Short-term DXA-derived and exploratory molecular changes after acupuncture in Mexican-Mestizo postmenopausal women with extreme BMD phenotypes: A pilot study.

OBJECTIVE: To explore short-term changes in dual-energy X-ray absorptiometry (DXA)-derived bone mineral density (BMD) after a standardized acupuncture program in Mexican postmenopausal women and to generate hypotheses regarding osteoporosis-related transcriptomic pathways and candidate peripheral-blood genes. METHODS: This exploratory pre-post pilot study included 19 women selected in two extreme-phenotype strata: low BMD (n = 13; T-score ≤ -2.5 at the lumbar spine or hip) and normal BMD (n = 6; T-score ≥ -1.0 at both sites). Women with intermediate T-scores were not enrolled. All participants received 24 acupuncture sessions over 3 months. Femoral neck and lumbar spine BMD and anthropometric variables were assessed before and after treatment. A published osteoporosis-versus-normal microarray dataset was reanalyzed, followed by pathway and network analyses and reverse transcription quantitative polymerase chain reaction (RT-qPCR) assessment of three prespecified candidate genes in peripheral blood mononuclear cells. RESULTS: The low-BMD stratum was older and had higher glucose and lipid concentrations at baseline. In the low-BMD stratum, measured femoral neck BMD was higher after treatment (nominal p < 0.001), as were lumbar spine BMD (nominal p = 0.037) and hip circumference (nominal p = 0.046). After Bonferroni correction across these three focal comparisons, only the femoral neck BMD difference remained below the corrected threshold of p < 0.0167. The exploratory microarray screen retained 478 probe-set records using an absolute log₂ fold-change threshold > 0.5 and nominal p < 0.05; none met an FDR threshold < 0.05. Pathway enrichment was therefore considered hypothesis-generating, and RT-qPCR differences involving FNIP2, SAMD9L, and FGD2 were observed. After Bonferroni correction for nine tests, eight comparisons remained significant, whereas the FNIP2 pre-post comparison was significant only at the nominal level (p = 0.032). CONCLUSIONS: The program was feasible and was associated with short-term changes in measured DXA values and candidate molecular signals. However, the small nonrandomized sample, extreme-phenotype selection, absence of a sham or untreated comparator, baseline imbalance, and lack of a facility- and site-specific least significant change preclude causal or definitive efficacy conclusions. Larger randomized, sham-controlled trials with validated DXA precision estimates are required.