Stress shielding of the proximal humerus following total shoulder arthroplasty (TSA) can promote unfavorable bone remodeling, especially for osteoporotic patients. The objective of this finite element (FE) study was to determine if a hollow, rather than solid, titanium stem can mitigate this effect for healthy, osteopenic, and osteoporotic bone. Using a population-based model of the humerus, representative average healthy, osteopenic, and osteoporotic humerus FE models were created. For each model, changes in bone and implant stresses following TSA were evaluated for different loading scenarios and compared between solid versus hollow-stemmed implants. For cortical bone, using an implant decreased von Mises stress with respect to intact values up to 34.4%, with a more pronounced effect at more proximal slices. In the most proximal slice, based on changes in strain energy density, hollow-stemmed implants outperformed solid-stemmed ones through reducing cortical bone volume with resorption potential by 11.7% ± 2.1% (p = .01). For cortical bone in this slice, the percentage of bone with resorption potential for the osteoporotic bone was greater than the healthy bone by 8.0% ± 1.4% using the hollow-stemmed implant (p = .04). These results suggest a small improvement in bone-implant mechanics using hollow-stemmed humeral implants and indicate osteoporosis could exacerbate stress shielding to some extent. The hollow stems maintained adequate strength and using even thinner walls may further reduce stress shielding. After further developing these models, future studies could yield optimized implant designs tuned for varying bone qualities.