Table 3.

Results of the Quality Assessment

Study	Incomplete Outcome Data Addressed
Keller et al., 1999	Yes
Accursi, 2000	Yes
Morris et al., 2000	No
van Steenberghe et al., 2002	No
Dowell et al., 2007	Yes
Doyle et al., 2007	No
Alsaadi et al., 2008a	Yes
Alsaadi et al., 2008b	No
Tawil et al., 2008	No
Anner et al., 2010	No
Bell et al., 2011	No
Levin et al., 2011	Yes
Grandi et al., 2013a	No
Le et al., 2013	No

For all studies, sequence generation was not randomized; allocation concealment was inadequate; there was no blinding; and the estimated potential risk of bias was high.

^aThe study was a randomized clinical trial for immediately vs. loaded implants but not for diabetic and non-diabetic patients.