Three types of grade-4 titanium cylindrical nonself-tapping implants (codes 12S, 06D, and 06S) were specially designed and manufactured (Suwa Co., Ltd., Fujiyoshida, Yamanashi, Japan) (Fig. 1 and Table 1). The code 12S single-threaded implant served as a reference. Codes 06D and 06S were designed to double the thread length compared to 12S. Code 06D was a double-threated implant with the lead and lead angle equal to those of 12S, although the pitch was reduced twofold because of the second thread. Code 06S was a single-threaded implant, with the pitch, lead, and lead angle reduced twofold compared to those of 12S.

Artificial bone model

We employed a rigid polyurethane foam (18 cm × 4 cm × 13 cm, Solid Rigid Polyurethane Foam 20 pcf; Sawbones, Vashon, WA, USA) that mimics maxilla molar bone density (0.32 g/cc, similar to that of type 4) and physical properties (compressive strength, 8.4 MPa; tensile strength, 5.6 MPa; shear strength, 4.3 MPa; and coefficient of elasticity, 284 GPa). The insertion socket was 3.5 mm in diameter and 10.0 mm deep and was prepared using a drill (ASD-360, ASHINA, Hiroshima, Japan).

Measurement of torque

The implants were inserted with a 500g load at 15 rpm. Torque was measured (1 sample/ms) using a PC Torque Analyzer (TRQ-5DRU, Vectrix, Tokyo, Japan). The maximum IT value was acquired when implantation was complete, immediately after which the implant was removed using the same load and rotation speed, and torque was similarly measured. The removal torque (RT) value was acquired at the beginning of the test. Each implant was tested 10 times to achieve sufficient statistical power.

Assessment of implant stability

Implant stability quotient values (ISQ) were measured using a wireless resonance frequency analyzer (Osstell Mentor, Osstell AB, Gothenburg, Sweden). Each implant was connected to an Osstell SmartPeg (Osstell AB) that transmitted four times at different directions, and 12 measurements were performed using each implant.

Microscopic examination of contact interfaces

The contact interfaces of the artificial bone (hereafter referred to as “bone”) and implant body were microscopically examined following a published procedure. Briefly, two blocks of artificial bone (2-cm wide, 1-cm deep, and 3-cm high) were assembled into a prismatic column and circumferentially fixated with a metal jig. The insertion socket (3.5-mm diameter, 10.0-mm deep) was introduced into the center of the junction using a drill (ASD-360, ASHINA). The implants were inserted with a 500g load at 15 rpm. After insertion terminated, the two bone blocks were simultaneously removed from the metal jags and prismatic column and then separated to observe the contact interfaces. A digital microanalyzer (VHX-1000, Keyence, Osaka, Japan) was used to view the contact interface. Images were morphometrically analyzed using image processing and analysis software (PopImaging, Digital being kids Ltd., Yokohama, Kanagawa, Japan).

Statistical analysis

Numerical data are presented as the mean ± standard deviation (SD). Data were analyzed using a two-tailed Student t test to compare two groups and the Tukey-Kramer method for multiple comparisons (JMP14, SAS Institute Japan, Tokyo, Japan). P < 0.05 indicates a significant difference.