Methods

Thirty nine children born before the gestational age of 37 weeks were enlisted in a study to assess the effect of early diet on BMC.4 Except for minor illnesses, the children were free from any chronic diseases. BMC of the left forearm, specifically at a site one third of the ulnar length measured proximally from the distal ulnar styloid (one third site), was measured by single photon absorptiometry5 at 3 months of age in 16 of these infants (mean (SD) gestational age 29.6 (3.0) weeks and birth weight 1263 (415) g). At 9–11 years of age, BMC was measured in all children by dual energy x ray absorptiometry (Norland XR-26, Norland Corp, Fort Atkinson, Wisconsin, USA)6 at the lumbar spine (L2–L4) and distal left forearm in addition to the one third site. The mean (SD) gestational age of the whole group was 31.3 (3.2) weeks and the mean (SD) birth weight was 1483 (442) g.

Genomic DNA was extracted and purified from EDTA blood samples using a QIAamp Blood Kit (Qiagen GmbH, Hilden, Germany). For restriction fragment length polymorphism studies of vitamin D receptor (VDR) and oestrogen receptor (ERα) genes carrying the polymorphic restriction sites, genomic DNA was amplified by polymerase chain reaction (PCR) using gene specific primers.7 ,8 After amplification, the PCR products were digested with restriction endonucleases (ApaI, TaqI, andFokI for VDR;PvuII for ERα), and the resulting fragments were separated in 2% agarose gel. The alleles were defined by capital letters in the absence of a restriction site and small letters in the presence of a given restriction site. The 569 bp fragment of genomic DNA containing the polymorphic portion (a G to T substitution) of the Sp1 binding site in intron 1 of the collagen type I α1 (COLIA1) gene was amplified by PCR under similar reaction conditions to those described above. The upstream primer was 5'-GGAAAGTAAAGCCAGGGATG-3' and the downstream primer was 5'-AACTCCAACCTCAGCCCATT-3'. After amplification, the PCR product was digested with PflMI endonuclease producing 379 bp and 190 bp fragments. The PflMI recognises the sequence CCANNN/NTGG where the last G represents the polymorphic site of the Sp1 binding region.9 According to Grant and coworkers,10 the allele S was designated as absence of the G to T transition (the presence of a restriction site in this study).

The polymorphic dinucleotide repeat upstream of the transcription start site of the insulin-like growth factor (IGF)-I gene was also amplified using gene specific primers.11 One primer was fluorescently labelled (Cy5; Amersham Pharmacia Biotech, Little Chalfont, Bucks, UK). The variation in length of labelled PCR products was analysed using an ALFexpress DNA sequencer and ALFwin Fragment Analyser software (Amersham Pharmacia Biotech AB, Uppsala, Sweden).

Between group differences were tested by analysis of covariance. The genotype was used as a factor and the weight at the time of the BMC measurement was used as a covariate. For the analysis of the effect of the IGF-I genotype, all different base pair lengths were separately analysed. The children were classified into three groups: (a) homozygotes for a certain length of DNA; (b) heterozygotes for that length of DNA; (c) those who did not have that length of DNA. Analysis of variance was performed with these groups for all basepair lengths. When the effect of different genotypes on the change in BMC was being evaluated, the difference between the BMC at 3 months and 9–11 years of age was used as a factor and the change in body weight during the same time period was used as a covariate. Because of the multiple comparisons, the α level was set at 0.01. The prevalences of VDR genotypes were compared by cross tabulation and χ² test.

The study was approved by the ethics committee of Tampere University Hospital, and written informed consent was obtained from the parents.