Adaptations in maternal calcium (Ca) economy occur during pregnancy to meet fetal demand and by birth a typical new-born contains 25-30g Ca. At the individual-level if Ca is mobilised it is important to identify where this occurs to identify any potential impact on fracture risk. At present, pregnancy is not considered a risk factor for osteoporosis but data are few. Trabecular bone is more metabolically active than cortical bone in part due to its structure, orientation and much greater surface area. During lactation, studies have consistently reported significant transient bone mineral mobilisation from trabecular-rich axial sites such as the hip and spine, while newer peripheral quantitative computed tomography (pQCT) techniques have found trabecular microarchitectural change in the appendicular skeleton also. My primary objective was to explore whether a similar pregnancy-induced bone mineral mobilisation could be observed in two distinct cohorts of pregnant women using novel pQCT techniques from mid- to late-pregnancy. These techniques provide non-invasive measurements of bone density, mass, geometry, distribution and strength at the radius and tibia.

This thesis aims to determine:​ whether pregnancy-induced changes occur in a) the trabecular compartment, b) the cortical compartment, c) any potential predictors of a) and b);​ d) to explore the correlation and agreement of pQCT techniques in-vivo. Single-slice and highresolution (HR) pQCT data were used to characterise maternal appendicular skeletal change during pregnancy in two studies in contrasting populations:​ 1) a resource poor subsistence farming Sub-Saharan African community with an habitually low Ca intake and high parity;​ 2) an affluent Cambridge based population, where Ca intake is higher but parity much lower. In The Gambia, existing pQCT data obtained at 14 and 30 weeks of pregnancy from women (n=811), aged 18-45 and accustomed to low habitual Ca intake, were analysed. In Cambridge, UK, I designed a study where pregnantwomen (n=53) and non-pregnant non-lactating (NPNL, n=37) controls aged 30-45 years were scanned with pQCT and HRpQCT at 14 and 36 weeks. These data were modelled to explore the maternal response to pregnancy in these distinct populations.

Contrary to my primary hypothesis there was no evidence of trabecular bone mobilisation in Gambian women, however in contrast in the UK study a -0.5 SD decrease in trabecular vBMD was observed. In The Gambia total vBMD did not change at the radius or tibia, while in Cambridge both techniques detected decreases in total vBMD at the distal tibia. Changes in cortical bone were documented during pregnancy in both populations. In Gambian women small but significant increases in cortical vBMD and BMC were observed at the radius and tibia. In Cambridge at the tibia HRpQCT showed a decrease in cortical vBMD and cortical thickness with an increase in cortical porosity. There were fewer pregnancy-induced changes at the radius but at the cortical- rich diaphysis cortical thickness decreased and endosteal circumference increased. No consistent predictors of these changes were found in either population.

These data from two contrasting populations show a “one size fits all” approach cannot be applied to the maternal skeletal response to pregnancy. In Gambian women with a habitually low Ca intake we have observed the conservation of bone mineral in the appendicular skeleton into the third trimester. In contrast in the Cambridge women evidence of trabecular and cortical mobilisation was observed at the distal tibia. At the radius changes were confined to the cortical-rich proximal radius and supported endosteal resorption during pregnancy. These data suggest that the maternal skeleton is a significant source of fetal Ca in a population accustomed to higher dietary Ca intakes. Further work is needed to determine what this mobilisation means at the individual level and to determine if the mobilised bone mineral is restored postpartum or whether there are lasting consequences for the woman’s bone health.