Dual-energy X-ray absorptiometry (DEXA) of the spine has shown that phenylketonuric children develop osteopenia [1, 4,5]. While the exact reasons for low bone mineral density remain unclear, many explanations have been proposed: the diet being too restrictive [5], poor dietary compliance [2] or a direct effect of the disease [4]. In order to determine the role of the control of phenylalanine (Phe) blood levels in the pathogenesis of osteopenia, we performed a retrospective study of Phe levels from birth up to the present date in 13 children (5 females, 7 males) with classical phenylketonuria, diagnosed in the neonatal period by the French neonatal screening programme. Amino acid supplementation was started upon diagnosis and regularly adapted, based on the serum Phe levels.We compared the Phe values with the spine bone mineral density (BMD). DEXA was performed at the age of 12 years (range 5-21 years) using a Hologic QDR 4500A instrument. Spine BMD were interpreted according to French references values [3] and expressed as Z-score for gender and age. Patients were divided according to their spine BMD results into two groups: patients with osteopenia (Z-score <)1; n=8) and controls (Z-score ‡)1; n=5). Spine Z-scores were -2.39±0.88 (range -3.60 to -1.40) and 0.28±0.83 (range -0.70 to 1.40) for osteopenic patients and controls, respectively. All the Phe levels recorded in the patients' medical notes since the 1st month following birth up to the date the DEXA was done were used for the statistical analysis. The non-parametric MannWhitney test was used to compare continuous data between the two groups.We found that the age when DEXA was performed and the estimators of Phe levels position were similar between osteopenic patients and controls (Table 1). Nevertheless, although the difference was not significant, both the 3rd and 10th percentiles of Phe levels are lower and the percentage of Phe levels <2 mg/dl was higher in osteopenic patients. For the analysis of Phe level variations, we used the mean of cumulative variations which was defined as the sum of the differences between two successive Phe levels divided by the number of the differences studying per patient: i.e. Mean of cumulative variations = (S Phe X -Phe X+1)/(n-1); where Phe X, Phe X+1 are successive Phe levels, n, the number of the successive Phe levels studying per patient. Despite the small size of the population, osteopenic patients had a higher mean of cumulative variations (3.1±0.4 mg/dl) than controls (2.5±0.3 mg/dL, P=0.006). Linear regression analysis revealed a significant negative correlation of BMD spine (Z-score) with mean of cumulative variations (r=)0.61, P=0.025). These results suggest that the variations of serum Phe levels may contribute to osteopenia in phenylketonuric children. References 1. Allen JR, Humphries IRJ, Waters DL, Roberts DCK, Lipson AH, Howman-Giles RG, Gaskin KJ (1994) Decreased bone mineral density in children with phenyketonuria. Am J Clin Nutr 59: 419-422 2. Al-Qadreh A, Schulpis KH, Athanasopoulou H, Mengreli C, Ska...