Purpose: Screwed joints are widely used in implant dentistry, and their failure is a significant problem that may be related to loss of preload. Preload is the compressive force generated across a joint when a screw is tightened and is responsible for keeping a joint closed. For a given torque, preload is limited by the frictional resistance of the contacting screw threads, flange, and opposing joint surfaces. This study tested the hypothesis that following correct placement, prosthetic gold screws lose preload over time. Materials and Methods: The study used standard Nobel Biocare components. Strain gauges mounted on a standard abutment formed a transducer to measure preload. Five sets of new prosthetic gold screws, gold cylinders, and standard abutment screws were assembled in turn on the top of an implant body, using the transducer abutment. The gold screws were tightened with a Nobel Biocare Torque Controller set to 10 Ncm, when it had an output of 12.06 ± 0.8 Ncm with 95% repeatability. Preload was monitored for 15 hours; then the screws were removed and examined under a scanning electron microscope. Results: Preload ranged from 157.5 to 488.9 N (mean 319.6 N), with a mean reduction over 15 hours of 24.9%, with 40.2% of this occurring within 10 seconds of tightening. Discussion: Torsional relaxation of the screw shaft, embedment relaxation, and localized plastic deformation of the gold alloy and opposing titanium threads were the most likely explanation for this phenomenon. Conclusions: New prosthetic gold screws suffer significant loss of preload following placement.