J Periodontol • July 2004

Guided Tissue Regeneration WithBioabsorbable Barriers. II. Long-TermResults in Infrabony DefectsPeter Eickholz,* Diana-Maria Krigar,* Bernadette Pretzl,* Harald Steinbrenner,* Christof Dörfer,* and Ti-Sun Kim*

957

Background: The aim of this 5-year randomized controlledclinical trial was to evaluate the long-term results after guidedtissue regeneration (GTR) therapy of infrabony defects using twobioabsorbable barriers.

Methods: Fifteen pairs of contralateral infrabony defects in 15patients with moderate to severe periodontitis were treated. Eachpatient received one polydioxanon (test: T) and one polylactideacetyltributyl citrate (control: C) barrier by random assignment.At baseline, 12, and 60 ± 3 months after surgery clinical param-eters and standardized radiographs were obtained. Vertical bonelevels (PBL-V) were measured during surgery and 60 ± 3 monthslater by transgingival bone sounding.

Results: Thirteen patients were available for the 60-monthexaminations. Twelve and 60 ± 3 months after GTR, statisticallysignificant (P ≤0.001) vertical attachment (CAL-V) gain was foundin both groups (T12: 3.5 ± 1.5 mm; T60: 2.2 ± 1.8 mm; C12:4.0 ± 0.9 mm; C60: 2.4 ± 1.0 mm). However, from 12 to 60months after therapy both groups experienced significant CAL-Vloss (P <0.05): two defects in the test group and three in the con-trol group had CAL-V loss >2 mm compared to the 12-monthreexamination. Twelve and 60 ± 3 months after surgery, statisti-cally significant (P <0.05) radiographic bony fill was found in bothgroups (T12: 1.2 ± 1.3 mm; T60: 1.5 ± 2.2 mm; C12: 0.9 ± 1.4 mm;C60: 1.0 ± 1.6 mm). Further, 60 months after surgery significant(P <0.05) PBL-V gain was found in both groups (test: 1.8 ± 2.3mm; control: 2.2 ± 1.8 mm). The study failed to show statisticallysignificant differences between test and control regarding CAL-Vand PBL-V gain 60 months after surgery.

Conclusion: CAL-V gain achieved after GTR therapy in infrabonydefects using both bioabsorbable barriers was stable after 5 yearsin 21 of 26 defects (81%). J Periodontol 2004;75:957-965.

KEY WORDSClinical trials, controlled; clinical trials, randomized; follow-upstudies; guided bone regeneration; membranes, artificial;membranes, barrier; membranes, bioabsorbable; radiography,dental.

* Section of Periodontology, Department of Conservative Dentistry, Clinic for Oral, Dental,and Maxillofacial Diseases, University Hospital Heidelberg, Heidelberg, Germany.

Whereas conventional periodontaltherapy results predominantly inreparative healing, the potential

of bioabsorbable barrier membranes tofacilitate periodontal regeneration has beendemonstrated in several histometric stud-ies in animals1-3 and some histologicalstudies in humans.4,5 Clinical studies haveobserved more favorable results afterGTR therapy using bioabsorbable barriersthan conventional surgery in vertical bonydefects.6-12 Clinical studies comparingGTR therapy of infrabony lesions usingdifferent bioabsorbable barriers demon-strated similar results for short-term peri-ods: 5 to 6 months;13 12 months.10,14,15

Results after longer observation periods arescarce.16 Up to now there are only long-term studies (≥5 years after surgery)17

comparing the results after implantationof non-resorbable and bioabsorbable bar-riers in Class II furcation defects18 and ininfrabony defects.19

The objective of the present random-ized controlled clinical trial was to assessand clinically and radiographically com-pare long-term results and stability of GTRregenerative periodontal therapy usingbioabsorbable barriers in a population whohad been previously evaluated after 613

and 1214 months.

MATERIALS AND METHODSStudy PopulationFifteen patients (12 female; age range 22to 64 years, mean 42.1 ± 12.8), underperiodontal treatment at the Section ofPeriodontology, Department of Conserva-

30127.qxd 7/8/04 8:13 AM Page 957

958

Long-Term Results of GTR With Bioabsorbable Barriers Volume 75 • Number 7

tive Dentistry, Clinic for Oral, Dental, and MaxillofacialDiseases, University Hospital Heidelberg participated inthe present study.13,14 Nine patients were diagnosed withchronic periodontitis and six patients exhibited aggres-sive periodontitis. Each patient contributed one pair ofsimilar contralateral infrabony defects. Two defects weretaken as a pair when they fulfilled the following criteria:Both defects were of the same extent, located con-tralaterally in the same jaw (maxilla or mandible) andaffecting the same type of tooth (molar, bicuspid, orcanine). After enrollment, the study protocol, risks, ben-efits, and procedures were explained and informed con-sent obtained. All patients were asked about current andpast cigarette consumption. The study was approved bythe Institutional Review Board for Human Studies of theUniversity of Heidelberg.

Patients who currently smoked or had quit smokingfor less than 5 years were classified as smokers.20 Allpatients were retrospectively tested for the interleukin(IL)-1β polymorphism using a newly available test† kitat the 60 ± 3 month evaluation; i.e., a foam swab wasmoved over cheek mucosa for 20 seconds and thesample cells were sent to the laboratory for analysis.

Clinical ExaminationsThe gingival index (GI) and plaque index (PI)21 wereassessed at six sites per tooth (mesio-buccal, mid-buccal, disto-buccal, disto-lingual, mid-lingual, mesio-lingual) for each of the teeth scheduled for GTR therapy.Probing depths (PD) and vertical clinical attachmentlevels (CAL-V) were measured to the nearest 0.5 mmusing a pressure controlled straight, rigid, periodontalprobe‡ at the baseline and 12-month examinations.The cemento-enamel junction (CEJ) was used as areference for the CAL-V measurements. If the CEJ wasdestroyed by a restoration (filling, crown), the marginof this restoration served as the reference. All clinicalparameters for each patient were assessed by one offour examiners (CD, TSK, HS, or PE) who was maskedto the barrier material at the baseline and 12-monthexaminations. The clinical parameters (GI, PD, CAL-V,PI) were assessed again 60 ± 3 months post-surgeryusing a simple manual rigid probe§ by a single exam-iner (PE). During this examination, vertical probingbone level (PBL-V) was assessed at six sites per toothin all but two of the infrabony defects (one patientdeclined transgingival probing) using the straight peri-odontal probe under local anesthesia.

Radiographic ExaminationsAfter initial treatment (oral hygiene instructions, pro-fessional tooth cleaning, and subgingival scaling), bonestatus around teeth exhibiting symmetrical defects wasassessed through standardized radiographs using mod-ified film holders� which have been previouslydescribed.22,23 Intraoral dental films¶ size 2 wereexposed to an x-ray source# and developed under stan-

dardized conditions.** These individualized film hold-ers were used to obtain radiographs 12 and 60 ± 3months after surgery.

Periodontal SurgeryPeriodontal surgery was begun on the defect locatedon the patient’s left side. Treatment of the defect on theright side was performed at the same appointment or 1to 3 weeks later. Treatment assignment was madeaccording to a table of random digits;24 one side receiveda bioabsorbable polylactide acetyltributyl citrate bar-rier†† (control) and the other a bioabsorbable polydiox-anon barrier‡‡ (test). Neither barrier membrane used inthis clinical trial is commercially available. The experi-mental polydioxanon membrane was never marketed,and the polylactide membrane has been withdrawn.

Following an intracrevicular incision, a muco-periosteal flap was reflected exposing the bony marginof the defect by 5 mm.25 The flap was designed accord-ing to the modified papilla preservation technique toobtain primary closure of the barriers.26 After completeremoval of inflammatory granulation tissue, the rootsurfaces were thoroughly scaled and root planed. Thedistance from the CEJ to the alveolar crest (AC) andCEJ to the most apical extension of the bony defect(BD) (vertical bone level: PBL-V) was measured usingthe above-mentioned manual probe. The depth of theinfrabony component of the defects (INFRA) was cal-culated as the difference of the distances PBL-V minusthe smaller score of the buccal and lingual CEJ to ACmeasurements. The bony lesion was covered by thebarrier, overlapping the margin of the bony defect by3 mm. The membrane was adapted to the root surfaceby a suture around the root trunk. The mucoperiostealflap was repositioned to cover the membranes com-pletely. The flaps were sutured with polyglactin 910suture material.§§ GTR surgery was performed by oneof four authors (CD, TSK, HS, PE).

To prevent perisurgical infection 14 patients took 3 gamoxicillin 30 to 60 minutes before surgery. One patientwho had been identified as harboring Actinobacillus actin-omycetemcomitans subgingivally underwent both surg-ical procedures within a week, during which she took375 mg amoxicillin and 250 mg metronidazole threetimes daily. All patients rinsed with a 0.12% chlorhexi-dine gluconate solution� � for 2 minutes, two times daily

† Geno Type PRT Parodontitis-Risiko-Test, Hain Life Science GmbH,Nehren, Germany.

‡ TPS probe (universal explorer), IvoclarVivadent AG, Schaan,Liechtenstein.

§ PCPUNC15, Hu-Friedy, Chicago, IL.� VIP 2 Positioning, UpRad Corp., Ft. Lauderdale, FL.¶ Ultraspeed, Eastman Kodak Co., Rochester, NY.# Heliodent 70, 70 kV, 7 mA, Siemens, Bensheim, Germany.** Periomat, Dürr Dental GmbH, Bietigheim-Bissingen, Germany.†† Guidor Matrix Barrier, Guidor AB, Huddinge, Sweden.‡‡ Mempol, Ethicon GmbH & Co. KG, Norderstedt, Germany.§§ Vicryl, Ethicon GmbH & Co. KG.� � Chlorhexidin Mundspüllösung, Oral-B Laboratories GmbH, Frankfurt,

Germany.

30127.qxd 7/8/04 8:13 AM Page 958

for 5 to 7 weeks after surgery. Duringthis time, patients were instructed torefrain from mechanical plaque con-trol on the treated teeth and were seenat least every other week for plaquecontrol and gentle cleaning of the sur-gical areas. If barrier exposure wasnoted the patient was advised to usea 1% chlorhexidine gluconate gel¶¶

once a day. Thereafter, patients wereplaced on a maintenance scheduleincluding oral hygiene instruction andprofessional tooth cleaning once every3 months for the first 2 years. Later,most patients were seen every 4 to 6months. Each maintenance visit in-cluded the assessment of a full mouthbleeding (gingival bleeding index,GBI)27 and plaque score (plaque con-trol record, PCR).28 Patients who com-plied with at least one maintenencevisit a year were classified as regularattendees.18,19

Radiographic EvaluationAll radiographs were digitized usinga flatbed scanner## with 600 × 1,200dpi resolution, 10 bit gray valuesand then transferred to a personalcomputer*** and displayed on ascreen.††† Digital manipulations andmeasurement of linear distanceswere performed using a computerprogram.29‡‡‡ All radiographs wereevaluated under 8.5-fold magnifica-tion. Images were calibrated usingthe actual length of the metal wire asreference, which was entered into the computer pro-gram. A region of interest (ROI) that contained theentire defect was selected. For each of the 78 radi-ographs, the parameters described below were mea-sured. Bony defect depth (BD) was defined as the mostcoronal point where the periodontal ligament showeda continuous width. If no periodontal ligament could beidentified, the point where the projection of the alveo-lar crest (AC) crossed the root surface was taken asa landmark.30 If both structures could be identified onone defect, the landmark defined by the periodontal lig-ament was taken as BD and the crossing of the sil-houette of the alveolar crest with the root surface wasdefined as AC. If several bony contours could be iden-tified, the most apical that crossed the root was definedas BD and the most coronal as AC.23 Examples of theradiographs are shown in Figure 1.

Radiographs were randomized by the clinical exam-iner (PE) and numbered from 1 to 78.31 Thereafter, they

959

J Periodontol • July 2004 Eickholz, Krigar, Pretzl, Steinbrenner, Dörfer, Kim

were numerically analyzed beginning with number 1 bytwo examiners (D-MK, BP) masked to the clinical resultsand to the time a radiograph had been taken. All radio-graphic measurements were repeated after 1 week.

Statistical AnalysisThe main outcome variable chosen to compare thetherapeutic effects of test and control membranes waschange in CAL-V. PD change was considered a sec-ondary endpoint. All other clinical and radiographicparameters were considered control variables.

The patient was defined as the statistical unit. For eachpatient and barrier material, the deepest site within a ver-tical defect at baseline was evaluated and examined after12 and 60 ± 3 months. To reduce measurement error

¶¶ Corsodyl Gel, SmithKline Beecham, Bühl, Germany.## Epson Expression 1680pro, Seiko Epson Corp., Hirooka, Japan.*** Precision 340, Dell Computer Corp., Round Rock, TX.††† Nokia 447Xi, Nokia Display Products, Salo, Finland.‡‡‡ Friacom 2.5, Friadent AG, Mannheim, Germany.

Figure 1.Radiographic images of two contralateral angular infrabony defects (#9). Right mandibularsecond bicuspid: A) before; B) 12; and C) 60 months after surgery.Twelve months after surgerythe infrabony defect has almost filled up to the baseline alveolar crest (AC) with minimal boneresorption of the bone at the distal aspect of the adjacent tooth. From 12 to 60 months there issome levelling of bone with little further fill mesial of the premolar and some resorption at theadjacent tooth. Note the bony healing mesial of the first mandibular molar that was treatedconventionally. Left mandibular second bicuspid: D) before; E) 12; and F) 60 months aftersurgery.Twelve months after surgery the infrabony defect has filled to a level coronal of thebaseline alveolar crest (AC) with resorption of the interproximal bone ridge between premolarand first molar.This situation is stable 60 months after therapy. (CEJ: cemento-enamel junction;BD: most apical extention of bony defect.)

30127.qxd 7/8/04 8:13 AM Page 959

960

Long-Term Results of GTR With Bioabsorbable Barriers Volume 75 • Number 7

the means of the replicate measurements (CEJ-AC, CEJ-BD) of both radiographic examiners were entered intoanalysis. All parameters were tested for normal distribu-tion using the Kolmogorov-Smirnov/Lilliefors test. Forthe comparison of test and control treatment, the changesfrom baseline to 12 months as well as 60 ± 3 monthswere calculated as differences, and these differences fortest and control were compared by the paired t test orthe Wilcoxon signed rank test (GI and PI only). In addition,95% confidence intervals (CI) of the mean differencebetween the therapeutical results after 12 months as wellas 60 ± 3 months for the primary and secondary outcomevariables were calculated. Statistical analysis was per-formed using a computer program.§§§

RESULTSThirteen patients exhibiting 13 pairs of similar con-tralateral infrabony defects (10 predominantly 2-walland 16 predominantly 3-wall defects) were availablefor the 60 ± 3 months examination. One patient diedafter the 24-month maintenance examination due toreasons not related to this study. Another patient didnot respond to several invitations to the examination.The following teeth were included in this study: fourcanine (mandibular), eight bicuspids (six maxillary,two mandibular), and 14 molars (four maxillary, 10 §§§ Systat for Windows, version 10.0, Systat Inc., Evanston, IL.

mandibular). This information and other characteris-tics (age, defect site, membrane exposure, mainte-nance visits, mean full-mouth bleeding [GBI] andplaque [PCR] scores, smoking status, and IL-β poly-morphism scores) of patients who attended the 60 ±3 month examination are shown in Table 1.

Clinical ParametersThe healing phase passed uneventfully for all defectsexcept for one infrabony defect that developed an infec-tion 1 week after implantation of a polydioxanon bar-rier. After a prescription of 250 mg amoxicillin twice aday for 2 weeks, the infection disappeared (#6, test).At baseline, PI was low in both groups (test: 0.08 ±0.28; control: 0.23 ± 0.60). There was a small increaseof PI at the 12-month examination in both groups (test:0.46 ± 0.78; control: 0.62 ± 0.87) which remained stableat 60 months (test: 0.46 ± 0.78; control: 0.39 ± 0.77).There were no significant changes between the exam-inations or the treatment groups regarding PI. Therewere no significant differences in GI between bothgroups at baseline (test: 1.69 ± 0.75; control:1.54 ± 0.88). Twelve months after surgery, both groupsshowed statistically significant GI reduction (P <0.01)(test: 0.39 ± 0.77; control: 0.31 ± 0.75). However, 60

Table 1.

Patient and Healing Characteristics

MembraneExposure

Teeth (days after GTR)Regular Number Mean ± SD Mean ± SD IL-1β

Patient Age Test Control Defect Site Test Control Recalls of Recalls GBI PCR Smoking Polymorphism*

1 59 31 19 Mesial/distal 14 7 Yes 14 2.1 ± 2.4 18.1 ± 6.0 Never −

2 49 14 3 Mesial/distal – 7 Yes 8 1.3 ± 1.8 32.1 ± 17.4 Never −

3 45 13 5 Distal/mesial 14 14 No 10 5.7 ± 4.3 13.2 ± 5.1 Never −

4 54 19 30 Mesial 14 – Yes 12 1.3 ± 1.6 30.3 ± 11.4 Never −

5 35 21 27 Distal 7 – Yes 9 6.7 ± 4.6 27.0 ± 21.2 Active −

6 31 12 5 Mesial 7 28 No 9 0.8 ± 13.0 12.8 ± 10.3 Never +

7 27 5 13 Distal – 28 Yes 14 1.7 ± 2.1 5.2 ± 1.9 Never −

8 50 27 22 Mesial 14 14 No 8 7.4 ± 5.1 15.1 ± 11.2 Never −

9 32 29 20 Mesial/distal 7 14 Yes 11 4.2 ± 4.1 13.9 ± 7.6 Never −

10 22 19 30 Distal 28 – Yes 13 6.6 ± 4.0 18.8 ± 10.3 Never −

11 37 14 3 Distal 14 14 Yes 11 4.5 ± 3.5 28.9 ± 16.9 Former −

12 64 13 4 Distal 7 7 Yes 12 6.7 ± 4.9 19.3 ± 4.7 Active −

13 42 22 27 Mesial 28 7 Yes 12 9.3 ± 6.6 24.9 ± 11.1 Active +

* − indicates negative; + indicates positive.

30127.qxd 7/8/04 8:13 AM Page 960

961

J Periodontol • July 2004 Eickholz, Krigar, Pretzl, Steinbrenner, Dörfer, Kim

months after surgery, there were significantly morefavorable GI scores in the test group (P <0.05) (test:0.46 ± 0.89; control: 1.15 ± 0.99). Mean and standarddeviation of PD and CAL-V at baseline and 12 and 60months after surgery, and differences between test andcontrol are presented in Tables 2 and 3. Twelve and 60months after surgery, a statistically significant PD reduc-tion (P <0.001) and vertical attachment gain (P ≤0.001)could be observed for both test and control. In thecontrol group, a significant deterioration of PD wasobserved from 12 to 60 months (P <0.05), while the

PD increase in the test group was not significant (Table2). In both groups, statistically significant (P <0.05)mean attachment losses were observed from 12 to 60months (Table 3). No statistically significant differencesin PD or CAL-V between test and control could beobserved at baseline or 12 or 60 months post-surgically.

Table 4 gives the CAL-V at baseline, 12, and 60months after surgery for each patient and defect. Fivedefects showed attachment loss of >2 mm from 12 to60 months. Table 5 gives the distribution of CAL-V gainafter 12 and 60 months.

Bone ParametersA minor, statistically insignificant increase in the dis-tance from the CEJ to AC was observed for both testand control 12 months after surgery. For polydioxanon,the distance from CEJ to AC decreased from 12 to 60months; in the control group, a statistically significant(P <0.05) absorption of the AC was found 60 monthsafter surgery (Table 6). A statistically significant bonyfill, assessed as a decrease of the distance from CEJto BD (P <0.05), could be observed in both groups 12and 60 months after surgery (Table 6). Further, a sig-nificant (P <0.05) bony fill, assessed by intrasurgi-cal/bone sounding measurements, was found for bothgroups (Table 7). The analysis failed to reveal statis-tically significant differences between test and controlregarding radiographic and bone sounding parameters(Tables 6 and 7). There was a statistically significantcorrelation between bony fill as assessed by radiographsand bone sounding (PBL-V) (R2 = 0.24, P = 0.016).

DISCUSSIONSeveral studies have revealed more favorable clinicalresults after GTR therapy of infrabony defects withnon-resorbable or bioabsorbable membranes than afterconventional surgery.6-12 A meta-analysis of 28 stud-ies with a total of 847 infrabony defects treated pri-marily with access flap as a control in comparison todifferent regenerative methods revealed after a heal-ing period of at least 6 months a mean CAL-V gain of1.78 mm.32 Both groups in the current study showedmore favorable CAL-V gains 12 months as well as 60months after regenerative therapy with polydioxanon(12 months: 3.46 mm; 60 months: 2.23 mm) andpolylactide (12 months: 4.00 mm; 60 months: 2.39mm). Most clinical studies comparing CAL-V after ther-apy of infrabony defects using different regenerativemethods (non-resorbable and bioabsorbable barriers orenamel matrix protein) failed to observe statisticallysignificant or clinically relevant differences for obser-vation periods up to 4 years.13-16,33-38 Hence, thisstudy was undertaken to examine the long-term resultsof a 24-month longitudinal randomized controlled clin-ical trial comparing the results of GTR therapy usingtwo different bioabsorbable barriers in infrabony defectsthat had been published earlier.13,14

Table 2.

Probing Depth (mm) (13 patients)

Polydioxanon Polylactide(test) (control) Difference P

Baseline 7.31 ± 1.20 7.35 ± 1.94 0.04 ± 1.80 0.940

12 months 3.15 ± 1.35 2.42 ± 0.91 0.73 ± 1.18 0.046Change −4.16 ± 1.56 −4.93 ± 2.01 0.77 ± 2.10 0.211P <0.001 <0.00195% CI −0.50 to 2.04

60 months 3.69 ± 1.49 3.39 ± 1.12 0.31 ± 1.38 0.436Change −3.62 ± 1.61 −3.96 ± 1.349 −0.35 ± 1.81 0.503P <0.001 <0.00195% CI −1.44 to 0.75

Change 0.54 ± 1.68 0.97 ± 1.16 −0.42 ± 2.05 0.47112-60 monthsP 0.269 0.01195% CI −1.66 to 0.82

Table 3.

Vertical Probing Attachment Level (mm)(13 patients)

Polydioxanon Polylactide(test) (control) Difference P

Baseline 8.31 ± 1.32 8.35 ± 1.61 −0.04 ± 1.20 0.910

12 months 4.85 ± 1.56 4.35 ± 1.70 0.50 ± 1.53 0.261Change 3.46 ± 1.52 4.00 ± 0.91 −0.54 ± 1.69 0.273P <0.001 <0.00195% CI −1.56 to 0.48

60 months 6.08 ± 2.17 5.96 ± 1.87 0.12 ± 1.70 0.811Change 2.23 ± 1.80 2.39 ± 1.00 −0.16 ± 1.85 0.770P 0.001 <0.00195% CI −1.27 to 0.97

Change −1.23 ± 1.93 −1.61 ± 1.06 0.38 ± 2.24 0.54712-60 monthsP 0.040 <0.00195% CI −0.97 to 1.74

30127.qxd 7/8/04 8:13 AM Page 961

962

Long-Term Results of GTR With Bioabsorbable Barriers Volume 75 • Number 7

(CAL-V: 11.1 mm; INFRA: 7.2 mm).40

Characteristics of infrabony defects havebeen shown to influence periodontal heal-ing.9,31,39 Hence, these differences are inaccordance with the observation that baselineCAL-V and height of the infrabony compo-nent correlated positively with verticalattachment gain after GTR therapy. The fewstudies that report on long-term stabilityafter GTR therapy observed stability19 ormoderate CAL-V loss of 1.0 mm19 and 1.2mm,41 respectively. This is in accordancewith the CAL-V loss observed in the cur-rent test group. However, in the polylactidegroup a CAL-V loss of 1.62 mm wasobserved from 12 to 60 months. However,attachment loss from 12 to 60 months isnot distributed equally over all patients.Cortellini et al. reported that patientcharacteristics such as irregular attendanceat maintenance visits, smoking, and a gen-eral tendency toward attachment loss (loserpatients) accounted for most attachmentlosses.41 Other authors tried to explain thistendency to attachment loss by the pres-ence of the IL-1 gene polymorphisms.42 Inthe current study attachment loss >2 mmwas observed only in five of 26 defects: twoin the test (4, 8) and three in the control

Table 4.

Vertical Attachment Levels (mm) for Each Patient

Polydioxanon (test) Polylactide (control)

Patient Baseline 12 Months 60 Months Baseline 12 Months 60 Months

1 10.0 4.5 5.5 10.0 4.0 7.5*

2 9.0 4.0 4.0 9.0 5.0 8.0*

3 8.0 6.0 6.5 8.5 5.0 6.5

4 9.0 4.0 11.0* 9.0 7.0 8.0

5 9.5 6.0 7.0 10.0 5.0 7.0

6 8.5 6.5 7.0 7.0 3.0 6.0*

7 7.0 4.5 5.0 6.5 2.5 4.0

8 9.0 3.0 5.5* 6.0 2.5 4.0

9 8.5 6.0 6.0 9.5 5.5 6.0

10 5.0 2.0 3.0 6.0 2.0 3.0

11 7.0 3.0 3.0 7.0 2.5 3.0

12 9.0 5.5 7.5 10.0 6.0 8.0

13 8.5 7.5 7.5 10.0 6.5 6.5

* Loss of attachment >2 mm from 12 to 60 months after surgery.

Table 5.

Distribution of Vertical Probing AttachmentGain (13 patients)

VerticalPolydioxanon (test) Polylactide (control)

Attachment Gain 12 Months 60 Months 12 Months 60 Months

≤2 mm 3 8 1 6

>2 to ≤4 mm 6 3 9 7

>4 to ≤6 mm 4 2 3 –

Table 6.

Distance From Cemento-Enamel Junctionto Alveolar Crest (AC) and Bony Defect (BD)(mm) (13 patients)

Polydioxanon Polylactide(test) (control) Difference P

CEJ-ACBaseline 5.12 ± 2.15 5.13 ± 2.22 −0.01 ± 2.56 0.98312 months 5.38 ± 1.75 5.69 ± 2.23 −0.31 ± 2.12 0.604Change −0.26 ± 1.38 −0.56 ± 1.17 0.30 ± 1.51 0.488P 0.505 0.108

60 months 5.11 ± 1.82 5.86 ± 2.46 −0.75 ± 2.37 0.343Change −0.01 ± 1.81 −0.73 ± 0.99 −0.74 ± 1.85 0.240P 0.857 <0.05

CEJ-BDBaseline 8.86 ± 2.56 9.35 ± 2.99 −0.51 ± 2.27 0.44312 months 7.70 ± 2.80 8.41 ± 2.75 −0.71 ± 2.21 0.272Change 1.16 ± 1.31 0.94 ± 1.41 0.22 ± 1.89 0.697P <0.01 <0.05

60 months 7.31 ± 3.18 8.33 ± 3.12 −1.02 ± 2.77 0.213Change 1.55 ± 2.19 1.02 ± 1.63 0.53 ± 2.83 0.527P <0.05 <0.05

This clinical trial revealed similar or slightly bet-ter CAL-V gain in infrabony defects 12 months aftersurgery compared to other results reported for thesame time in infrabony defects with similar or morefavorable baseline characteristics: 3.04 mm (baselineCAL-V: 10.3 mm; baseline INFRA: 5.9 mm);9 3.0 mm(CAL-V: 8.3 mm; INFRA: 4.0 mm);34 3.1 mm (CAL-V: 7.8 mm; INFRA: 3.6 to 4.3 mm);10 4.1 mm (base-line CAL-V: 10.3 mm; INFRA: 5.8 mm).39 However,Cortellini et al. observed more favorable CAL-V gainsin defects with more pronounced baseline verticalattachment loss and infrabony defect depth: expandedpolytetrafuoroethylene: 5.2 mm (CAL-V: 10.8 mm;INFRA: 7.0 mm);40 bioabsorbable barrier: 4.6 mm

30127.qxd 7/8/04 8:13 AM Page 962

963

J Periodontol • July 2004 Eickholz, Krigar, Pretzl, Steinbrenner, Dörfer, Kim

Because neither barrier membrane usedin this clinical trial is commercially available,the results of this study cannot be used todecide which membrane material is moreeffective. However, the study adds to the smalldata pool of long-term results after regener-ative periodontal therapy.

Four authors (CD, TSK, HS, and PE) col-lected the clinical parameters and performedthe periodontal surgery. To prevent biaswithin a patient, surgery was performed byone author and clinical measurements byanother who was masked for the membraneused; this introduced the problem of inter-examiner reproducibility. The pressure-controlled probe might have provided somestandardization. The study, however, wasdesigned as a split-mouth investigation withboth sides measured by the same investigatorin any one case. Interexaminer reproducibil-ity, therefore, did not affect the comparisonof the two membranes. All measurements atthe 60-month examination were performedby one examiner (PE). This may have intro-duced bias. However, 5 years after surgery itwas very unlikely that this examiner still knewthe allocation of the different barrier materi-als when he performed the measurements.

Generally, radiographic vertical bone gainas assessed by radiographs and bone sound-ing was less pronounced than CAL-V gain.An earlier study demonstrated that connec-tive tissue attachment and bone level are twodifferent aspects of periodontal healing.8

Interestingly, whereas a significant CAL-Vloss was noted from 12 to 60 ± 3 months,

after 60 ± 3 months there was an increased bony fillcompared to 12 months post-surgery. However, thebony gain observed between 12 and 60 ± 3 monthsfailed to reach statistical significance.

The sample size (13 patients) of this study was toosmall to show equivalence of both barrier materialswith sufficient test power after 60 ± 3 months. For aclinically relevant difference (δ >1 mm), with a type Ierror α <0.05 and a standard deviation of the differ-ence between test and control for CAL-V gain of s =1.85 mm, a test power of 44% was calculated. A dif-ference in CAL-V gain between test and control after60 ± 3 months of 0.15 mm may certainly be consid-ered clinically irrelevant.

Within the limitations of the present study, we maydraw the following conclusions: The vertical attach-ment gains and bony fill achieved after GTR therapyusing bioabsorbable barriers can be maintained sta-ble up to 5 years after surgery in most infrabonydefects.

Table 7.

Depth of Infrabony Component (INFRA) at Baselineand Vertical Probing Bone Level (PBL-V) at Baselineand 60 Months (mm)

Polydioxanon (test) Polylactide (control)

INFRA PBL-V PBL-V INFRA PBL-V PBL-VPatient Baseline Baseline 60 Months Baseline Baseline 60 Months

1 7.5 10.5 5.0 6.5 11.5 8.5

2 4.0 8.0 5.5 2.0 10.0 9.5

3 3.0 9.0 8.5 5.0 7.0 7.5

4 5.0 9.0 12.0 5.0 10.0 10.0

5 5.0 8.0 10.0 5.0 12.5 7.5

6 5.0 10.5 * 4.0 6.5 *

7 3.0 6.5 6.5 3.0 5.5 4.0

8 4.0 8.0 7.0 6.0 9.0 5.0

9 4.5 10.0 7.0 5.5 10.0 7.0

10 6.0 6.5 4.0 3.5 6.0 4.0

11 5.0 8.0 3.5 4.0 7.0 4.0

12 5.5 10.0 9.0 10.0 14.5 9.5

13 6.5 9.5 9.5 6.0 12.0 9.0

Mean ± SD 4.9 ± 1.3 8.5 ± 1.3 6.8 ± 2.6 5.0 ± 2.0 9.3 ± 3.0 7.1 ± 2.3

PBL-V gain 1.8 ± 2.3† 2.2 ± 1.8†

* Did not participate in 60-month bone sounding.† Statistically significant difference from baseline (P <0.05).

group (1, 2, 6). Neither of patients 6 or 8 attendedmaintenance visits regularly, and patient 6 was IL-1positive. However, patient 4, who experienced themost pronounced CAL-V loss from 12 to 60 months,and patient 2 merely showed less stable individualoral hygiene (PCR >30%). For patient 1, no risk fac-tor on the patient level could explain the instabilityof one defect. Further, all patients experiencing >2mm CAL-V loss at one defect exhibited stability atthe contralateral site. Thus, it seems to be difficult toexplain these attachment losses by risk factors(smoking, IL-1, irregular maintenance) on a patientlevel alone. All sites with CAL-V loss >2 mm experi-enced membrane exposure. However, more stable sitesshowed early membrane exposure, and exposure hasbeen shown to predominantly affect short-term ratherthan long-term results.42 Perhaps the early exposuresites that suffered from CAL-V loss were those whereinfection control by 1% chlorhexidine gel was notsufficient.

30127.qxd 7/8/04 8:13 AM Page 963

964

Long-Term Results of GTR With Bioabsorbable Barriers Volume 75 • Number 7

REFERENCES1. Magnusson I, Batich C, Collins BR. New attachment

formation following controlled tissue regeneration usingbioabsorbable membranes. J Periodontol 1988;59:1-6.

2. Kon S, Ruben MP, Bloom AA, Mardam-Bey W, Boffa J.Regeneration of periodontal ligament using resorbableand nonresorbable membranes: Clinical, histological,and histometric study in dogs. Int J Periodontics Restora-tive Dent 1991;11:59-71.

3. Caffesse RG, Nasjleti CE, Morrison EC, Sanchez R.Guided tissue regeneration: Comparison of bioab-sorbable and non-bioabsorbable membranes. Histologicand histometric study in dogs. J Periodontol 1994;65:583-591.

4. Zappa U. Resorbable membranes (II). Periodontal regene-ration using resorbable membranes – histologic aspects(in German). Schweiz Monatsschr Zahnmed 1991;101:1321-1324.

5. Windisch P, Sculean A, Klein F, et al. Comparison of clin-ical, radiographic, and histometric measurements fol-lowing treatment with guided tissue regeneration or withenamel matrix proteins in human periodontal defects.J Periodontol 2002;73:409-417.

6. Cortellini P, Pini Prato GP, Tonetti MS. Periodontal regen-eration of human infrabony defects with bioresorbablemembranes. A controlled clinical trial. J Periodontol1996;67:217-223.

7. Eickholz P, Lenhard M, Benn DK, Staehle HJ. Periodontalsurgery of vertical bony defects with or without syntheticbioabsorbable barriers. 12 months’ results. J Periodontol1998;69:1210-1217.

8. Eickholz P, Hausmann E. Evidence for healing of interproxi-mal infrabony defects after conventional and regenerativetherapy: Digital radiography and clinical measurements.J Periodontal Res 1998;33:156-165.

9. Tonetti MS, Cortellini P, Suvan JE, et al. Generalizabilityof the added benefits of guided tissue regeneration in thetreatment of deep infrabony defects. Evaluation in amulti-center randomized controlled clinical trial. J Peri-odontol 1998;69:1183-1192.

10. Pontoriero R, Wennström J, Lindhe J. The use of bar-rier membranes and enamel matrix proteins in thetreatment of angular bone defects. A prospective con-trolled clinical study. J Clin Periodontol 1999;26:833-840.

11. Cortellini P, Tonetti MS, Lang NP, et al. The simplifiedpapilla preservation flap in the regenerative treatment ofdeep infrabony defects: Clinical outcomes and postop-erative morbidity. J Periodontol 2001;72:1702-1712.

12. Needleman I, Tucker R, Giedrys-Leeper E, WorthingtonH. A systematic review of guided tissue regeneration forperiodontal infrabony defects. J Periodontal Res 2002;37:380-388.

13. Dörfer C, Kim T-S, Steinbrenner H, Holle R, Eickholz P.Regenerative periodontal surgery in interproximal infra-bony defects with biodegradable barriers. J Clin Peri-odontol 2000;27:162-168.

14. Eickholz P, Kim T-S, Steinbrenner H, Dörfer C, Holle R.Guided tissue regeneration with bioabsorbable barriers:Intrabony defects and Class II furcations. J Periodontol2000 2000;71:999-1008.

15. Christgau M, Bader N, Schmalz G, Hiller K-A, Wenzel A.GTR therapy of infrabony defects using 2 different biore-sorbable membranes: 12-months results. J Clin Peri-odontol 1998;25:499-509.

16. Sculean A, Donos N, Miliauskaite A, Arweiler N,Brecx M. Treatment of infrabony defects with enamel

matrix proteins or bioabsorbable membranes. A 4-yearfollow-up split-mouth study. J Periodontol 2001;72:1695-1701.

17. Machtei EE. Outcome variables in the study of peri-odontal regeneration. Ann Periodontol 1997;2:229-239.

18. Eickholz P, Kim T-S, Holle R, Hausmann E. Long-termresults of guided tissue regeneration-therapy with non-resorbable and bioabsorbable barriers. I. Class II furca-tions. J Periodontol 2001;72:35-42.

19. Kim T-S, Holle R, Hausmann E, Eickholz P. Long-term results of guided tissue regeneration therapywith non-resorbable and bioabsorbable barriers. II. Caseseries of infrabony defects. J Periodontol 2002;73:460-469.

20. Lang NP, Brägger U, Salvi G, Tonetti MS. Supportiveperiodontal therapy (SPT). In: Lindhe J, Karring T, LangNP, eds. Clinical Periodontology and Implant Dentistry,4th ed. Copenhagen: Blackwell Munksgaard; 2003:781-805.

21. Löe H. The gingival index, the plaque index and theretention index system. J Periodontol 1967;38:610-616.

22. Duckworth JE, Judy PF, Goodson JM, Socransky SS. Amethod for the geometric and densitometric standardiza-tion of intraoral radiographs. J Periodontol 1983;54:435-440.

23. Eickholz P, Benn DK, Staehle HJ. Radiographic evaluationof bone regeneration following periodontal surgery withor without ePTFE barriers. J Periodontol 1996;67:379-385.

24. Werner J. Statistics in Medicine (in German). Munich:Urban & Schwarzenberg; 1984:13-14,253-257.

25. Zappa U. Resorbable membranes (I). Periodontal regen-eration using resorbable membranes–clinical aspects(in German). Schweiz Monatsschr Zahnmed 1991;101:1147-1151.

26. Cortellini P, Pini Prato G, Tonetti MS. The modified papillapreservation technique. A new surgical approach for in-terproximal regenerative procedures. J Periodontol 1995;66:261-266.

27. Ainamo J, Bay I. Problems and proposals for recordinggingivitis and plaque. Int Dent J 1975;25:229-235.

28. O’Leary TJ, Drake RB, Naylor JE. The plaque controlrecord. J Periodontol 1972;43:38.

29. Wolf B, von Bethlenfalvy E, Hassfeld S, Staehle HJ,Eickholz P. Reliability of assessing interproximal bone lossby digital radiography: Infrabony defects. J Clin Periodontol2001;28:869-878.

30. Benn DK. A review of the reliability of radiographic meas-urements in estimating alveolar bone changes. J Clin Peri-odontol 1990;17:14-21.

31. Klein F, Kim TS, Hassfeld S, et al. Radiographic defectdepth and width for prognosis and description of peri-odontal healing of infrabony defects. J Periodontol 2001;72:1639-1646.

32. Lang NP. Focus on infrabony defects – conservativetherapy. Periodontol 2000 2000;22:51-58.

33. Eickholz P, Kim T-S, Holle R. GTR with non-resorbableand bioabsorbable barriers. 6 months’ results. J Clin Peri-odontol 1997;24:92-101.

34. Cortellini P, Pini Prato G, Tonetti MS. Periodontal regen-eration of human infrabony defects with titaniumreinforced membranes. A controlled clinical study. JPeriodontol 1995;66:797-803.

35. Christgau M, Schmalz G, Reich E, Wenzel A. Clinical andradiographical split-mouth-study on resorbable versusnon-resorbable GTR-membranes. J Clin Periodontol 1995;22:306-315.

36. Caffesse RG, Mota L, Quinones C, Morrison E. Clinicalcomparison of resorbable and non-resorbable barriers

30127.qxd 7/8/04 8:13 AM Page 964

965

J Periodontol • July 2004 Eickholz, Krigar, Pretzl, Steinbrenner, Dörfer, Kim

for guided tissue regeneration. J Clin Periodontol 1997;24:747-752.

37. Christgau M, Schmalz G, Wenzel A, Hiller K-A. Periodontalregeneration of intrabony defects with resorbable andnon-resorbable membranes: 30-months results. J ClinPeriodontol 1997;24:17-27.

38. Eickholz P, Kim T-S, Holle R. Regenerative periodontalsurgery with non-resorbable and bioabsorbable barriers:Results after 24 months. J Clin Periodontol 1998;25:666-676.

39. Cortellini P, Carnevale G, Sanz M, Tonetti MS. Treatmentof deep and shallow infrabony defects. A multicenter ran-domized controlled clinical trial. J Clin Periodontol 1998;25:981-987.

40. Cortellini P, Pini Prato GP, Tonetti MS. Periodontal regen-eration of human infrabony defects with bioresorbablemembranes. A controlled clinical trial. J Periodontol 1996;67:217-223.

41. Cortellini P, Pini Prato GP, Tonetti MS. Long-term stabilityof clinical attachment following guided tissue regenera-tion and conventional therapy. J Clin Periodontol 1996;23:106-111.

42. De Sanctis M, Zucchelli G. Interleukin-1 gene poly-morphisms and long-term stability following guidedtissue regeneration therapy. J Periodontol 2000;71:606-613.

Correspondence: Dr. Peter Eickholz, Sektion Parodontolo-gie, Poliklinik für Zahnerhaltungskunde, Im NeuenheimerFeld 400, D-69120 Heidelberg, Germany. E-mail: peter_eickholz@med.uni-heidelberg.de.

Accepted for publication November 13, 2003.

30127.qxd 7/8/04 8:13 AM Page 965