M. Maddalone & M. Gagliani
Restorative & Endodontic Department, School of Dentistry, University of Milan, Milan, Italy.

Introduction.
Surgical endodontics is a reliable method for the treatment of teeth with periapical lesions that do not respond to conventional root canal treatment (Harrison 1992, Gutmann & Harrison 1994, Danin et al.19 96, Briggs & Scott1997) or when orthograde retreatment is not feasible (el-Swiah & Walker 1996, Wada et al.1 998, Danin et al.1999, von Arxet al.20 01).
Successful outcomes have been reported in over 80% of cases in recent years (Rud et al.197 2a, Friedman et al. 1991, Jesslen et al.1995, August 1996, Chong et al.1 997, Mehlha ff et al.19 97, Bader & Lejeune 1998, Delano et al. 1998, Rud et al.2 001).T his high success rate may be due to modern surgical techniques (Carr 1992 a, b, Carr 1997), magnifying devices (Rubinstein 1997, Bahcall et al.1 999), microsurgery instruments (Kim 1997, von Arx & Walker 2000), ultrasonic retrotips (Min et al. 1997, Lin et al.1998, Morgan & Marshall1999, Gray et al. 2000, Peters et al.2 001) and improved root-end  filling materials (Adamo et al.1999, Rud et al.20 01).
The purpose of this paper is to report a 3-year follow up of a group of patients whose teeth were treated for periapical lesions of endodontic origin using modern surgical endodontic procedures.

Materials and methods.

Patient selection and inclusion/exclusion criteria.
One hundred and fifty-four teeth with periapical lesions, diagnosed radiographically and programmed for treatment with periradicular surgery in 101 patients, referred over a period of 2 years to the Department of Dentistry at the San Paolo Hospital in Milan, Italy, were included. Eight teeth in eight patients were extracted during the procedure because of vertical root fracture or lateral perforation; these were eliminated from the study. Thus, a final total of 146 teeth in 93 patients were included. Patients were recalled with in a 36-month time period; radiographs were taken at 1, 3, 6, 12, 24 and 36- month intervals. Some patients failed to regularly attend the follow-up visits and 26 teeth (17.8%) of 14 patients (15.05%) were eliminated because the patient did not attend three out of four recalls until the final year. All the patients included came to at least one of the fifth and sixth control appointments, 95% after 24 months and100% at 36 months.
Therefore, a group of 79 patients (Table 1) were followed over a 36-month time period. Amongst the120 teeth, 62 were anterior, 30 premolars and 28 molars (Table 2). All the teeth had root fillings with in 2.5 mm of the radiographic apex. The lesions affected one or more apices and varied in maximum diameter between 3 and12 mm. Clinical signs and symptoms were recorded preoperatively and also at each follow-up appointment. Clinical signs or symptoms after the first review were classified within the unsuccessful case group.

Table 1. Distribution by gender and age of the patients.

Surgical procedures.
All procedures were performed by one surgeon. Magnification loupes (4x) were used for all cases. An intrasulcular incision flap was used. Approximately 2-3 mm of the root apices were resected and the presence of gutta-percha and/or cement in the canal was con firmed.
Root end preparation using an ultrasonic technique was performed (PiezonMaster 400 and dedicated retro tips EMS, Nyon, Switzerland). The ultrasonic unit was used on medium power setting (Morgan & Marshall 1999). The canals were prepared to a depth of 3 mm using specifically designed tips (Berutti CT5, EMS).A zinc oxide-eugenol cement modified with EBA (Super Seal, Ogna Pharmaceuticals, Milan, Italy) was used as the root end  filling material. No space making or space maintenance biomaterials were placed in the wound. Bleeding was induced in the  field if it was absent at the end of the  filling and cleaning procedures (Gutmann 1993, Kim & Rethnam1997).N on-reabsorbable silk was used for suturing.
Antibiotics were prescribed starting 1 day before surgery and during the 4 days postoperatively (1 g amoxicillin every 12 h). Nonsteroidal anti-inflammatory drugs for 1 or 2 days were administered to limit postoperative pain A 4-mg dose of betamethasone disodiophosphate was administered immediately after the procedure to control swelling.

Radiographic centering.
Customized  film holders and ultra speed X-ray  film (Eastman Kodak Company, Rochester, NY, USA) were used throughout the study with a paralleling technique (Forsberg & Halse1997, Delano et al.19 98, Patel et al.2000). Radiographs were taken prior to surgery, after root end preparation, after root end filling but before suturing and at each follow-up appointment. Lesions were examined radiologically at 1, 3, 6, 12, 24 and 36-month intervals.

Radiographic examination.
Two endodontists with at least 10 years of clinical experience in endodontics were calibrated (Molven et al. 1987, 1996). All radiographs were independently examined twice (interval of almost 20 days) by each observer under 2x magnification and assigned to the appropriate category at the corresponding time (Figs 1-4).
Multirooted teeth were categorized on the basis of the root with the least favourable result. In the cases in which the two observers were not in agreement, the worst result was recorded. All the radiographs used to monitor healing of each case were evaluated according to the classification introduced by Molven et al.(1 987, 1996). In this way, each tooth was classified into one of the following categories:

• complete healing;
• incomplete healing;
• uncertain healing;
• unsatisfactory healing (failure).

All the clinical records were supplied to the observers. Teeth with clinical signs and symptoms reported at each of the control visits were categorized as failures. On the contrary, the first two categories were always associated with no clinical signs or symptoms.

Observers’ reproducibility assessment.
Inter- and intraobserver analyses were performed using Kappa statistics (Brunette1996).

Tooth.
All cases were divided into three groups based on the type of teeth, i.e. anterior, premolars and molars (Table 2).

Table 2. Distribution of the teeth included in the study and outcomes.

Post restoration.
The presence of posts was taken into account. For this reason the sample was further divided into two subgroups, teeth with (post group) or without posts (no-post group).

Statistical analysis.
The teeth were divided in to anterior, premolar and molar, and a Kruskal-Wallis test was performed at a significance level of P < 0.05 to assess differences between the groups. The sample was then divided into post and no-post groups as reported above, and the analysis was performed using the nonparametric Mann-Whitney U test at a significant level for P < 0.05. In addition, to simplify the comparison between post and no-post groups, complete and incomplete healing were added to the ‘successful’ cases and the other two categories (uncertain and unsatisfactory healing) to the ‘unsuccessful’cases. A chi-square was performed to compare the results.