J Oral Maxlllofac 50:210-217,
A Comparative Study of the Efficacy and Morbidity of Five Techniques for Ridge Augmentation of the Mandible PAUL MERCIER, DDS, FRCD(C),* HONGZHANG HUANG, DDS, MSc,t JOE CHOLEWA, DMD,$ AND SLOBODAN DJOKOVIC, DDSS Severely atrophic mandibles were augmented in 254 patients using five different surgical procedures, followed by total lowering of the floor of the mouth, vestibuloplasty, and skin graft. Patients were evaluated on a long-term basis to compare efficacy of treatment subjectively by recording patient complaints and by objectively assessing ridge form clinically and radiographically. Patients, in general, were pleased with the short- and long-term results of all five procedures. The ideal ridge form, comma-shaped, was obtained at a much higher rate with the three procedures using hydroxylapatite (HA) alone or in combination with collagen (HA-PFC). Ridge height loss both anteriorly and posteriorly was more severe in the two visor groups with bone graft or with HA than in those with only the alloplast. Labial sensory alterations more frequently occurred in the visor bone graft group, and wound dehiscence and lack of skin graft take were seen most often in the HA-PFC groups. Despite great differences in the quality of ridges obtained and in the number of complications between the visor groups and the alloplast groups, general patient satisfaction with any of the five procedures could be explained by the severity of the original complaints and because muscle interference has been eliminated in all cases by lowering of the floor of the mouth.
The main problem with mandibular dentures is that bone resorption and loss of ridge contour significantly affect their stability and retention. An obvious solution to severe bone loss is replacement using autogeneous or homogeneous bone grafts or alloplastic materials such as hydroxylapatite (HA). Because iliac or rib onlay grafts are also susceptible to resorption, newer surgical techniques have been introduced. One method of ridge augmentation involves elevating segments of the mandible by vertical (visor) or hor-
izontal (sandwich) osteotomies. These segments are left attached to soft tissues to assure blood circulation and to minimize bone resorption. Cancellous chips or corticocancellous blocks of bone and HA are used to supplement these procedures. Another approach to bone replacement is to augment the ridge solely with HA in particulate form or as a composite with collagen (HAPFC). The latter is used in blocks that become soft and malleable when saturated with blood. These blocks solve the problem of dispersion when HA particles alone cannot be maintained in place because of the unfavorable anatomic conditions present. It is the object of this article to compare the efficacy and morbidity of five different ridge augmentation procedures used to treat cases of severe atrophy of the mandible.
Received from the Maxillary Atrophy Clinic, St Mary‘s Hospital. Montreal, Quebec. Canada. * Oral Maxillofacial Surgeon and Director. t Research Fellow: Deputy Director. Department of Oral and Maxillofacial Sureerv. Facultv of Stomatoloav. Hubei Medical University. Wuhan. China. $ Clinical Associate in Prosthodontics. Address correspondcncc and reprint requests to Dr Mercier: Maxillary Atrophy Clinic. St Mary’s Hospital (‘enter. 3830 Ave Lacombe. Montreal, Quebec H3T I M5. Canada. I
of Oral and Maxillofacial
Two hundred fifty-four patients were selected from a larger group of 777 patients involved in a lo-year
MERCIER ET AL
prospective study on risks and benefits following residual alveolar ridge reconstruction. The main criterion for including patients in this study was the presence of severe or extremely severe atrophy of the mandible. The majority of cases belonged to the latter group where the basal bone was resorbed. Two hundred thirty-eight patients were female and 16 were male, with a mean age of 49 years. They had been wearing complete mandibular dentures for an average of 24 years. Patients were evaluated clinically and radiographically at yearly intervals postoperatively. The observation periods ranged from a minimum of 1 year (14 cases) to a maximum of 10 years (23 cases). The mean observation period for the whole group was 4.8 years. Twenty-six cases were lost to follow-up between the 2- and 5-year period, 5 between the 5- and 7-year period, and 3 between the 7- and IO-year period. Five techniques were investigated: group A (63 cases), total visor augmentation with a bone graft; group B (17 cases), anterior visor augmentation with posterior HA particles (Alveograf, Sterling Wintrop, Rensselaer, NY); group C (85 cases), total ridge augmentation with HA particles (Alveograf); group D (27 cases), total ridge augmentation with three HA-PFC blocks (Alveoform, Collagen Cot-p, Palo Alto, CA); group E (62 cases), augmentation with an anterior HA-PFC block (Alveoform) and posterior HA particles (Alveograf). The five augmentation procedures were not available at the same time. They were selected according to their periods of development and also in accordance with the residual anatomic structures present. When HA became available in particulate form, the total visor osteotomy with bone graft procedure was simplified by confining the bone cut and segment elevation to the anterior region in cases where the entire ridge could not be built up with HA particles. Limited soft-tissue dissection prior to insertion of the HA in the posterior tunnels usually would keep the material from migrating to adjacent tissues. When the preformed HA-PFC blocks were introduced, the augmentation technique was further simplified and performed on an outpatient basis by replacing the anterior visor osteotomy with a curved anterior block supplemented with HA particles posteriorly. When a concavity bordered by protruding genial tubercles was also present in the anterior region, the ridge was totally augmented with HA particles. Finally, when both the anterior and posterior regions were unfavorable for particle containment, three preformed HA-PFC blocks, one curved and two straight, were used. Extremely severe atrophy was also an indication for this approach in view of the potential for bony ingrowth with this material. With all five methods, a secondary skin graft procedure with release of the floor of the mouth and the vestibular musculature was initizlly planned and considered an essential step in securing retentive ridge forms. The surgical procedures
were done by the same surgeon and the prostheses made by the same prosthodontic team, usually within 6 weeks of the last surgery. The efficacy of treatment was analyzed subjectively by recording patients’ complaints, denture discomfort and stability, esthetics, and inadequate mastication of food and gastrointestinal (GI) symptoms before treatment and after 1, 3, and 6 years. Total elimination or a remarkable decrease in complaints after treatment both were rated as absence of complaints. The efficacy of the treatment also was evaluated clinically by an assessment of the ridge form obtained: concave or flat, inverted V, inverted ll, or comma-shaped. The last form was defined as the ideal ridge form for denture retention. It involves a wide, convex surface with deep lingual undercuts and a fixed mucobuccal fold (Fig 1). Ridge resorption was also assessed by measuring mandibular height on lateral cephalograms (Fig 2) before and immediately after augmentation, after 5 months, and at 1, 2, 3, 4, 5, 7, and 10 years. Two sites were selected because they showed the least measurement distortion, 20 and 30 mm from point 0, the projection of the pogonion on the mandibular plane. These heights, ht 20 and ht 30, corresponded to the caninefirst premolar and second premolar-first molar areas. The anterior region was also evaluated by measuring the angulation of the anterior external slope. Treatment complications were recorded at five different periods: after augmentation (surgery 1), after ridge extension with skin graft (surgery 2), and after 1, 3, and 6 years. These involved wound dehiscence with or without release of a bone sequestrum or loss of alloplastic material, bone or implant displacement, and lingual and/or labial neurosensory alterations. Any complaints of anesthesia or dysesthesia were confirmed by objective tests. Paresthesia was reported as a minor sensory change. Skin graft take was evaluated at the time of splint removal. At yearly recall visits, the skin graft donor site was examined for the presence of hypertrophic scars. The oral recipient site was also evaluated for the development of multiple foci of hyperkeratosis or keratolysis. The data collected were analyzed using the dBase IV Data Management System (Ashton-Tate, Torrance, CA). Nonparametric data for gastrointestinal symptom evolution after surgery were statistically analyzed with the x2 method. Results PATIENTSYMPTOMS
A major reduction in denture discomfort and improvement in denture stability and facial esthetics was reported by most patients for all five procedures both on short- and long-term evaluation (Figs 3-5). However, the long-term gain in denture stability was greater in
FIVE TECHNIQUES FOR RIDGE AUGMENTATION OF MANDIBLE
FIGURE 1. Comma-shaped reconstructed ridge.
the groups where an alloplast instead of a bone graft was used. The best results in relief of denture discomfort were seen with group C, where only HA particles were used. Decreased denture discomfort was associated with improved denture stability. Symptoms in the upper (57%) and lower (42%) digestive tracts were reported preoperatively, as well as the habit of swallowing food unchewed (84%) (Figs 68). Gastrointestinal symptoms and masticatory evolution after surgery were analyzed statistically using contingency tables (x2). The positive effect of oral surgical rehabilitation on the GI symptoms and on mastication was found to be highly significant at the 1-year visit (P 5 .OOOI). Further reduction of these symptoms, especially those in the lower digestive tract with a relief of constipation, was reported on a long-term basis. RIDGE FORM
76% of groups C, D, and E and only in 11% of groups A and B. Group C rated best at 83%. The second ideal form, the inverted U-shaped one, was produced twice as frequently with procedure D and almost three times as much with procedure B than with C and E. The least retentive type, the small inverted V, was obtained at a much greater rate in the visor groups than with those in which alloplasts were used (A + B = 39% vs C + D f E = 2%). RIDGE AUGMENTATION
The average increase in ridge height (Table 2) was from 10.9 to 19.0 mm (+75%), with the group D procedure providing the largest gain, 92%, and group B the least augmentation (58%). The gain with the total HA-PFC procedure (D), 8.6 mm, corresponded closely
The best ridge forms were obtained in the nonvisor groups (Table 1). Comma-shaped ridges were seen in
Total _ HA
D Told HA-PFC
E 113 HA-PFC1
FIGURE 2. Diagram of lateral cephalometric measurement technique of ridge height and anterior external slope angle.
FIGURE 3. Denture discomfort before and after treatment.
MERCIER ET AL
0 FIGURE 4.
Denture stability before and after treatment.
Although the observation periods and number of individuals in each group were unequal, observations on patterns of resorption (or compaction) can be made for both the anterior (slope) and midposterior (ht 20 to 30) regions of the mandible (Figs 9, 10). The pure particulate group (C) showed much less compaction in the initial stage (5 to 6 months) than the pure HA-PFC group (D) in accordance with the nature of the material used (ridge height loss: C = 3%, D = 14%; slope loss: C = 4%, D = 17%). However, the loss with the dense particulate group (C) increased constantly to reach the same plateau as that of group D after a few years. The
0 Vbor - HA
D Total HA-PFC
FIGURE 6. ment.
to the height of the block inserted, 9 mm. Results were similar for the anterior slope angulation. There was an almost threefold increase for group D and a 48% increase for group B in comparison with 92% for group A. The smaller gain in group B is explained by the tendency for backward rotation of the short visor segment not supported by cancellous bone as in group A.
A Visor - Bone
E 113HA.PFCI 2!3 HA
I Pre 1
Gastric symptoms after eating before and after treat-
decrease does not appear to stabilize after the 5-year period, but the number of observations for this group and group B at this period is small to draw conclusions. The worst overall performance was displayed by the visor-bone graft groups, which also had the longest observation period (in some cases up to 10 years). After the 5-year period, some minor resorption was seen in the posterior regions. COMPLICATIONS
Stage 1 Surgery
Major labial neurosensory alterations were four times more frequent in the visor-bone graft group, which involved bone sectioning along the mandibular canal (Table 3). Wound dehiscence with or without material loss was most frequent with the total HA-PFC blocks followed in frequency by the combined block and particle procedure. Material displacement was not seen in the total block group, but occurred in one in two cases with the visor-HA procedure when the os-
Total’ - HA
1Total HA-PFC D 1 113 HA-PFC E 1 2/3HA I
Pie 1 3
Esthetics before and after treatment.
Intestinal symptoms before and after treatment.
A Visor. Bone
Total - HA
D Total HA-PFC
E l/3 HA-PFC/ 213 HA
FOR RIDGE AUGMENTATION
The overall incidence of keratolysis of skin grafts in the mouth was low for all groups. It was mostly seen in those visor cases that underwent severe postoperative resorption. Discussion RIDGE
FIGURE 8. treatment.
Habit of swallowing food unchewed before and after
teotomy and the insertion of particles were performed simultaneously. Stage 2 Surgery The incidence of major labial sensory changes after the first surgery did not change after lowering of the floor of the mouth, vestibuloplasty, and skin graft, except in group E. There was a higher incidence of labial paresthesia after this surgery for the visor-bone graft group. There were a few cases of lingual paresthesia that disappeared after a few months. Major lack of skin graft take, defined as a loss over 2 cm in length, occurred most frequently with the collagen groups. Material loss after the second procedure also was more frequent with these groups. Follow-up
In all groups except group D (two cases), there was continuous improvement in lip sensitivity at the l-year visit and thereafter. Hypertrophic scars at the donor site, some requiring treatment, were visible in 6% of the whole group at the l-year visit. This complication was reduced to 4% (one patient) after 3 years. Less apparent, but visible, donor sites, with the presence of pale or pigmented skin, were seen in 15% of cases at the 1-year visit; these continued to improve with time.
When comparing data, the total HA-PFC approach with three blocks stands out as the best method for an immediate increase of height in the anterior and posterior regions, but is second to particulate HA augmentation in securing ideal ridge form. This is due to the configuration of the U-shaped, preformed posterior blocks inserted. Less inward inclines on the lingual side are developed when the floor of the mouth is lowered than when HA particles, which can be better molded over the lingual side, are used. This disadvantage is counterbalanced in part by the development of large denture-bearing surfaces that enhance denture retention. The worst ridge forms were developed by the visor-bone graft procedure, which allows elevation of only a rigid, vertical osseous pedicled segment that does not provide the necessary lingual undercuts. No previous studies have had the opportunity to evaluate and compare diRerent ridge augmentation procedures for such a large group of patients on a longterm basis. We had determined earlier’ that the mean postoperative height loss using the visor osteotomy with the bone graft approach was 25% after 4 years, figures similar to those reported in Peterson’s study.* With longer observation periods, the same decreasing exponential curve was confirmed, with an additional 7% loss in the midbody region after 10 years. The anterior region is similarly affected, with an additional 6% loss. We also have shown previously in a comparative study3 of visor bone grafts and visor-HA procedures that there is less ridge height loss in the HA-augmented section than in the bone graft one, and similar losses for the anterior visor segments. Longer observation periods confirm that the intense bone remodeling period is almost terminated after 5 years. It is consistent with patterns of resorption seen after dental extraction, as
Preoperative and Postoperative Ridge Forms (%)
Flat or concave Inverted V Inverted U Comma
98 2 -
3 48 46 3
88 12 -
29 53 18
73 27 -
Post 2 15 83
92 4 4 -
90 10 -
2 3 18 77
Table 2. Immediate Postoperative Gain of Mandibular Height and Anterior Slope A
Ht 20-30 (mm) Pre Post Gain (%) Slope (degree) Pre Post Gain (%)
25.7 49.3 92
29.1 43.1 48
28.9 43.1 49
23.5 44.7 90
24.8 45.1 90
16.7 45.3 171
of the anterior crest moves in a posterior direction with a loss of anterior vertical slope that stabilizes after a few years. The major and unexpected finding of this study was the small but continuous decrease of ridge height seen in the HA particle segment of the visor-HA procedure and in the total particulate HA augmentation procedure. Our findings on HA particles compare with figures from other studies;4s54% to 10% loss of ridge height in the first year. Compaction of the material is being held responsible for these early changes, but after 1 or 2 years, the compaction period should have ceased and no more loss should have taken place unless the bone underneath resorbs. Since apparently this is not the case in those mandibles that have completed their postextraction resorption phase, questions must be asked about the resorbability of this material. Does the material truly resorb or is it replaced by tissues not dense enough to be detected on radiographs? Is the purity of HA a determinant factor? Does the addition of collagen provide more stability to the augmented ridge? These questions could be best answered by reentry procedures and by longer-term comparative studies using different brands of HA. In the only histologic study of HA-PFC, Mehlisch6 found bone infiltration throughout three of the five specimens analyzed. New bone was directly opposed to the particles without evidence of resorption or for-
postoperative loss of slope angle. A, Visor bone; D. total HA-PFC, E, HA-PFC and HA.
visor HA; c, total HA:
FIGURE 10. Postoperative loss of mandibular height (ht 20, 30). A, Visor bone; B, visor HA; c, total HA, D, total HA-PFC; E, HAPFC and HA.
eign body reaction. Donath,’ who had access to a whole segment of resected mandible that had been previously augmented, has noted a dissolution of particles, with granules of HA within the cytoplasm of macrophages. We found* in a recent radiologic study of a large group of mandibles augmented with both HA alone or HA-PFC, a phenomenon of fragmentation that sometimes leads to total disappearance of the material on the radiograph. A radiolucent zone that might suggest encapsulation and foreign body reaction is first observed with islands of material separating from the mass. However, these observations seldom correlate with the clinical findings. Of 380 cases analyzed, only 9 showed comparable major radiographic and clinical changes of ridge form. PATIENT
Improvement in denture efficacy is difficult to evaluate. The literature is replete with subjective studies reporting patient satisfaction or lack of complaints following surgical attempts at ridge reconstruction or just simply after replacement of ill-fitting prostheses. The fact that patients were generally pleased with treatments despite unequal surgical results attests to the severity of the functional handicap present that, in many instances, affects the general health of the individual. This study highlights the important role that masticatory function may play in the digestive processes; a role that is often ignored or challenged in the medical literature. The results were similar to those of another study9 that we did in association with a gastroenterologist. The severity of the masticatory handicap present is also underlined by the fact that the degree of satisfaction does not appear to be particularly affected by the incidence of complications. The most frequent complications were neurosensory alterations, mostly from the visor-bone graft procedure; wound dehiscence after insertion of the HA-collagen blocks; and ulceration usually with some release of particles at the time of skin
216 Table 3.
FIVE TECHNIQUES FOR RIDGE AUGMENTATION
Complications After Stage 1 and 2 Surgery and After 1,3, and 6 Years (%) Surgery
Follow-up 1 yr
Follow-up 3 yrf
Follow-up 6 yr+
ABCDEABCDEABCDEABCDEABCDE Sensory changes Major Minor Wound dehiscence and lack of skin graft take Major Minor Major loss HA or bone HA or bone displacement Keratolysis Visible donor site Major Minor
6 1 18 7
09 0 0
0004336 8 6 14
1 18 23
16 14 24 26 12
graft splint removal. All of these complications were seen less frequently in the HA particles group. Visible donor sites were also reported, but these sites improved steadily after the l-year visit so that only very few hypertrophic scars remained clearly visible after this time. The improvement was not similar for the labial sensory complication. The 49% incidence of major and minor changes increased with the visor-bone graft cases to 66% after the vestibuloplasty procedure, which is more difficult to perform than an alloplastic augmentation. After the l-year visit, the incidence decreased, but mostly in the 3-year period and less afterwards. The high incidence of wound dehiscence (33%) after HA-PFC block insertion, or of major ulceration of the crest (16%) at the time of skin graft splint removal, calls for discussion. Melhisch6 also reported frequent dehiscence (39%), but did not discuss it. As often is the case with HA, the delay in healing is usually limited to one small area and does not grossly affect ridge form, as it involves only the superior aspect of the crest and not the periphery of the augmented ridge. There is seldom loss of large portions of material, as healing takes place around each particle. However, because the incidence of this complication with the composite was three times that with HA particles alone, it raises some questions about this complication. Local rejection of a foreign animal protein by certain patients and the propensity for contamination by oral fluids of a spongelike material that must undergo considerable compaction in the early period after insertion are factors that could be held responsible. Another reason for delayed healing could be the quantity of material used in relation to the bone and periosteal surfaces present, as the blocks were often used in the most severely atrophic cases.
000400 5 18 13
Two complications, displacement of HA particles and visible changes in the keratin layer of the oral skin graft, seldom occurred. The incidence of the first complication, often quoted as the main disadvantage of HA in particulate form, was greatly reduced by the techniques used. Soft-tissue elevation was limited during insertion, as the ridge will take its final form only after the second procedure. Contrary to the findings of a recent study by Hillerup,” who reported 38% major and minor skin graft changes in the mouth, keratolysis was infrequently seen except in some cases in the visor groups that had undergone severe resorption. Instructions to patients to brush the skin graft regularly to eliminate accumulated desquamation, and annual prosthetic checkup visits, are probably responsible for the low incidence reported. The evolution that has taken place in ridge augmentation with HA in one form or another has given rise to procedures that are causing less morbidity to patients, with a much higher degree of predictability than with former approaches. The ease of augmenting any ridge, even the most severely atrophic one, on an outpatient basis cannot overshadow the fact that a second surgical phase is required unless contraindicated by the general health of the patient. Muscle detachment under direct vision to define ridge contour allows one to compensate for possible surgical problems created by the first surgery. If the ridge form is not well outlined, as seen with the visor-bone graft cases, at least the reconstructed ridge is free of muscle interference. The ridge remains broad and covered with tissues of better quality than the original fragile mucosa. Therein might be found a determinant factor, the skin graft procedure, to explain the discrepancies between overall patient satisfaction with any of the five ridge augmentation
MERCIER ET AL
procedures and differences in quality of surgical results obtained. The findings are in support of the opinion that the skin graft procedure is a necessity after ridge augmentation of the mandible. References 1. Mercier P, Zeltser C, Cholewa J, et al: Long term results of mandibular ridge augmentation by visor osteotomy with bone graft. J Oral Maxillofac Surg 45:997, 1987 2. Peterson LJ, Slade CW: Mandibular ridge augmentation by a modified visor osteotomy: A preliminary report. J Oral Surg 35:399, 1977 3. Mercier P, Zelstser C: Visor osteotomy augmentation of the mandible with posterior onlay bone graft or with hydroxylapatite: A comparative study. Oral Surg 62:25 1, 1986 4. Kent JN, Quinn JH, Zide MF, et al: Alveolar ridge augmentation
using non-resorbable hydroxylapatite with or without autogenous cancellous bone. J Oral MaxiBofac Surg 41:629, 1983 Block MS, Kent JN: Long term radiographic evaluation of hydroxylapatite augmented mandibular alveolar ridges. J Oral Maxillofac Surg 42:793, 1984 Mehlisch DR: Collagen-hydroxylapatite implant for augmenting deficient alveolar ridges: A 24-month clinical and histologic summary. Oral Surg 68:505, 1989 Donath K, Rohrer MD, Beck-Managetta J: Histologic evaluation of mandibular cross-section one year after augmentation with hydroxylapatite particles. Oral Surg 63:65 1, 1987 Huang H, Mercier P: Resorption or substitution of hydroxylapatite with or without Collagen in mandibular ridge reconstruction. J Dent Res 70:199 I (abstr 1954) Poitras P, Mercier P: Masticatory deficiency and GI symptoms: Evolution of GI symptoms after surgical reconstruction of atrophic mandibular ridge. Gastroenterology 96(5). 1989 (abstr 341) Hillerup S, Hjorting-Hansen E, Eriksen E, et al: Influence ofskin graft pathology on residual ridge reduction after mandibular vestibuloplasty. Int J Oral Maxillofac Surg 19:212, 1990