The purpose of this study is to compare the exposure rate of 3 different barrier types after a guided-bone regeneration procedure as well as to compare the percentage grafted bone dimensional loss with and without exposed barriers. Patient records from September 2007 to May 2015 were reviewed to identify subjects who had received a bone graft followed by implant placement procedure after the graft had completely healed. The subjects were divided into 3 groups: (1) resorbable barrier, (2) nonresorbable barrier, and (3) titanium-mesh barrier. Incidences of barrier exposure were recorded. Cone-beam computerized tomography images before treatment (T0), right after grafting (T1), and after healing (T2) were used to determine the percentage of grafted bone dimensional loss and am quantitative amount of grafted bone remaining (mm2). Three cross-sectioned areas, at 1-mm apart, of preplanned implant positions at the grafted site were measured using cone-beam computerized tomography to calculate the remaining grafted bone and grafted bone dimensional change. The exposure rate of all guided bone regeneration was 36.9%. The exposure rate of the resorbable barrier (23.3%) was significantly lower than titanium mesh (68.9%) and nonresorbable (72.7%; χ2, P < .001). The results of this study revealed that barrier types have a significant effect on the exposure rate. There was also a significant difference in grafted bone dimensional loss between sites with barrier exposure (58.3%) and sites with no barrier exposure (44.1%) during the healing period (Mann-Whitney U test, P = .008).
Introduction
Dental implants have been used successfully to replace missing dentition.1,2 In the event of insufficient alveolar ridge height or width, guided-bone regeneration (GBR)3,4 with a barrier is commonly used to reconstruct the alveolar ridge before implant placement.
Barriers used for GBR can be categorized as resorbable (RS), nonresorbable (NRS), and titanium mesh (TM).5,6 Although grafted bone tends to heal better with barrier,7,8 studies have reported high rates of barrier exposure, which increases the risk of infection and compromises the graft.4,9,10 Studies have shown a wide range of mean barrier exposure rates, regardless of barrier types used: RS, 14.4% (0%–29%); NRS, 15.1% (2.5%–45.5%), and TM, 23.5% (0%–80%).4,10–40
It has been reported that the level of complications differs with different barriers types when exposed.41,42 NRS are known for exposure and dehiscence, which lead to early barrier removal,27,43 compromising the survival of the grafted bone.7,44,45 Although a higher exposure rate (23.5%) has been reported with TM, the formation of a pseudo-periosteum layer directly underneath the exposed site sometimes protects the underlying bone graft from infection.28,39 TM exposure often results in less than the original grafted bone volume after the sites heal.39 Although RS has the least exposure and complications of the three barriers, while lacking rigidity, its barriers have limited application in cases with extensive vertical augmentation.46
Aside from type of barrier used during the graft, other factors may also influence the exposure rate. These factors include the graft material, location of the graft, purpose of grafting, surgeon experience, the patient's smoking habit, and the use of interim.
The methods studies have used to evaluate the consequence of graft barrier exposure were reentry for direct measurement, marginal bone level change, volumetric analysis, and clinical photograph evaluation.47–55 Even though various methods were used, many studies had small sample sizes with fewer than 25 samples.20–34,36–39 To date, there have been no published reports comparing the exposure rate and consequences on bone dimensional change (mm2) with the 3 different barrier types (RS, NRS, and TM).
The purposes of this retrospective study are to compare the exposure rate, timing, and complication management of three barrier types (RS, NRS, and TM) used in GBR procedures as well as its effect on grafted bone dimensional loss (GBDL) at the sites where implants were planned.
Materials and Methods
Patient selection
This retrospective study was approved by the institutional review board of Loma Linda University. The treatment records of patients who underwent GBR and subsequent implant placement between September 2007 and May 2015 were reviewed.
Inclusion criteria
Potential subjects for exposure rate comparison were required to meet all of the following criteria:
Older than 18 years at the beginning of treatment
Had undergone a GBR procedure without implant placed simultaneously using one of the following barriers: (1) RS (Bio-Gide, Geistlich Pharma, Wolhusen, Switzerland), (2) NRS (Cytoplast, Osteogenics Biomedical Inc., Lubbock, TX, and Gore-Tex, Goremedical, Flagstaff, AZ), or (3) TM (RidgeForm Mesh, Osteomed, Addison, TX; Figure 1).
Additional criterion for GBDL comparison:
- ○
Had undergone cone-beam computerized tomography (CBCT) before GBR procedure, right after the GBR procedure, and before or right after implant placement
- ○
(a) Resorbable (RS) barrier. (b) Titanium mesh (TM). (c) Nonresorbable (NRS) barrier.
(a) Resorbable (RS) barrier. (b) Titanium mesh (TM). (c) Nonresorbable (NRS) barrier.
Data collection
Patient demographics (age and gender), tobacco usage (smoking or nonsmoking), date of GBR, graft materials, purpose of grafting, barrier types (RS, NRS, or TM), exposure occurrence, time of exposure (early [<1 month] or delayed [>1 month]), exposure management (barrier removal, oral hygiene instruction, antibiotics or combination), and infection rate were recorded. Barrier exposure rates were reported in the following categories:
Types of barriers used: RS, NRS, or TM
Graft materials: autogenous (AU), allograft (AL), xenograft (XE), and combination (AU + AL, AU + XE, AL + XE, and AU + AL + XE)
Graft locations: maxillary posterior, maxillary anterior, mandibular anterior, and mandibular posterior
Purpose of grafting: horizontal (H), vertical (V), and combined (HV)
Surgeons' surgical experience: >5 years (E) or <5 years (LE)
Smoking habit
Report of interim prothesis usage: fixed provisional, complete denture, removable partial denture, or Essix retainer
Imaging measurements
Cross-sectioned images were measured using CBCT (i-CAT, Imaging Science International, Hatfield, Penn) before GBR (T0), right after GBR (T1), and before or right after implant placement (T2). The CBCT images were transferred and opened using an implant planning software program (InVivo 5.2, Anatomage Dental, San Jose, CA). The superimposition function was used to measure the same area from T0, T1, and T2 CBCT (Figure 2). A preplanned implant position (P1) was selected as the point of measurement based on the radiographic template. Additional implant positions were selected 1 mm mesial (P2) and distal (P3) to P1 (Figure 3).
Superimpositioning of 2 cone-beam computerized tomography images (blue and orange) by using the superimposition feature in InVivo 5.2.
Superimpositioning of 2 cone-beam computerized tomography images (blue and orange) by using the superimposition feature in InVivo 5.2.
Transverse view of the cone-beam computerized tomography image to determine the position of the sagittal section in which the measurement will be made. P1 is the preplanned implant position based on radiographic template. P2 is the position 1 mm mesially to P1. P3 is the position 1 mm distally to P1.
Transverse view of the cone-beam computerized tomography image to determine the position of the sagittal section in which the measurement will be made. P1 is the preplanned implant position based on radiographic template. P2 is the position 1 mm mesially to P1. P3 is the position 1 mm distally to P1.
(a) Cross-sectioned area before guided-bone regeneration (T0) at one of the preplanned implant positions. (b) Cross-sectioned area of the ridge with newly grafted bone (gray; T1). (c) Cross-sectioned area after healing with healed grafted bone (blue) and lost bone (red; T2).
(a) Cross-sectioned area before guided-bone regeneration (T0) at one of the preplanned implant positions. (b) Cross-sectioned area of the ridge with newly grafted bone (gray; T1). (c) Cross-sectioned area after healing with healed grafted bone (blue) and lost bone (red; T2).
A total of 232 patients (89 men and 143 women) with a mean age of 56.8 years (range, 21 to 88 years) were included in this study. A total of 271 GBR sites were categorized into 3 different groups: 193 RS, 45 TM, and 33 NRS (Figure 1). The mean QGBP, QGBR, and PGBDL of the exposed and nonexposed groups were calculated for each type of membrane. All measurements and data collections were performed by one examiner (K.P.).
Statistical analysis
The barrier exposure rates associated with each recorded parameter were represented using descriptive statistics. The chi-square test was used to compare the incidence of exposure of all barriers and the incidence of exposure among the three barriers. The PGBDL of the exposed groups and nonexposed groups among each barrier were compared using independent samples Mann-Whitney U test. The PGBDL among the exposed barriers was compared by independent-sample Kruskal-Wallis test. Intraexaminer reliability was tested using the SPSS intraclass correlation coefficient (ICC) and was considered adequate at .90. The level of statistical significance was set at P< .05. The methodology was reviewed and approved by an independent statistician.
Results
Measurements of T0, T1, and T2 cross-sectioned images of 27 randomly selected implant sites were collected for the intraexaminer reliability test. The measurements were repeated at least 2 weeks apart by one examiner (K.P.). The reliability of the measurement methods was expressed as an ICC of >.95, which suggests that the measurement methods were reliable and reproducible. The power analysis revealed >.90 with a current sample size of 271. All results were reviewed by an independent statistician.
Exposure rate
There was a total of 100 (36.9%) exposed barriers. Of these, early exposure (<1 month; 14.14 days [3–29 days]) was observed in 63 sites and late exposure (>1 month; 3.12 months [1–8 months]) in 37 sites. The overall incidence of barrier exposure of RS (23.3%) was significantly lower than that of NRS and TM (72.7% and 68.9, respectively; P < .001; Table 1). No significant different was found in the barrier exposure rates between NRS and TM (P = .71; Table 1).
The most frequently used graft material was AL (110 cases). The least common graft material used was XE by itself (7 cases). Aside from XE having a 0% exposure rate, the exposure rate of the other graft materials ranged from 31.5% to 66.7% (Table 2).
The exposure rate based on the location of the GBR is reported in Table 3. The highest exposure rate was found in the mandibular posterior (42.7%) and the lowest was found in the mandibular anterior position (29.4%; Table 3). The exposure rate based on the purpose of grafting is shown in Table 4. Horizontal augmentation has the least exposure rate at 30.1%, whereas vertical augmentation has the highest exposure rate at 65%.
Eighty-four GBR procedures (52 RS, 6 NRS, and 26 TM) were performed by surgeons with >5 years of experience (E group), whereas 187 GBR procedures (141 RS, 27 NRS, and 19 TM) were done by surgeons with <5 years of experience (LE group). Exposures were noted in 29 (34.5%) in the E group and 71 (37.9%) in the LE group. The exposure rate between experienced surgeon and surgeon with limited experience was not statistically significant (P = .295; Table 5).
Comparison of barrier exposure rate between experienced surgeons and surgeons with limited experience using χ2 test at α = .05

Fourteen GBR (9 RS, 1 NRS, 4 TM) were performed in smokers and 257 GBR (184 RS, 32 NRS, 41 TM) in nonsmokers. Despite the small sample size of smokers in this study, barrier exposure was almost 2 times greater in smokers (57.1%) than in nonsmokers (35.8%; Table 6).
Table 7 compares the barrier exposure rate from the use of different types of interim prostheses. There were 81 cases with interim prosthesis usage reported. The exposure rate for cases with interim prosthesis reported (37.0%) was similar to that of cases with no interim prosthesis usage (36.8%). The removable Essix retainer had the highest exposure rate at 51.5%.
Effect on grafted bone
A total of 89 GBR sites with 130 implants underwent CBCT at T0, T1, and T2 and were included for evaluation of grafted bone dimensional change. Of these, 56 GBR and 77 implants were with RS, 12 were with GBR, and 20 implants were with NRS, whereas 21 GBR and 33 implants were with TM. The mean healing time between T1 and T2 was 7 months (range, 2–20 months). The mean QGBP of each group is shown in Table 8. The barrier type with the highest mean QGBP was NRS (73.05 mm2). The barrier type with the least QGBP was RS (59.09 mm2). The mean QGBPs of exposed and nonexposed barriers were not significantly different for all barriers (P > .05; Table 8).
Comparison of quantitative grafted bone placed (mm2 ; T1) independent-sample t test at α = .05 (N = 130)

The comparisons of PGBCL are shown in Table 9. The overall PGBDL of the exposed group (58.3%) was significantly higher when compared with the PGBDL of the nonexposed group (44.13%; P = .008; Table 9). The NRS had the greatest PGBDL at 63.0% and the RS group and TM had PGBDLs of 49.9% and 42.5%, respectively. The PGBDL of exposed RS (54.4%) and nonexposed RS (48.7%) were not significantly different (P = .421; Table 9). The PGBDL of exposed NRS (68.7%) and exposed TM (52.1%) were significantly higher than that of nonexposed NRS (12.2%) and nonexposed TM (27.8%; NRS: P = .011, TM: P = .040; Table 9). No statistically significant difference in PGBDL was observed among different barriers in the exposed group (P = .24; Table 9). Correlations between QGBP and PGBDL were analyzed using the Pearson correlation test (Figure 5). There were no associations observed in any of the groups (P > .05; Figure 5a–e). The sample size of the nonexposed NRS was too small to convey a meaningful statistical interpretation (Figure 5f).
Comparison of percentage grafted bone dimensional loss (%) using independent-sample Kruskal-Wallis test and independent-sample Mann-Whitney U test (N = 130)

Pearson correlation between quantitative grafted bone placed (mm2) and percentage grafted bone dimensional loss (%).
Pearson correlation between quantitative grafted bone placed (mm2) and percentage grafted bone dimensional loss (%).
Of 100 exposed cases, 83 received initial treatment with chlorhexidine and antibiotics and 17 did not receive any treatment. Sixteen of 100 exposed cases and 3 of 171 nonexposed cases developed infections. Ten of the infections were RS, 6 were NRS, and 3 were TM. Similar initial treatment was given to the infected cases. Removal of barrier and graft and subsequently regraftings were done when the graft did not respond to the initial treatment. Six of 7 barrier removals were performed in infected cases. Three infected cases from the nonexposed group responded to the initial treatment and did not require further treatment.
Discussion
The 2 main principles for successful regeneration are exclusion of gingival tissue during the bone formation process and minimization of bacterial contamination of the grafted bone under the barrier. Regardless of the types of barrier used, studies have shown less bone regeneration when exposure occurs.7,44,45,56–60 In the present retrospective study, the exposure rate of the different types of barriers used for guided bone regeneration was observed, and the effect of exposure on the grafted bone was compared by evaluating PGBDL.
Previous short- and long-term studies conducted independently of each other and involving different type of barriers used during GBR have reported the following mean barrier exposure rates: RS, 14.4% (0%–29%); NRS, 15.1% (2.5%–45.5%); and TM, 23.5% (0%–80%).4,10–40 Comparatively, higher rates of barrier exposure (Table 1; RS, 23.7%; NRS, 68.9%; and TM, 72.7%) were observed in this retrospective study. The exposure rate of the RS group was significantly lower than that of the NRS and TM groups (P < .001; Table 1). The observed exposure rate in this study agreed with that of previous studies that a higher incidence of dehiscence, barrier exposure, and/or premature barrier removal are noted when NRS and TM barriers are used.61,62
High exposure of NRS and TM observed in the present study might be associated with the barrier characteristics. NRS has been shown to have cell-occlusive properties, which can obstruct and delay the healing mechanism of the wound, leading to exposure.61,63 Previous studies reported that insufficient fixation of TM could be the cause of exposure, and its rigidity could increase the risk of mucous membrane complication.28,29
Sites with autogenous grafts had the highest exposure rate (66.7%) in this study, while sites with xenografts had no exposure. However, the graft material may influence the exposure rate. The result from this study is inconclusive, as the barriers used on each type of graft material were not controlled.
Rocchietta et al41 reported that a similar complication rate can be found between maxillary augmentation (33.5%) and mandibular augmentation (31.1%). Results shown in Table 3 were similar to rates from previous studies. The exposure rate of the maxillary grafted site was 34.5%, whereas the exposure rate of the mandibular grafted site was 40.6% (Table 3); hence, the location of the graft might not contribute to the exposure rate.
A previous study by Donos and Rocchietta also reported that as many as 45.5% of vertical augmentation cases may experience complications, whereas horizontal augmentation has a lower complication rate (11.1%–24.4%).41,51 The result from this study coincided with the previous studies, as the barrier exposure of V (62.5%) was higher than of H (30.1%) and HV (45.1%; Table 4). These results may relate to the lower predictability of the vertical augmentation.41,51
It has been reported that the experience of the surgeon and the complexity of the surgical procedure have a strong influence on surgical success.64 In this study, surgeons' experience did not affect the barrier exposure rate, as the barrier exposure rate of the E group (34.5%) was not statistically significantly different from that of the LE group (37.9%; P = .295; Table 5). Perhaps the statistical similarity observed between the two groups of operators showed that types of barriers were more influential on the exposure rate in this study.
Smoking was reported to have a negative effect on GBR.53,65,66 A previous study showed that 94.7% of nonsmokers have a successful GBR compared with 62.5% of smokers.67 In this study, 57.1% of patients in the smoking group experienced barrier exposures, whereas only 35.8% of patients in the nonsmoking group did (Table 6). Although the sample size of the smoking group was too small to enable a statistical conclusion to be drawn, the rate of exposure of the smoking group was almost double that of the nonsmoking group. Patients must understand the effects of smoking on the healing process, and habit modification is recommended prior to surgery.
Cases with interim prosthesis usage reported exhibited similar exposure rate to cases with no report of interim prosthesis usage (Table 7). Based on the data from this study, the use of provisional may not affect the rate of exposure, although the type of provisional might. A removable form of provisional has been known to exert pressure on the grafted or implant site and may cause membrane exposure.68 The results showed that the exposure rates of the 3 removable interim prostheses (removable complete denture [28.6%], removable partial denture [27.8%], Essix retainer [51.5%]) were higher than that of the fixed provisional (26.1%; Table 7).
It was proposed that a larger graft can cause more tension to the soft tissue, which could lead to barrier exposure.4,69 However, the results of this study disagree and show that the QGBPs of the exposed and nonexposed groups were not statistically significantly different (P > .05; Table 8). Nevertheless, tension-free primary closure should be strived for, regardless of the size of graft placement.
It was reported that barrier exposure during healing had a major negative effect on GBR.7,44,45 Machtei53 reported that new bone formation for a nonexposed barrier was 6-fold greater than for the exposed site. Simion et al9 reported a significantly lower mean regeneration rate when the membranes were exposed (41.6%) compared with nonexposed membrane–treated sites (96.6%). The results of this study are in agreement with those of previous studies. The overall PGBDL of the exposed group (58.3%) was significantly higher than the PGBDL of the nonexposed group (44.1%; P < .05; Table 9). This study's findings also agreed with a previous study done conducted by Lizio,39 who found a positive correlation between exposure and grafted bone volume change.
There was no statistically significant difference in PGBDL among the exposed barrier types (P = .242; Table 9). However, there was a statistically significant difference in PGBDL among the nonexposed barrier types (P < .01; Table 9). Of the nonexposed barrier types, nonexposed RS had the highest PGBDL (48.7%); however, it was not significantly different when compared with the PGBDL of exposed RS (54.4%; P = .421; Table 9). The similarity in the PGBDL in the exposed RS groups and nonexposed RS group may be due to the poor mechanical properties and shorter resorption period of the RS barrier, resulting in a similar PGBDL in both the exposed and nonexposed groups.28 There were significant differences in PGBDL when comparing the exposed TM (52.1%) and NRS (68.7%) barrier to the PGBDL of the nonexposed TM (27.8%) and NRS (12.2%) barrier (P < .05; Table 9). Based on the data from this study, barrier exposure has a greater effect on PGBDL in TM and NRS barriers. When working with TM and NRS, practitioners should use the correct soft-tissue management technique to minimize the risk of exposure of the grafted site.
PGBDL was not associated with QGBP (P > .05; Figure 5a–f). This result applied to both the exposed and nonexposed groups of all barrier types. Nonexposed NRS was not statistically conclusive in the correlation tests because of the low sample size (Figure 5f). The scatter plot (Figure 5) revealed that up to 100% PGBDL was observed in all barrier types when exposed and in RS only when not exposed. Most of the nonexposed TM and NRS sites exhibited less than 50% PGBDL. Based on the results from this study, practitioners should be aware that PGBDL was not related to the amount of graft placed but rather to the exposure status of the barrier, especially when dealing with TM or NRS (Figure 5). The periosteum of the flap should be adequately released and evaluated before suturing to minimize flap tension and the risk of barrier exposure.4
Barrier exposure did not lead to definite failure of the augmentation, and an implant can still be placed in a site where membranes have been exposed. Treatments of choice for barrier exposure in this study were chlorohexidine rinse only or in conjunction with antibiotics, which have been mentioned in previous studies.39,66,67 Barrier exposure can cause bacteria colonies to form on the barrier,9 and chlorohexidine can help reduce the bacteria population.66 Maintaining appropriate oral hygiene with the method mentioned above is highly recommended when the barrier is exposed.
This retrospective study has some limitations. One limitation was the use of existing data. The current data set contained an uneven number of cases for each of the barrier groups and other parameters. Some groups had a much smaller sample size, which reduced the statistical power of the findings. Another limitation is the lack of uniformity in recording the procedure notes, which caused missing data and hence sample exclusion. Perhaps standardizing the chart recording, such as using a template note, may improve the availability of the data. On the other hand, one strength of this study was the long observation period, which allowed for a satisfactory sample size. Future studies should be conducted by incorporating another dimension (length) to measure the volume of the graft and should aim to control the type of graft material used. Furthermore, a future controlled prospective clinical study could improve data quality by eliminating the absence of data on confounding factors.
Conclusions
In this study, the cumulative barrier exposure rate was 36.9% among all 3 types of barriers. The results of this study suggest that a higher exposure rate is associated with the type of barriers used: RS (23.3%), NRS (72.7%), and TM (68.9%). For the other variables, either there was not a statistically significant difference in the exposure rate or there was not an adequate sample size to determine the statistical value.
Percentage grafted bone dimensional loss was significantly different between the overall exposed group (58.3%) and the nonexposed group (44.1%); however, there were no significant differences among the exposed groups of different barrier types: RS (54.4%), TM (52.1%), and NRS (68.7%). Barrier exposure of TM and NRS was more influential on PGBDL than barrier exposure of RS. No correlation was found between the quantitative value of grafted bone placed and the percentage of grafted bone loss (P > .05). Based on the results of this study, the exposure rate of TM or NRS and its effect is more severe when compared with RS. Flap management is critical to reduce the risk of exposure.
Abbreviations
- AU:
autogenous
- AL:
allograft
- B1:
quantitative grafted bone placed
- B2:
quantitative grafted bone remained
- B3:
quantitative bone dimensional loss
- CBCT:
cone-beam computerized tomography
- E:
experienced
- GBDL:
grafted bone dimensional loss
- GBR:
guided-bone regeneration
- H:
horizontal
- HV:
horizontal + vertical
- ICC:
intraclass correlation coefficient
- LE:
limited experience
- NRS:
nonresorbable barrier
- P1:
preplanned implant position
- P2:
1 mm mesial to P1
- P3:
1 mm distal to P1
- PGBDL:
percentage grafted bone dimensional loss
- QGBDL:
quantitative grafted bone dimensional loss
- QGBP:
quantitative grafted bone placed
- QGBR:
quantitative grafted bone remained
- RS:
resorbable barrier
- T0:
before GBR
- T1:
right after GBR
- T2:
before or right after implant placement
- TM:
titanium mesh
- V:
vertical
- XE:
xenograft
Note
The authors declare no conflict of interest. This research was conducted in partial fulfillment of a master of science in dentistry degree for Kiddee Poomprakobsri.