Objective: Conflicting evidence is reported about the effect of treatment on bone mineral density (BMD) in adults with celiac disease (CD). This study analyzed the effects on BMD induced by treatment with a calcium-rich, gluten-free diet in adults with nonsilent CD.
Methods: In 30 women and 11 men with newly diagnosed CD, BMD was measured at the right femur (femoral neck and right Ward's triangle) and the lumbar spine by dual-energy x-ray absorptiometry under untreated conditions (pretreatment) and after 1-yr treatment with a calcium-rich, gluten-free diet.
Results: On average, posttreatment BMD was greater than pretreatment BMD at the lumbar spine (mean +/- SE: 0.907 +/- 0.028 and 0.795 +/- 0.028 g/cm2, respectively; p < 0.001), the femoral neck (0.818 +/- 0.023 and 0.741 +/- 0.030 g/cm2, respectively; p = 0.002), and the Ward's triangle (0.703 +/- 0.025 and 0.654 +/- 0.025 g/cm2, respectively; p < 0.001). The greatest BMD change (percent of baseline) was observed at the lumbar spine (+14.1%), the smallest at the Ward's triangle (+7.5%). In the absence of appropriate controls, the BMD change expected in the patients under untreated conditions was estimated by regressing pretreatment BMD over duration of CD with control for gender and age at which CD became clinically evident. The regression coefficient of this analysis indicated that 1 yr of untreated CD was associated with a BMD decrease at the lumbar spine by 0.00570 g/cm2 (95% confidence interval -0.0103 to -0.0011 g/cm2). The 95% confidence interval of the treatment-induced change in BMD at the lumbar spine (+0.060 to +0.160 g/cm2) did not overlap the 95% confidence interval of the BMD change expected under untreated conditions. A large interindividual variability was observed in the BMD response to the treatment: in univariate and multivariate analyses, the treatment-induced change in BMD was significantly related to gender (greater in women than in men) and to pretreatment age and BMD.
Conclusions: The data show that BMD is increased by dietary treatment of CD in most but not all adult patients; pretreatment BMD, gender, and pretreatment age predict the bone response after a 1-yr treatment.