In order to achieve esthetically more satisfying results, it has been proposed to place ITI implants with their border between the rough and smooth surfaces below the level of the alveolar crest, thereby obtaining a submucosally located implant shoulder following healing. The aim of the present experimental study was to clinically and radiographically evaluate the tissue response to the placement of one-stage transmucosal implants with the border between the rough and the smooth surfaces sunk by 1 mm into a subcrestal location. 11 patients underwent comprehensive dental care including the placement of 2 implants of the ITI Dental Implant System in the same quadrant (test and control). Randomly assigned control implants were placed according to the manufacturer's instructions, i.e. the border between the rough titanium plasma-sprayed and the smooth polished surfaces precisely at the alveolar crest. At the test implant the apical border of the polished surface was placed approximately 1 mm below the alveolar crest. Probing bone levels were assessed at implant placement (baseline), 4 and 12 months later. Modified plaque and modified gingival indices were recorded at 1, 2, 3, 4 and 12 months. Clinical probing depth and "attachment" levels were measured at 4 and 12 months. All parameters were assessed at 6 sites around each implant. The mean for each implant was calculated and used for analysis. The Wilcoxon matched pairs signed rank test and the Student t-test were applied to detect differences over time and between the test and control implants. At baseline, a mean difference in probing bone level of -0.86 mm (SD 0.43 mm, p < 0.05) was found between test and control implants with the test implants being placed more deeply. Both test and control implants lost a significant amount of clinical bone height during the first 4 months (test 1.16 mm, p < 0.05; control 0.58 mm, p < 0.05). However, only the test implants significantly lost clinical bone height from 4-12 months (test 1.04 mm, p < 0.05; control 0.45 mm, p = 0.08). Overall, the test implants lost 2.26 mm and the control implants 1.02 mm of bone height during the first year of service. On the average, the test implants demonstrated a bone level of 0.38 mm lower than the controls at 12 months. Except for the modified gingival index at 4 months (mean difference 0.21, SD 0.19, p < 0.05), no clinical parameters yielded significant differences between test and control implants at any time. It is concluded that in addition to the crestal bone resorption occurring at implants placed under standard conditions, the bone adjacent to the polished surface of more deeply placed ITI implants is also lost over time. From a biological point of view, the placement of the border between the rough and the smooth surfaces into a subcrestal location should not be recommended.