in agreement with virtually all studies following BMD after weaning, reporting that BMD is restored 6–18 months postweaning.5,29–33 It has been proposed that this recovery cannot occur if pregnancies are closely spaced34 but published studies consistently show that women with many closely spaced pregnancies and extended lactation do not risk failure of bone recovery to prelactation levels.35,36 There are also data that suggest that the progesterone-only pill protects against this loss.37

Figure 1.

Relative changes in BMD during the first year following delivery in women with no breastfeeding and in women with breastfeeding below or above six months duration.

Can supplementation or any other intervention influence BMD during lactation?

Calcium intake is of importance during lactation. An increased maternal demand for calcium occurs during lactation, as 200 mg of calcium is transferred daily in the mother's milk to the infant during full breastfeeding. This leads to a total calcium transfer via the breast milk (during one lactation period of three to six months) greater than the calcium transferred across the placenta during pregnancy.21 Maternal absorption and excretion of calcium are altered postpartum, so that the calcium absorption and urinary calcium excretion return to prepregnancy levels shortly after delivery.4,8,38 Hypothetically, there could be a shortage of calcium for the skeleton, which would lead to decreased BMD and it is possible that calcium and vitamin D supplementation during lactation might affect BMD. Randomised controlled studies of calcium supplements given to lactating women with high and low calcium intake have shown no effects on bone turnover markers.8,21,22,38 The studies using the BMD as end point suggest that extra calcium has little or no positive effect on BMD loss during lactation, and any effect is usually described as transient with no long term benefits.38–40 Additionally, the few studies that report a small effect of calcium postpartum have usually found the same beneficial effect in non-lactating mothers,8,43,44 suggesting that calcium supplement does not influence the specific lactation-induced BMD loss. This view is further strengthened by the findings that breastfeeding women with low calcium intake did not benefit from calcium supplement44 and the fact that vitamin D requirements appear to be no greater in lactating than non-lactating women.45

The long term effects of pregnancies and lactation on BMD

Several observational and case–control studies of sibling sets suggest that women with a history of multiple pregnancies and long periods of lactation actually have higher BMD than women with few or no children (Table 3). The data on association of BMD with number of pregnancies are conflicting. Studies suggested that number of pregnancies was not correlated with BMD6 has been confirmed in other studies51,52 and this opinion has been stated in a recently published review including 23 different citations.9

Table 3.  Influence of parity upon BMD.
Region of body measuredReferenceNumber of womenPercentage of BMD changes in parous versus nulliparous women
  1. All changes significant at 1% level compared with baseline except those indicated by an asterisk (*).

Spine46825+4.2
471605+4.0
697+0.8*
481855+3.8
Hip/pelvis46825+5.4
471605+4.0
491091+5.5
697+3.2*
481855+5.2
Distal radius501652+3.8*

In contrast, several other studies have reported higher BMD with increasing parity.46–48,50,53–55 These studies included between 21754 and 312647 women from diverse racial groups and all showed higher up to the range of 3% in the total body, 8% in the femoral neck and 4% in the leg (Table 3). This association is further strengthened by one study that demonstrated a dose–response relationship, so that the lowest BMD was found in nulliparous women, intermediate values in primiparae and the highest value in women with two or more children.46

These findings appear confusing because both pregnancy and a lactation are followed by a BMD loss. It therefore appears that these findings are of minor clinical importance in the long term and other factors overshadow this loss so that mothers with many children and a long period of lactation have similar or higher BMD than their peers who have not given birth. Presently, we can only speculate as regards the causality, but changes in lifestyle associated with having a large family seem most likely to account for the outcome.

The long term effects of pregnancies and lactation on fracture risk

There are a number of observational and case–control studies that evaluate the fracture risk in relation to pregnancies and lactation.1,48,49,56–61 These studies suggest that women with a history of multiple pregnancies and a long period of lactation have no different, or lower, fracture risk than their peers with no children (Fig. 2). Furthermore, women who had breastfed for more than two years did not have a higher fracture risk than women who had never breastfed.56 These data are supported by a study of 70-year-old Swedish women52; the Study of Osteoporotic Fracture (SOF), which included 9704 American women over 65 years of age62; and the Mediterranean Osteoporosis Study (MEDOS), a cross sectional study including 2086 women from 14 centres in six countries in Southern Europe with a hip fracture and 3532 non-fractured controls.63

Figure 2.

Relative risk of sustaining a fracture in postmenopausal parous compared with nulliparous women, presented with first author numbers of women included, type of fracture evaluated and numbers of pregnancies included, with 95% confidence interval.

In fact, several studies demonstrate significant reductions in fracture risk with increasing parity. The fracture risk after two or more deliveries was only half that of nulliparous women in a study of 1855 postmenopausal women.48 Reduction in risk of between 6% and 35% has been reported by others.49,58–61,64,65 Thus, most studies suggest that there is no relationship or an inverse relationship between parity and fracture risk and that there is no association between lactation and fracture risk. When fracture is used as the end point, we achieve stronger data than when using a surrogate end point like BMD. So, having many children will probably lead to not only a high BMD but also reduced fracture risk. This view is further supported when data infer that the association between parity and reduced hip fracture risk may operate through a different mechanism than is detected by BMD.66,67 The BMD loss found to be associated with a pregnancy and lactation must be regarded as a finding without clinical relevance, as the deficit recovers after weaning.

In summary, most studies are in agreement with the suggestion that pregnancy and lactation lead to a BMD loss of up to 5%, but that this loss is reversed after weaning. Studies of elderly women who have had multiple pregnancies and a long total duration of lactation conclude that multiparous women have similar or higher BMD and similar or lower fracture risk than women with no history of pregnancy. The causality for this cannot be proven by any existing data, but discrepancies in lifestyle factors probably explain, at least partly, the discrepancies.

Ancillary