The commonly used thermoluminescent dosimeter TLD-100 (Harshaw Chemical Company, Solon, OH) responds not only to photons and electrons, but also to neutrons that are produced during high-energy therapies. As a result, TLD-100 measurements outside of the treatment field are suspect when high-energy radiation is used. Although alternatives such as TLD-700 do not respond to neutrons, specialty dosimeters of this kind are expensive and are not routinely used in most clinics. In the current study, we examined the accuracy of TLD-100 in measuring the out-of-field photon dose as a function of treatment energy. To determine the accuracy of TLD-100 as compared with TLD-700, TLD-100 was irradiated outside of the treatment field by medical accelerators operated at 6, 10, 15, and 18 MV. In an effort to eliminate the response of TLD-100 to neutrons, TLD capsules were encased in varying thicknesses of cadmium foil (0.25 - 0.75 mm) before being irradiated at 18 MV. The out-of-field TLD-100 was found to be accurate at 6 MV and 10 MV, but to be substantially over-responsive at 15 MV and 18 MV (by up to 1063% relative to TLD-700). By wrapping the TLD-100 in up to 0.75 mm of cadmium, it was possible to drastically reduce (down to 39% on average) the over-response of the TLD-100; however, total removal of the over-responsiveness was not possible. Although TLD-100 is well suited for measuring out-of-field dose at energies as high as 10 MV, at higher energies (15 MV or greater), this dosimeter over-responds substantially and should not be used. Although encasing the TLD in cadmium minimized over-response to a degree, the reduction was not sufficient to make TLD-100 viable for measuring out-of-field dose at high treatment energies.