Patient Recruitment. One hundred and twenty-six postmenopausal women were recruited from new subjects referred to the Bone Density Unit, Department of Obstetrics and Gynaecology, St. Luke’s Hospital, G’Mangia, Malta, for an osteoporosis risk evaluation. All were healthy Caucasians between the ages of 40-75 years [mean ± standard deviation (SD): 55.6 ± 7.1 years]. Informed consent was obtained from all participants and the study was approved by the Research Ethics Committee of the Faculty of Medicine and Surgery, University of Malta, Msida, Malta. All participants answered a detailed questionnaire concerning medical conditions and the use of medications, family history of osteoporosis and dietary/ lifestyle habits, and only those with no medical conditions and not on treatments known to affect bone metabolism, were included. Blood for DNA analysis and biochemical markers of bone turnover was collected from each participant.

According to World Health Organization (WHO) criteria [31] for both lumbar and femoral BMD, 30 (24.4%) of the participants were osteoporotic at the lumbar spine (LS) (t-score <–2.5), 41 (33.3%) were osteopenic ( t-score <–1.0 to –2.5) and 52 (42.2%) were normal. At the femoral neck (FN), four (3.2%) were osteoporotic, 45 (36.3%) osteopenic and 75 (60.5%) were normal. The sample was divided into two groups, one of normal individuals (41.1%) having both lumbar and femoral t-score >–1.0 and another of individuals who were osteopenic and/or osteoporotic at one or both anatomical sites (58.9%). The MTHFR C677T polymorphism was also analyzed in a random sample of 254 cord blood DNA obtained from neonates to determine the frequency of this polymorphism in the Maltese population at birth.

Genetic Analysis. Genomic DNA was extracted and purified from peripheral blood leucocytes by the salting-out procedure [32]. Fragments within the COL1A1 and MTHFR genes were amplified by polymerase chain reaction (PCR), using appropriate primers as previously described [11,12,22]. All reactions were performed in a total volume of 25 mL where amplification was performed using a ‘hot-start’ PCR protocol for the COL1A1 Sp1 and MTHFR C677T polymorphisms. Amplification of the COL1A1 –1997 G→T fragment was done using a hot-start/touchdown protocol starting from an annealing temperature of 60°C decreasing by 1°C every two cycles until it reached 50°C, followed by 15 cycles at an annealing temperature of 53°C. The correct size of the amplified products was evaluated by electrophoresis of 5 mL PCR product on a 1.0% agarose gel. The PCR products (10 mL) were digested using the restriction enzymes MscI (COL1A1 Sp1), BsaI (COL1A1 –1997 G→T) and HinfI (MTHFR C677T) according to the manufacturer’s instructions (New England Biolabs, Beverly, MA, USA). Nomenclature used for the COL1A1 –1997 G→T and MTHFR C677T refers to the actual nucleotide changes in the respective polymorphisms. Since the same nucleotide changes occur in both COL1A1 –1997 G→T and Sp1 polymorphisms, the Sp1 variant was labeled SS→Ss→ss (for nucleotide changes GG→GT→TT). Re-analysis of 20% of samples was performed to check for accuracy of genotyping. No discrepancies were found from the original genotyping.

Biochemical Markers of Bone Turnover. Blood and urine samples were collected and analyzed for biochemical markers of bone turnover. Urinary deoxypyridinoline (DpD) crosslinks and serum procollagen type I were used as markers of bone resorption and formation, respectively (Metra Biosystems Inc., San Diego, CA, USA).

Bone Densitometry. Bone mineral density was measured at the LS (L2-L4) and FN using a Norland 486 Dual-Energy X-ray Absorptiometer (Norland Medical Systems Inc., New York, NY, USA). The BMD measurements were done at the Bone Density Unit, St. Luke’s Hospital, G’Mangia, Malta.

Statistical Analysis. The normality of the population was tested using the non parametric Kolmogorov-Smirnov test. The chi-square test was used to test for Hardy-Weinberg and for differences in genotype frequencies between normal individuals and those with low BMD. Comparisons of continuous measurements were performed using the one-way analysis of variance (ANOVA), independent sample t-test or alternatively, the Kruskal-Wallis and Mann-Whitney tests were used for those variables not showing a normal distribution. Adjustments for age, body mass index (BMI) and years since menopause (YSM) were performed using a generalized linear model (GLM) univariate analysis of variance. Linear regression was used to test for allele dose effect. All statistical tests were performed using the Statistical Package for Social Sciences (SPSS) for Windows, student version 9.0 (SPSS, Chicago, IL, USA). The EH software by Terwillinger and Ott [33] was used to construct haplotypes of polymorphisms within the COL1A1 gene, to test for association between alleles (linkage disequilibrium) and for association of haplotypes with low BMD by comparing normal women with those who had a low BMD.