The morbidity caused by bone graft harvesting and the delay in the final treatment due to the time necessary for bone incorporation triggered the development of techniques without grafting as an option for the treatment of patients with edentulous jaws . Brånemark in 1998 developed a novel technique for placing implants in the zygomatic bone to treat severely atrophic maxilla without the need for grafting, which was later modified by Stella and Warner . The later minimized the presence of the implant into the maxillary sinus, improving the emergence of the implant, since it allowed a more vertical angle than the original technique. Many prospective and retrospective studies [1, 5, 6, 12, 14–17, 22–24] showed good results by using the original technique, while only few researches [11, 18] discuss Stella and Warner’s technique. So, this retrospective study aimed to evaluate Stella and Warner’s technique, contributing to a greater scientific validation.
Fourteen patients who underwent placement of zygomatic implants were evaluated over a period ranging from 15 to 53 months, where 100 % survival rate of conventional and zygomatic implants involved in the rehabilitation was observed. This represented a survival rate compatible with Brånemark’s studies [14, 19] that showed a survival rate for zygomatic implants of 94.2 to 100 % after 5 to 10 years and 12 years of follow-up, respectively. Different authors [1, 6, 11, 15–18, 21, 22, 24–26] reported a survival rate for these implants between 96 and 100 %. For conventional implants, some studies [11, 15, 16, 18, 21, 22] reported a survival rate ranging between 95 and 100 % . The findings of our study demonstrated that the technique of Stella and Warner is fairly predictable with survival and success rates compatible with the ones described in the literature, independently on the technique used.
The bone level after loading the implants was one of the criteria used in this study to assess the survival rate of conventional implants involved in the rehabilitation. As described by Hirsch et al. , Aparicio et al. , and Farzad et al. , this criteria was assessed by periapical radiographs obtained by the parallelism technique combined with panoramic radiographs for conventional implants and CT scans for zygomatic implants. The bone loss was defined as a vertical change of bone level measured from the most inferior line of implant exposure. All previous studies have demonstrated satisfactory sustained levels over a period of 60 months of the load application.
One aspect to be considered is that our research is a retrospective cohort study, making it difficult to find a standardization of radiographs that could accurately determine the annual bone loss as described by Farzad et al. , especially for conventional implants. Sometimes when the conventional implant is slightly tilted to the palate in cases of anterior maxillary atrophy, it is difficult to obtain adequate periapical radiographs by the parallelism technique. Therefore, the methodology suggested by Lang and Lindhe  was used with reference to the implant bone level that should not be less than two thirds of its total length in order to have a satisfactory osseointegration. Additionally, the radiographic criteria recommended by Buser were used to analyze the presence/absence of persistent radiolucency around the implant.
For greater accuracy to assess implant osseointegration, percussion and immobility tests are described in the literature [15, 16]. Performing these tests requires the removal of fixed prosthesis, which in our view, would be justified only in cases of necessity for prosthesis replacement or when mismatches or gaps was observed in the rehabilitation and/or implants, since the osseointegration loss can also be verified by signs indicating decreased bone volume around the implant, periimplant radiolucency, the presence of spontaneous pain on palpation, local redness, and the presence of purulent secretion [6, 26, 27]. According to Von Krammer , periimplantar mobility is generally associated with periimplant radiolucency and this monitoring method has the advantage of not requiring removal of the prosthesis during the evaluation.
Our study demonstrated the absence of pain as well as of pus or bleeding on probing and palpation for both zygomatic and conventional implants, with good bone level for conventional implants. No periimplant radiolucency was noted around the conventional implants and in the apical portion of the zygomatic implants. These findings are similar to the studies of Stiévenart and Malevez , Peñarrocha et al. , and Davo et al. .
Some authors [6, 16, 17] reported inflammation of the soft tissue around the implants, being mucosa redness one of the signs. Farzad et al.  reported some degree of inflammation found in 14 of 22 zygomatic implants installed by the original technique. The soft tissue found around the implant appeared to be susceptible to infection justified by the authors by the increase in the number of patients with problems in performing a correct oral hygiene. In our study, only one patient had redness in the palatal region of the left zygomatic, being also associated with poor hygiene. In our opinion, probably a factor that contributes to the health of periimplant zygomatic implants is a sulcus depth within an acceptable range, which is able to promote self-cleaning by the patient during brushing and/or gum massage.
The probing depth in the palatal mucosa of zygomatic implant can be considered normal up to 5 mm, consisting of parakeratinized epithelium which is not comparable to the normal depth of the sulcus around a conventional implant . In this study, both the zygomatic and conventional implants had a mean probing depth within normal limits, ranging from 2 to 3 mm, which is considered satisfactory. The technique of Stella and Warner allows a more vertical emergency profile favoring a less deep sulcus due to a more open angle obtained. Other studies have reported the presence of problems with oral tissue in the region of the zygomatic implants, including infection and swelling, usually associated with loss of implant apical osseointegration [6, 17, 28]. Hirsch et al.  have reported the presence of hyperplasia, mucositis, and infection in eight patients in a total of ten throughout the monitoring period.
Although certain criteria to evaluate osseointegration were considered, the study used the concept of survival of the implants instead of the success. Survival is a more general term, considering only the implant is still in the oral cavity, without analyzing the quality of its function and maintenance of its support. The success rate is applied to those fitting into established and applied criteria of installed implant . Due to various factors, from the type of test, as well as the difficulty to access all implants and characteristics inherent when working with atrophic maxilla without reconstruction, we found that survival, which is a concept widely used in the literature, would fit best for this research.
The possibility of zygomatic implants causing or favoring sinus disease due to the exposure of the implants within the maxillary sinus is an important issue. A number of studies describe the occurrence of maxillary sinusitis in patients with zygomatic implants [4, 6, 8, 10–12, 15, 17, 18, 25, 26]. This finding has been attributed to perforation of the sinus membrane [17, 26], a lack of contact between the implant and the surrounding bone crest , the migration of bacteria from the oral cavity to the maxillary sinus due to communication between these structures [7, 17], and preexisting sinus conditions from the clinical and radiographic standpoint [4, 5, 7, 8, 10, 27].
According to Stiévenart and Malevez , the incidence of sinusitis ranges from 14 to 30 %. In a previous retrospective study involving patients submitted to the Stella and Warner’s technique, Peñarrocha et al.  found two cases of sinusitis among the 42 implants; one case was treated with antibiotics and the other was submitted to the removal of the implant. Becktor et al.  report that patients with oral-sinus communication may develop suppuration with or without sinusitis. In such cases, treatment consists of the administration of antibiotics and/or the repositioning of the soft tissue and maintenance of a stable zygomatic implant, with no reports of the recurrence of sinusitis [6, 14, 25]. Brånemark  found fistula in five patients both before and after the connection of the abutment in 1 year of follow-up. Three patients exhibited specific symptoms of sinusitis, such as nighttime pain, unilateral pain in bad weather, and obstruction of the sinus. The existence of a small amount of residual bone in the alveolar crest associated with an implant placement technique with minor destruction of the sinus region can determine a more favorable prognosis for these complications.
The risk of the development of maxillary sinusitis associated with zygomatic implants installed using the original technique is reported to be low to moderate . Few data have been published regarding this risk in relation to the Stella and Warner’s technique. According to Peñarrocha et al. , the small slot in the zygomatic-maxillary region diminishes the likelihood of maxillary sinusitis, reporting a 4.7 % rate of occurrence of this complication in 12 months of follow-up. In the present study, the implants had two position patterns: those that partially invaded the maxillary sinus and those that were positioned alongside but completely outside the sinus due to the anatomy of the zygomatic-maxillary region encountered. However, the slot that characterizes this technique was made in both cases. The clinical and imaging findings demonstrated no cases of maxillary sinusitis in the follow-up.
According to the Brazilian Guidelines for Sinusitis , the clinical exam has sensitivity and specificity of 69 and 79 %, respectively, which makes the use of complementary diagnostic tools necessary. A number of authors report the use of computed tomography for the diagnosis of sinusitis. Nakai et al.  performed this exam 6 months following the placement of 15 zygomatic implants in nine patients and found an absence of signs and symptoms of sinusitis. Maló et al.  evaluated the association between zygomatic implants and maxillary sinusitis using sinusoscopy on 14 patients and found no cases of infection or inflammation of the mucosa surrounding the implants, demonstrating that titanium implants are compatible with the health and normal function of the maxillary sinus. However, the studies cited employed the original technique.
In a systematic review, Chrcanovic and Abreu  report that immobility of zygomatic implants is one of the main factors contributing to the homeostasis of the maxillary sinus. This immobility is accomplished by adequate anchorage of the implant in the zygomatic bone and, when possible, the maxillary bone as well as a firm connection with the overdentures. The rigorous selection of patients with no history of active sinus disease is another important factor and was confirmed in the present sample through preoperative computed tomography, following the normal routine of the hospital at which this study was carried out.
Computed tomography is currently the method of choice for the determination of sinusitis. A number of scoring systems have been proposed for this purpose, most of which are based on the presence and extent of inflammation in the interior of the paranasal sinuses. The Lund-McKay  scoring system is an objective method for the evaluation of opacification of the sinuses on tomograms that eliminates the occurrence of false positives or negatives. A clinical exam and computed tomography performed by an otolaryngologist allows a precise diagnosis of sinus disease, which can present in a similar manner without necessarily being maxillary sinusitis. Moreover, a number of studies have demonstrated that cone-beam computed tomography (as employed in the present study) is a good imaging tool for the evaluation of sinus disease.
The prostheses supported by zygomatic implants have a special design due to the location and a more palatal emergence profile of the implants in position when compared to conventional implants. This situation can hinder the tongue position and hygiene of the prosthesis and interfere with function [14, 31]. Some studies [6, 16, 31] conducted an assessment of the level of patient satisfaction on the prosthesis supported by zygomatic implants, demonstrating good levels of acceptance. Farzad et al.  evaluated the satisfaction of patients undergoing placement of zygomatic implants by Stella and Warner’s technique and compared with a group rehabilitated with full fixed prosthesis without zygomatic implants also using VAS.
No statistically significant differences have been found considering the different aspects analyzed, except with respect to aesthetics. In our study, there were significant differences in both overall satisfaction as the specific items assessed showed better results in total fixed prosthesis without zygomatic implants, although group I, represented by the PTF with zygomatic implants, has achieved good averages, except in the ease of entry for cleaning the prosthesis [15, 20, 31].
Farzad et al.  in their assessment of patient satisfaction after rehabilitation did not describe changes in speech. However, in our study three patients rehabilitated with zygomatic implants complained of difficulty in the ability to speak, especially when pronouncing words with the letter “s”. Nakai et al.  also reported the presence of patients complaining about speech, one patient complained for 3 months and the other one for 2 weeks, both after installation of the prosthesis. Brånemark et al.  and Nakai et al.  correlated problems in speaking with the design of the installed prostheses in patients with zygomatic implants which differs from those who are treated with conventional implants with or without the need for grafting.
Hirsch et al.  evaluated the satisfaction at the time of insertion of fixed prostheses and after 1 year of follow-up in 76 patients treated with 124 zygomatic implants. Complete satisfaction was observed with the cosmetic and functional results in 80 % of these patients, in both time frame analyzed. Farzad et al.  also used a VAS to assess patients’ response to treatment with zygomatic implants, describing difficult to chew and less satisfaction with respect to aesthetics, that can be related to the subjectivity of the analyses. In our study, both groups of patients presented good results with respect to aesthetics and function, but the conventional implant group showed the highest rate for both questions. For the group with zygomatic implants, two patients in the cosmetic item reported that the prosthesis did not show the expected results, which may have been influenced by the individual’s subjective opinion.
Analyzing masticatory function and stability, the group without zygomatic implants showed better results that can be explained by the fact that 85.7 % of the total antagonists are fixed implant prostheses or natural dentition, against 57.14 % in group I.