Table 5 Threshold effect analysis of plasma total homocysteine on lumbar bone mineral density using a two-piecewise linear regression model
Lumbar bone mineral density | Adjusted ß (95% CI) |
|---|---|
Plasma total homocysteine | |
Males aged 30-49 years | |
Fitting by standard linear model | 0.0018 −0.0012, 0.0048) |
Fitting by two-piecewise linear model | |
Inflection point | 9.0 |
Plasma total homocysteine <9.0 (umol/L) | 0.0042 (−0.0015, 0.0099) |
Plasma total homocysteine >9.0 (umol/L) | -0.0005 (−.0059, 0.0050) |
Log-likelihood ratio | 0.323 |
Males aged 50-69 years | |
Fitting by standard linear model | 0.0027 (−0.0009, 0.0063) |
Fitting by two-piecewise linear model | |
Inflection point | 10.0 |
Plasma total homocysteine <10.0 (umol/L) | 0.0098 (0.0032, 0.0163) |
Plasma total homocysteine >10.0 (umol/L) | −0.0053 −0.0125, 0.0019) |
Log-likelihood ratio | 0.011 |