The overall implant survival rate of 95% is slightly lower than the reported mean survival rates of the concept of tilted implants and immediate loading in edentulous jaws [11] but still close to them and maybe more comparable to investigations in which implants were also immediately loaded in the edentulous maxilla which were partly placed in fresh extraction sites [15, 16]. Nevertheless, what is remarkable is the two lost tilted implants (n = 2 of 20). In some reviews, there seems hardly to be a difference in the survival rate between axial and tilted implants [5, 11]. The potentially higher implant loss rate in this study might be due to the limited number of tilted implants.

In 30% of the patients (n = 3 of 10), one implant failed. There might be some reasons which could be responsible:

In the present study, one implant failed with a low primary stability. That confirms the assumption that a high primary stability is an important precondition for immediate loading [10]. However, the two other lost implants had high stability parameters. As shown by the low AUC values, the ISQ and PT values were unspecific parameters and unsuitable as a predictor for the risk of non-osseointegration in this collective, and this is in line with other studies [17, 18].

Another failure occurred in a situation where the provisional prostheses broke twice so that this implant might have been overloaded. Two of the failed three implants were completely or partially inserted in fresh extraction sockets, and studies have shown that this is an additional risk for implant failure in immediate loading in edentulous maxillae [16, 19].

That we have not found a significant difference in bone loss between straight and tilted implants is in line with the literature [5, 20]. In both reviews, the differences in bone loss after 12 months are in a range below a tenths of a millimeter and most probably not clinically relevant. It should be taken into account that the level of evidence of most studies is rather low due to the lack of randomized studies and the non-systematic use of a standardized technique to obtain a reproducible bone loss measurement [5, 11, 20]. This is a limit of the present study as well with a single cohort and measurements on digital panoramic radiographs and with irregular time intervals of the second measurement. This could explain that in some cases, even a bone growth was measured (up to 0.4 mm). Another limit of this study is the rather small patient group.