The Role of Timing of Loading on Later Marginal Bone Loss Around Dental Implants: A Retrospective Clinical Study

Different concepts of loading for dental implants have been proposed in the literature. The original protocol included healing periods of 3 and 6 months before loading for the mandible and maxilla, respectively.1–4 More recently, healing periods have been reduced to minimize treatment time. Immediate/early loading therefore is an attractive treatment modality, and numerous clinical follow-up studies have been published on this topic during the past 10 years.5 On the basis of reviews of the literature, it has generally been concluded that this is a safe procedure in the anterior mandible, and immediate/early loading for other indications is not well documented.5,6 

One of the most important success criteria that should be evaluated is the marginal bone level around implants.7 Long-term clinical evaluation of dental implants and their superstructure is crucial, and measurement of marginal bone loss (MBL) over time is regarded as a sensitive tool for evaluation of the clinical performance of implants because gradual loss of marginal bone eventually will lead to implant failure. A pathologic decrease in bone level could lead to loss of bone anchorage of the implant. During the first year, marginal bone resorption of a maximum of 1.5 mm has been accepted, whereas MBL of 0.2 mm annually is considered acceptable for subsequent years.8,9 

In a recent review comparing bone reactions of immediate loaded implants vs those of conventional loaded ones, the following findings were discussed.10 Immediately loaded implants showed approximately 0.6 mm bone loss in the first 12-month period, and the same amount or more in the second year. In contrast, conventionally loaded implants exhibited almost the same loss as immediately loaded implants in the first year but smaller magnitudes of loss in the second year. Early-loaded implants showed the least amount of bone loss in the first and second 12-month periods. It is unknown from this review whether bone loss continues beyond 24 months or becomes stable. This limits the ability to draw any conclusions beyond this period, and long-term follow-up studies are awaited.

The aim of this retrospective study was to assess the relation between the MBL around dental implants after a loading period of 36 months and the timing of loading and draw conclusions.

A total of 87 patients with 252 implants, who appeared in routine recall sessions consecutively at 6, 12, 24, and 36 months after loading, were included in the study group. Information was given to each patient regarding alternative treatment options. All subjects were required to be (1) at least 18 years old, (2) able to read and sign the informed consent document, (3) physically and psychologically able to tolerate conventional surgical and restorative procedures, and (4) willing to return for follow-up examinations as outlined by the investigators.

All implants were placed by the same surgeon and were restored by several prosthodontists at the University of Istanbul, Faculty of Dentistry, Department of Prosthodontics.

The fixed restorations and mandibular overdentures were evaluated separately.

The loaded implants were investigated in 2 groups: smooth collar implants (Straumann, Waldenburg, Switzerland; Swissplus, Zimmer Dental, Carlsbad, Calif) and microgeometry neck implants (Astratech, Mölndal, Sweden; Biolok-Biohorizons, Birmingham, Ala).

Implant sites included all interforaminal mandibular anterior positions for overdenture (OVD) support in the overdenture group (51 patients: 34 female and 17 male; mean age, 59.39 ± 9.99 years, with 126 implants), whereas all mandibular and maxillary anterior and posterior positions were found in the fixed restored group (36 patients: 21 female patients with 73 implants and 15 male patients with 53 implants; mean age, 54.97 ± 12.24 years, with 126 implants). In the fixed restored group, 106 implants were splinted, whereas 20 have been single. In the OVD group, the implants had no neighboring teeth, but in the fixed restored group, the neighboring teeth were recorded so their influence on MBL could be evaluated. A total of 81 implants had no neighboring teeth, 32 implants had 1 neighboring tooth, and 13 implants had teeth on both sides (mesially and distally).

The loading period was 36 months for all evaluated implants. The time span from implant placement to beginning of loading of the implants in the removable restored group was evaluated in 4 classifications (<60 days, 60–90 days, 91–140 days, >140 days), and the fixed restored group in 3 (<91 days, 91–180 days, >180 days). The groups were selected in such a manner that the implant counts were evenly distributed for better statistical assessment.

Recalls were routinely performed 6, 12, 24, and 36 months after loading. At each recall session, a clinical examination was performed by the same examiner.

The following clinical parameters were recorded: plaque index (PI), sulcus bleeding index (SBI), peri-implant probing depth (PD), and peri-implant MBL. The PI was recorded visually with the aid of a probe moved along the abutment surfaces. For each implant, the presence of plaque was scored on 4 surfaces (0  =  no plaque; 1  =  plaque on probe; 2  =  plaque on implant seen by the naked eye; 3  =  abundance of soft matter).11 The SBI, defined as bleeding elicited 20 seconds after a periodontal probe is moved 1 mm into the mucosal sulcus parallel to the abutment wall, was scored (0  =  no bleeding; 1  =  isolated bleeding spots visible; 2  =  blood forms a confluent red line on margin; 3  =  heavy or profuse bleeding). PD was measured with a periodontal probe (PW, Hu-Friedy, Chicago, Ill) at 4 sites around each abutment. After baseline scores were recorded, the periodontal examination was performed at each recall session, and the scores were recorded.

Panoramic radiographs (Planmeca, Proline XC, Helsinki, Finland) were taken preoperatively, immediately after surgery, immediately after loading, and at every recall session. In cases of insufficient quality, intraoral radiographs were taken as well. Mesial and distal marginal bone levels of all implants were determined at baseline and recall evaluations. Measurements were obtained from images of successive radiographs, which were scanned and digitized (Epson 1680 Pro, Seiko Epson Cooperation, Nagano, Japan) before, and were analyzed at ×20 magnification with the use of a software program (CorelDraw 11.0, Corel Corp and Coral Ltd, Ottawa, Canada) (Figure 1). The known diameter of the implant at the collar region according to the manufacturer's dimensions of the respective implants was used as a reference point. The distance from the supracrestally widest part of the implant to crestal bone level was measured on the magnified images. To account for variability, the implant dimension (width) was measured and compared with the documentation dimensions; ratios were calculated to adjust for distortion. Bone levels were determined by applying a distortion coefficient (true bone height is equal to true implant width multiplied by bone height as measured on the radiograph, which is then divided by the implant diameter measured on the radiograph). The actual bone level measurement was performed independently by 2 examiners (a prosthodontist and a specialist in oral and maxillofacial radiology). The average from the 2 examiner calculations was used as the marginal bone level value. The level at which the marginal bone seemed to be attached was assessed by visual evaluation at the distal and mesial surfaces of all implants.

Statistical analyses were utilized in this study to assess mean marginal bone level changes at 6, 12, 24, and 36 months, as well as to explore the potential effects of the timing of loading on bone loss. For statistical analysis of the results, NCSS 2007 and PASS 2008 Statistical Software (NCSS, Kaysville, Utah) was used. Descriptive statistics (means and standard deviations for continuous variables, frequencies for categorical variables) was applied for the entire sample of implants and subsequently for each type of implant. Comparison of quantitative data was accomplished with 1-way analysis of variance (ANOVA), and the group causing the difference was identified with the use of Tukey's honestly significant difference (HSD) test. The parameters of 2 groups were compared with the Student t test. For analysis of repetitious measurements, variance analysis was used, whereas for determination of the group causing the significance, the paired sample t test was chosen. Results were evaluated at a significance level of P < .05.

A total of 87 patients with 252 implants were evaluated in this retrospective study. No incidences of excessive bone loss at implants or peri-implant inflammation were noted.

In terms of timing of loading, a statistically significant difference was noted between sixth month distal bone loss averages (P < .01) (Table 1). At the sixth month, distal bone loss levels of the implants where the loading started after 140 days were significantly higher than bone loss levels of the implants loaded earlier than 60 days (P  =  .001; P < .01), between the 60th and 90th days (P  =  .037; P < .05), and between the 91st and 140th days (P  =  .001; P < .01). No significant difference was reported between the other loading timing groups in terms of distal bone loss rate at the sixth month (P > .05).

In terms of loading timing, a statistically significant difference was seen between 12th month distal bone loss averages (P < .01). At the 12th month, distal bone loss levels of the implants where loading started after 140 days were significantly higher than bone loss levels of implants loaded earlier than 60 days (P  =  .001; P < .01), and between the 91st and 140th days (P  =  .001; P < .01). At the 12th month, distal bone loss levels of the implants where loading started between the 60th and 90th days were significantly higher than bone loss levels of implants loaded earlier than 60 days (P  =  .001; P < .01), and between the 90th and 140th days (P  =  .024; P < .05). No significant difference between the other loading timing groups in terms of distal bone loss rate was noted at the 12th month (P > .05).

In terms of loading timing, a statistically significant difference was observed between 24th month distal bone loss averages (P < .01). At the 24th month, distal bone loss levels of the implants where the loading started after 140 days were significantly higher than bone loss levels of the implants loaded earlier than 60 days (P  =  .001; P < .01), and between the 91st and 140th days (P  =  .001; P < .01). At the 24th month, the distal bone loss levels of the implants where the loading started between the 60th and 90th days were significantly higher than the bone loss levels of implants loaded earlier than 60 days (P  =  .001; P < .01), and between the 91st and 140th days (P  =  .025; P < .05). No significant difference was reported between the other loading timing groups in terms of distal bone loss rate at the 24th month (P > .05).

In terms of loading timing, a statistically significant difference was seen between the 36th month distal bone loss averages (P < .01). At the 36th month, the distal bone loss levels of the implants where the loading started after 140 days were significantly higher than the bone loss levels of implants loaded earlier than 60 days (P  =  .001; P < .01), and between the 91st and 140th days (P  =  .001; P < .01). At the 36th month, the distal bone loss levels of the implants where the loading started between the 60th and 90th days were significantly higher than the bone loss levels of implants loaded earlier than 60 days (P  =  .001; P < .01), and between the 91st and 140th days (P  =  .033; P < .05). No significant difference between the other loading timing groups was seen in terms of the distal bone loss rate at the 36th month (P > .05).

In terms of loading time, a statistically significant difference was observed between the sixth month mesial bone loss averages (P < .01). At the sixth month, mesial bone loss levels of the implants where loading started after 140 days were significantly higher than the bone loss levels of implants loaded earlier than 60 days (P  =  .001; P < .01), between the 60th and 90th days (P  =  .029; P < .05), and between the 91st and 140th days (P  =  .001; P < .01). No significant difference was noted between the other loading timing groups in terms of the mesial bone loss rate at the sixth month (P > .05).

In terms of loading timing, a statistically significant difference was observed between the 12th month mesial bone loss averages (P < .01). At the 12th month, mesial bone loss levels of the implants where the loading started after 140 days were significantly higher than the bone loss level of implants loaded earlier than 60 days (P  =  .001; P < .01), and between the 91st and 140th days (P  =  .001; P < .01). At the 12th month, the mesial bone loss levels of the implants where implants have been loaded after 60 and before 90 days were significantly higher than the bone loss levels of implants loaded earlier than 60 days (P  =  .003; P < .01). No significant difference was reported between the other loading timing groups in terms of distal bone loss rate at the 12th month (P > .05).

In terms of loading timing, a statistically significant difference was observed between the 24th month mesial bone loss averages (P < .01). At the 24th month, mesial bone loss levels of the implants where the loading started after 140 days were significantly higher than the bone loss levels of implants loaded earlier than 60 days (P  =  .001; P < .01), and between the 91st and 140th days (P  =  .001; P < .01). At the 24th month, the mesial bone loss levels of implants loaded between the 60th and 90th days were significantly higher than the bone loss levels of implants loaded earlier than 60 days (P  =  .001; P < .01), and between the 91st and 140th days (P  =  .014; P < .05). No significant difference was noted between the other loading timing groups in terms of mesial bone loss rate at the 24th month (P > .05).

In terms of loading timing, a statistically significant difference was reported between the 36th month mesial bone loss averages (P < .01). At the 36th month, the mesial bone loss levels of implants loaded after 140 days were significantly higher than the bone loss levels of implants loaded earlier than 60 days (P  =  .001; P < .01), and between the 91st and 140th days (P  =  .001; P < .01). At the 36th month, the mesial bone loss levels of implants loaded between the 60th and 90th days were significantly higher than the bone loss levels of implants loaded earlier than 60 days (P  =  .001; P < .01). No significant difference was noted between the other loading timing groups in terms of mesial bone loss rate at the 36th month (P > .05).

In terms of loading timing, no statistically significant difference was observed between 6th, 12th, 24th, and 36th month distal and mesial bone loss averages (P > .05) (Table 2).

In the fixed restored group, 106 implants were splinted, whereas 20 have been single. Statistical analysis showed no statistically significant association between the 2 groups in terms of loading timing and later MBL (Table 3).

Evaluation of the implants in terms of the neighboring teeth situation showed that a neighboring tooth significantly decreased later MBL on the related side of the implant (P < .01) (Table 4).

Evaluation of smooth collar implants (Straumann; Zimmer Dental) and microgeometry neck implants (Astratech; Biolok-Biohorizons) showed no statistically significant association between loading timing and MBL (Tables 5 through 7).

The aim of this study was to document the influence of several variables on the MBL around implants supporting mandibular overdentures. Since it was proposed that bone-anchored prostheses could be sustained in the oral environment for a lifetime,2 it is important to know the effects of different factors on bone resorption. It was emphasized that it is worth noting that even the proposed rate of bone loss of less than 0.2 mm/y may be too liberal for young implant patients who could lose up to 8 mm of bone over the ensuing 40 years.12 On the basis of clinical observations, bone loss ranging between 1 and 2.6 mm has been reported to occur around the margin of successfully osseointegrated dental implants.13–15 In spite of lack of consensus, the values generally accepted as a reasonable guideline for bone loss since the late 80s are less than 1.5 mm for the first year post loading of the implants, with less than 0.2 mm of additional loss for each following year.16,17 

Although it was stated that immediately and early loaded implants may be at greater risk of failure than conventionally loaded ones,18 several randomized controlled trial studies and review articles have shown no important differences in terms of prosthesis failure, implant failure, and marginal bone level changes between the different loading strategies6,19—immediate vs conventional loading,20–29 early vs conventional loading,30–32 and immediate vs early loading.33–37 Based on a review of the current literature, it can be concluded that immediate or early placement of implants may be a viable alternative to delayed placement.38 Although a tendency toward higher losses where implants were also immediately loaded was pointed out in a review of the related literature,39 the results of several clinical studies have indicated that immediate loading of implants placed in immediate extraction sites can be carried out successfully.40–42 

As is well known, bone responds to local mechanical forces with changes in internal architecture and external shape. Cortical and trabecular bone is modified by modeling and remodeling. The cellular behavior of bone cells is determined largely by the mechanical environment of strain or deformation of the bone cells.43 As was described by Frost,44 disuse of bone increases remodeling and results in decreased bone mass, similar to an overload situation. The remodeling rate is the period of time needed for new bone to replace existing bone and allows for the adaptation of bone to its environment, for example, next to an implant. Lamellar bone forms at a rate of 1 to 5 µm each day, whereas woven bone can form at rates of more than 60 µm each day. Hence a higher bone turnover rate is taking place for woven bone. Three types of bone are found next to an osseointegrated implant: lamellar, woven, and composite bone. Lamellar bone is the best organized, most highly mineralized, and strongest of all bone types. It is the load-bearing bone that is most desired next to an implant.45 Woven bone is also called immature bone because it is unorganized and less mineralized and has less strength than other types. Composite bone is a combination of lamellar and woven bone. After an implant is inserted, remodeling in the implant-bone interface allows formation of a viable bone. By the end of 4 months of a maturation phase bone next to a healing implant interface, osteoblasts deposit about 70% of the mineral found in mature vital bone. The remaining 30% of mineral deposition occurs during secondary mineralization over the next 8-month period. Once the bone has healed and the implant is loaded, the interface remodels again, as influenced by its local strain environment, and this remodeling process is continuous.46 Crestal bone loss is reported to be the result not only of overload situations, but also of disuse atrophy as a consequence of smooth collar designs, causing reduced strain transmission to neighboring bone.47–50 This point is important in explaining the results of this present study. However, in terms of loading timing, the implants with microgeometry neck design when compared with smooth collar implants did not show less MBL in this study. The microthreads in the implant neck are designed to conduct functional forces to the surrounding bone in a favorable way. Because of missing functional forces during the unloaded period, the bone seems to recede, irrespective of collar design. Delay of loading of implants seems to increase MBL around dental implants caused by disuse atrophy. A possible reason for the indifference in the fixed restoration supporting implant group could be the support of neighboring teeth. Implants placed next to teeth have shown less bone loss at the tooth-facing side than at the free end or implant-facing side (Figures 2 and 3, Table 4). Teeth must be conducting functional forces to the surrounding bone, thus protecting it from disuse atrophy. Once the bone has healed and the implant is loaded, the interface remodels again, as influenced by its local strain environment. A major secondary remodeling action due to late loading could be the reason of higher MBL observed later.

Radiography plays an important role in clinical routine practice and in research projects evaluating dental implants. Measurements of marginal alveolar bone level changes that occur over time in radiographs have been reported to be important parameters.51 Panoramic radiographs are widely used for evaluation of the condition of the bone around implants supporting mandibular OVDs.52–57 Annual bone loss of 0.2–0.32 mm, measured from panoramic radiographs, has been reported. Different methods have been used to assess bone height in the implant region—from just counting the number of threads on screw-type implants, to taking measurements by using a computer-based interactive image analysis system. However, the accuracy of the computerized method was reported to be higher than that of conventional methods, such as the use of a magnifying glass or a sliding gauge by several authors.58,59 Another author pointed to the fact that intraoral direct digital radiographs are not an equivalent substitute for conventional radiographs in evaluating alveolar bone levels.60 Currently, the best object to use when measuring marginal bone levels around implants seems to be the scanned and digitized conventional radiograph,61,62 which was used in the present study, too. The bone loss documented in this study was caused by a reduction in bone levels at the mesial and distal sides of the implants, ignoring the so-called saucerization of the crestal bone around the neck of the implants caused by eventual excessive dynamic loading,63 because only 2-dimensional imaging was used. To be able to gain data especially about vestibular crestal bone changes, a 3-dimensional imaging technique, such as volumetric tomography, is necessary. Nevertheless, a panoramic radiograph includes both of the jaws and the teeth and is a simple examination method.64 Although most panoramic machines provide varied and unreliable magnification, intraoral radiographs have a higher resolution and they are more time consuming to obtain.65 Furthermore, the anatomy of the patient, for example, an extremely resorbed mandible or the inclination of the implant, can make placement of the intraoral radiograph impossible. Alternatively, panoramic exposure offers ease of operation and a shorter working time. Given that the radiographs were of a high quality, Åkesson concluded that for assessment of MBL around the teeth, the radiographic examination of choice would be the panoramic radiograph,66 which is in accordance with findings of a study by Persson et al.67 A recent study by Kullman et al68 also showed that panoramic radiographs were as reliable as conventional intraoral radiographs when used to assess the point of bone attachment to implant threads. Interobserver agreement on MBL assessment from intraoral and panoramic radiographs was studied in several investigations.69,70 It was stated that reliability can be improved with multiple readings by a single observer or, even better, by letting several observers make several, independent readings, which limits the probability that a single observer may be an outlier. In light of these studies, we decided to use panoramic radiographs in routine recall sessions of all patients and to supplement them in cases of insufficient quality with intraoral radiographs. To ensure reliable bone loss measurements, a prosthodontist and a specialist in oral and maxillofacial radiology assessed the bone level on all radiographs.

Within the limitations of this study, our results encourage early loading because delay of loading of implants seems to increase MBL around dental implants caused by disuse atrophy obviously seen in removable denture supporting implants. Implants next to natural teeth seem to be protected by functional forces conducted to the crestal and alveolar bone. So that more reliable conclusions can be drawn, higher numbers of implants should be observed in future studies.

ANOVA

analysis of variance

HSD

honestly significant difference test

MBL

marginal bone loss

OVD

overdenture

PD

probing depth

PI

plaque index

SBI

sulcus bleeding index

The team would like to thank Dr Sinan Horasan from the Teknodent Radiographic Imaging Center, Istanbul, Turkey, for his kind efforts in digitalizing the radiographs and taking the second measurements as a maxillofacial radiologist.