The purpose of this article is to present a new surgical technique – the double-barrel approach – employed for the extraction of dental implants from the maxillary sinus, and to compare it with the classic canine fossa puncture (CFP). The surgical steps in the performance of the double-barrel approach (DBA) are systematically described. In order to compare the two techniques (CFP versus DBA), a prospective, randomized study was carried out between 2004 and 2011, comparing the length of duration of the procedure and the incidence of adverse effects. The average duration was 6.3 ± 5.2 min for the DBA group and 14.8 ± 8.4 min for the CFP group ( P = 0.0001, Mann–Whitney test). Nevertheless, the incidence of adverse effects within the first postoperative week did not differ between the two groups. We have demonstrated that the double-barrel approach is faster, has the same incidence of adverse effects, provides excellent visualization, and permits a more accurate withdrawal of implants in comparison with the classic CFP. It is minimally invasive, requires no sutures, and minimizes the risks to local vascular and neural structures.
Migration of dental implants into the maxillary sinus, although an unusual but important complication, has lately been reported with increasing frequency. Even asymptomatic foreign bodies should be removed due to sequelae related to chronic inflammation. Furthermore, implant migration has been reported from the maxillary sinus to the ethmoid sinus, sphenoid sinus, and nasopharynx. Hence, judicious removal is imperative even in asymptomatic patients.
Foreign bodies can be removed from the maxillary sinus by three methods: (1) suction from the socket of an extracted tooth, (2) Caldwell–Luc approach, and (3) by the use of endoscopes. Endoscopic removal is performed either by transnasal route or by canine fossa puncture (CFP).
Endoscopic removal of dental implants has been described extensively. However, suction recovery and earlier endoscopic retrievals, while minimally invasive, are blind approaches. A Caldwell–Luc approach offers recovery under direct vision, but is associated with significant morbidity (cheek and lip oedema, face numbness, neuralgia, and collapse of the sinus cavity).
The purpose of this article is to illustrate a new surgical technique employed for the extraction of dental implants from the antrum – the double-barrel approach (DBA). Additionally, DBA was compared to the CFP technique in a prospective, randomized study. The length of duration of the procedures and incidence of adverse effects were compared between the two techniques.
Materials and methods
In order to compare the two techniques – CFP versus DBA – a prospective, randomized study was carried out between 2004 and 2011. The study had institutional review board approval. All patients were provided with thorough explanations of the two surgical techniques and signed an informed consent form. Patients with migrated dental implants within the maxillary sinus, with or without associated sinusitis, were included. Patients presenting oro-antral communications and patients with pre-existing paresthesia of the upper teeth or gums were excluded. Each patient had a preoperative sinus computed tomography (CT) to determine the position of the implant (see Fig. 1 ). Patients were randomly allocated (using a combination of triple and quadruple randomization blocks) to one of the two groups. Each patient was followed-up in the outpatient clinic for at least 1 year after surgery.
Surgery was performed only by the author (SA). The DBA technique can be performed under local anaesthesia and makes use of two trocars. The area of the oral vestibule between the superior canine and premolar tooth was topically anaesthetized with 10% lidocaine spray. After 5 min, the maxillary anterior wall periosteum was infiltrated with 5 ml of 1% lidocaine and adrenaline 1:100,000. In order to lessen the probability of anterior superior alveolar nerve injury, new anatomic landmarks illustrated by Robinson and Wormald were used: the tip of the first trocar was positioned at the point of intersection of a vertical line drawn through the pupil and a horizontal line drawn through the floor of the nose. The trocar was rotated parallel to the sagittal plane to penetrate the anterior wall of the maxillary sinus. The trocar was removed from the sheath and sinus diagnostic endoscopy was performed using the 45° endoscope. If secretions were noted, aspiration was done through the sheath. Afterwards, a second trocar was inserted through the anterior wall in close proximity to the first one; hence the technique is termed ‘double-barrel’ (see Fig. 2 ). Under visual control, biopsy forceps were introduced through the second sheath (see Fig. 3 ). The foreign body was grasped under endoscopic control and removed by extracting the forceps together with the cannula of the trocar. In sinusitis, copious irrigation of the sinus with saline solution was performed to break up any remaining infective materials. Ice was placed over the cheek area postoperatively to reduce the bruising and swelling associated with this approach. Patients with sinusitis were discharged on general antibiotics for 12 days and topical corticosteroids for 6 weeks.
In the control group, the classic CFP was performed. The implant was visualized by means of the 0° endoscope. Consequently, the trocar sleeve was placed in a direct line with the foreign body. Afterwards, maintaining the very same position of the sleeve, the endoscope was withdrawn and biopsy forceps were introduced. The dental implant was grasped blindly and removed by retrieving the forceps together with the trocar. Occasionally the process had to be repeated because the grasp was lost while the implant was passing through the maxillary sinus aperture.
Preoperatively, demographic data (age, gender) and the time interval between injury and implant retrieval were recorded. Intraoperatively, the duration of the procedure and the dental fixture characteristics (length, diameter) were assessed by the assistant surgeon. During the first postoperative week the author conducted a telephone survey of every patient to determine the rate of adverse outcomes: facial numbness, cheek pain, cheek swelling, teeth numbness, and gingival problems.
The statistical analysis was performed using SPSS version 10.0 (SPSS Inc., Chicago, IL, USA). Data were expressed as the mean ± standard deviation (SD). The distribution of the data was calculated by Smirnoff–Kolmogorov test. A parametric test such as the two-sample Student’s t -test was useful for data that followed a normal distribution. A non-parametric test such as the Mann–Whitney U -test was applied for data that did not follow a normal distribution. The occurrence of complications and the distribution of baseline features were compared between the two surgical groups by χ 2 test. P -values of <0.05 were considered significant.
According to the literature, the incidence of facial numbness within the first week after CFP is around 15%. Since DBA is a more distressing procedure in comparison to CFP, it was assumed that it would yield at least 25% facial paresthesia and pain. Based on 80% power to detect a statistically significant difference ( P = 0.05), it was calculated that at least 22 patients would be required for every group, and in order to raise the power, 25 cases were included in each group.
The time between injury and retrieval was separated into three intervals: less than 1 month, between 1 and 3 months, and between 3 and 12 months. The length of the implants was classified as short (6–9 mm), medium (10–12 mm), or long (13–16 mm).
A comparison of the baseline characteristics in the two surgical groups is presented in Table 1 . As demonstrated in Table 1 , randomization was efficient since the two groups were matched for age, gender, dental fixture characteristics, and time between injury and retrieval. The only difference between the two groups consisted of the length of the surgical procedure. Extraction of the implant using the DBA was significantly quicker than withdrawal by CFP: average time 6.3 ± 5.2 min for the DBA group and 14.8 ± 8.4 min for the CFP group ( P = 0.0001, Mann–Whitney test). Nevertheless, the incidence of adverse effects within the first week after the procedure was comparable between the two groups (see Table 2 ). After 3 months, paresthesia was reported by only 4% of the patients. Among the group of 50 patients, nine presented associated sinusitis. At the last follow-up, all patients were symptom-free and no endoscopic or CT signs of sinusitis were noted.
n = 25
n = 25
|Age, years, mean ± SD
|43.5 ± 17.6
|50.2 ± 20.5
|Student’s t -test
|χ 2 test
|χ 2 test
|χ 2 test
|Implant diameter, mean ± SD
|3.8 ± 0.49
|3.7 ± 0.48
|Student’s t -test