Clinical performance of additively manufactured subperiosteal implants: a systematic review

The use of subperiosteal implants (SI) was originally described by Dahl [1] in 1943, but gained relevance after the publication of Goldberg & Gerskoff [2] at the end of the 1940s. In the 1960s the basis of the so-called osseointegration was partially enlightened [3]. This scientific breakthrough, allowed implant dentistry to evolve from an experimental treatment to the current highly predictable option to replace missing teeth [3]. During this step-by-step transformation, SI have evolved like root-shape implants have drastically made. First SI were manufactured in Vitallium [4], (60% cobalt, 20% chromium, 5% molybdenum, and traces of other substances) [5] and were designed to support complete dentures (mostly removable). New SI designs were described and clinically assessed between the 1970s and the late 1990s. [6] At that time, acceptable 5-year results were documented (95% [7, 8] to 100% [9]) but long-term results regarding survival were less favorable (79% at 10 years [7], 76% at 10 years [8], 75% at 6 years [9], 67% at 10 years [10]). A less convenient for the patient two-time surgical approach was also required then. At the first surgery, a wide flap was raised to allow direct analogical bone surface impressions. During the second one, a casted Cr–Co alloy (or others) framework was adapted and placed beneath the mucoperiosteum without anchoring elements (such as osteosynthesis screws) in most of the cases. Lack of fitting and/or stability, unfavorable biomechanical design, and the use of unsuitable materials to achieve osseointegration, increased the risk of infection, implant exposition, and failure [4, 11]. In case of implant exposition, former SI designs impaired partial or full removal [12]. Probably, this further jeopardized both implant function and esthetics.

SI lost popularity among dental practitioners for a long period, but recent advances in computer-aided design (CAD), computer-aided manufacturing (CAM), the development of new materials (new Ti alloys or Polyether-ether-ketone or PEEK), improvements in surface treatments and a deeper understanding of bone biomechanical principles, have brought SI to a new scenario [13‐18]. Thus, modern CAD designed and additively manufactured SI could provide advantages over former SI such as enabling one-time surgery and immediate loading, better fitting or surgical time reduction [14‐16].

Frequently, clinicians must face the challenge of treating cases of severe bone atrophy or bone resection. Advances in root-shape implants design and size (short, extra-short and narrow implants) have provided new solutions or have enhanced older ones, for the treatment of different types of bone atrophy [19‐21]. Different accessory surgical techniques for recovering the lost bone volume were also developed and improved to treat those patients where root-shape implants could not be placed directly [22‐26]. Among them, guided bone regeneration (GBR), maxillary sinus and nasal floor augmentation, inlay or onlay bone grafting, distraction osteogenesis, nerve lateralization or others have been routinely employed with a varying degree of clinical success [22‐26]. The use of zygomatic implants could be also a reliable option for the treatment of those patients with severe posterior maxillary atrophy [27]. The success rate and the incidence and severity of postoperative complications using this type of implants is dependent of clinician expertise [28].

Modern SI have been claimed to present some advantages to treat certain patients with bone atrophy over the above-mentioned techniques. The elimination of bone donor area morbidity (in the case of autologous bone grafting need), the possibility of ambulatory realization and reduction of surgical time, are among the reported benefits for patients [14]. Modern SI also provide an option of treatment for patients with extreme bone defects due to oncologic disease treatment or trauma [29, 30]. On the other hand, the digital resources (devices and software) required to design and manufacture SI are not accessible to all professionals, and the clinical performance of SI is still not well evidenced. This systematic review attempts to assess the clinical performance of modern additively manufactured SI by analyzing their survival and complications rate data available in the literature.

A systematic review was carried out following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement recommendations [31] to answer the following the PICO questions: “In patients with bone atrophy (P), do additively manufactured subperiosteal implants (I), compared to subperiosteal implants manufactured following traditional approaches (c), present satisfactory implant survival and complication rates(O)?”. The aim of this review was to answer with the best available evidence, this question to help clinicians when planning the treatment of patients with maxillary or mandibular bone atrophy.

CAD designed additively manufactured SI presented satisfactory survival (97.8%) in the short-term (weighted mean follow-up time 21.4 months; mean range 1 to 74 months), but there is a paucity of data on their success rates and medium- or long-term clinical behavior. Available data are coming only from observational studies (cohort studies and case series), including 227 unilateral/bilateral implants (in 227 patients). From them, 70/227 (31%) came from the same retrospective study [43] and 55/227 (24%) came from two multi-center studies using same type of implant and performed by the same international team [36, 39]. Another 35/227 (15.4%) came from 3 single-center studies including patients treated with the same type of implant at the same center [47‐49].

The most frequent complications reported are those related to soft tissues. Hereby, partial exposure of the framework seems to be the most frequent complication, although this seems not to conditionate the survival in the short-term. New designs could allow to remove exposed parts or prosthetic posts in an easier and safer way than in former designs [13, 39, 44]. Although this fact has not yet been specifically evidenced in the literature, this improvement could positively influence the success of modern SI.

From those patients where the use of a provisional prosthesis was stated, 5.2% suffered a fracture of the interim prosthesis. Despite no further information is available to analyze the reasons, to ensure a good passive fitting of the prosthesis, to carefully adjust the occlusion and to reinforce the framework of interim prosthesis seems advisable for these patients as it is for those wearing conventional root-shape implants [51, 52].

Although the location of the implant was not specified in 65 patients [43], there was a noticeably higher number of maxillary than mandibular implants (142:20). Furthermore, 93/142 (65.4%) of maxillary implants had been manufactured following the same two specific design concept, and material [36, 39, 46‐49]. From those stated to have been placed in the mandible (20), 11 supported partial rehabilitations, so the extrapolation of the results of this review to full-arch mandible SI must be very prudently performed.

Bone implant fitting during surgery (in the 4 assessed studies) [39‐42, 44] was satisfactory in mostly all cases. However, it was only assessed in 55 implants and the way of rating this outcome was based on personal feedback and potentially subjective. In those patients where fitting was unsatisfactory, the time of surgery was increased to make the implant fit properly to the bone contour. Dimitroulis et al. [44] noted that longer time (> 3 months) between the CT scan and delivery of the SI (what could cause further bone remodeling) or CT slices greater than 1 mm (which reduced the accuracy and tolerance of the device) could influence misfitting.

The main reason for implantation in the selected studies was bone atrophy. A Cawood–Howell atrophy type V or higher was an inclusion criteria in 5/13 studies [36, 39, 41, 44, 47], including 94/227 (41.5%) SI. In 24 patients (10.6%) it was clarified that a resective/maxillectomy had been previously performed. The studies from Mangano et al. [40] and Nemtoi et al. [42] included patients with a residual bone < 10 mm and regenerative bone surgery unwillingness on the part of the patient. In these two last studies the advantages and disadvantages of SI over the use of extra-short implants (≤ 6.5 mm) without the need for ancillary bone regenerative procedures could be arguable in the absence of more specific information about each specific case. Extra-short root-shape implants have evidenced in recent systematic reviews and meta-analysis, similar clinical performance to standard-length ones in terms of marginal bone loss (MBL), technical complications or implant survival [52‐57]. The possibility of placing an immediate prosthesis, the peculiarities of the type of bone defect in each specific case or the experience degree of the surgeons may have influenced this decision, although the real reasons are unclear. The same can be argued to the study from Mounir et al. [45] where an inclusion criteria was enough bone volume to room standard root-shape implants with at least 3 mm of diameter and 8 mm of length.

Despite SI have a long history, their use is secondary to the use of endo-osseous root-shape implants, both in terms of experience and evidence. Until two decades ago, they were mainly used to support mandibular full-arch removable prostheses [4, 7‐9, 58]. In oldest designs, SI were not directly anchored to the bone with osteosynthesis screws or other systems (to avoid movement), were manufactured by casting, required a two-time surgical procedure (first one to take direct impressions of the bone) and good bone–implant fitting was complex to achieve [4, 6‐11]. Studies with these oldest designs and materials showed poor clinical results in the medium- or long-term [6]. Between the 1980s and the 1990s, success rates at 5-year ranging from 90% [10] to 100% [9] were reported but survival rates decreased at 6 years (75%) [10], 10 years (87%) [59] or 13 years (78%) [59]. Furthermore, most of articles did not include other results (in addition to implant survival) that would allow a reliable assessment of the success rate or the degree of patient satisfaction. Considering studies from 90s onwards, a 10-year survival rate of 79% was reported by Yanase et al. [7] and 76% by Bodine et al. [8] A 6-year evaluation performed by Ferrer et al. [60] revealed a 92.5% success of SI including design innovations and an 84% success for SI with classical designs.

Aforementioned paucity of data does not allow to compare medium- or long-term clinical behavior of modern additively manufactured SI and former ones. In any case, several improvements have been incorporated that could be helpful to improve survival, success and/or patient patient´s satisfaction degree, but this is yet to be evidenced. Among these improvements, a better understanding of the of biomechanics trough finite elements studies has allowed to reduce stress accumulation on bones, implants, abutments, and prosthetic frameworks [61, 62]. Golec [63] anticipated in 1986 the use of CAD/CAM to eliminate the need for surgical bone impression. Since then, several improvements in CBCT definition and additive manufacture refinement were needed to obtain more precise frameworks (reducing misfitting and/or micromovements) [34, 64, 65]. Surface features are also involved in the optimization of SI–bone surface interactions. A higher number of the implants included in this revision were porous (rough) on the bony face to promote osteointegration, and smooth (polish) on the soft tissue face to prevent biofilm colonization [13, 34, 66]. Modern manufacturing and new materials resistance allowed to reduce the thickness of the framework up to 0.7- or 0.8-mm [41, 42]. Further than weight lightening, this reduction seems helpful to prevent exposition. On the other hand, small connections also may lead to more fractures, although the limits of thinning are yet to be studied in more depth.

In relation to SI design too, 4–6 prosthetic posts (implant-prosthesis connectors) were preferred in most studies. All these improvements could have contributed to maintain bone and soft tissue stability. In this sense, Van den Borre et al. [37] performed a radiographic evaluation of modern SI and observed acceptable bone remodeling in the underlying bone (mean negative bone remodeling over six reference points on the crest: 0.26 mm ± SD 0.65 mm; mean bone remodeling at the supporting bone at the wings and basal frame: 0.088 mm ± SD 0.29 mm).

No differences in clinical performance between cemented and screw-retained fixed prostheses could be demonstrated. This is not to say that the choice of one retention system or the other lacked clinical significance. From a technical point of view, screw-retained prostheses offer a critical advantage in terms of retrievability. In patients at risk or with a history of previous malignance, screw-retained prosthesis facilitates the mandatory periodical check-up of the tissues underneath fixed rehabilitations [67, 68]. The same rationale can be applied to patients with soft tissue complications, in which screw-retention allows prosthetic removal and recontouring. [42]

Limited information on the performance of SI additively manufactured with other materials different from Ti alloys (PEEK or other materials) is available on the literature. As these materials could be considered very experimental, the group of 5 PEEK SI of the study of Mounir et al. [45] was excluded in the present review. Apart from the data of this study, a case series of 4 edentulous patients was published by Elsawy et al. [35] reporting survival of all the maxillary SI and no complications after a 12-month follow-up period. All the PEEK SI in their study had been manufactured with a 5-axis milling machine, therefore the study did not match the eligibility criteria of the present systematic review.

In summary, modern additively manufactured SI present good survival in the short-time but they still present a notable number of soft tissue complications. Comparing to traditional casted SI, soft-tissue complications could be probably more easily solvable (or containable in extension) as new CAD designs enables a simpler implant trimming and partial removing of the implant. This could reduce the influence of soft-tissue complications on implant survival. Nevertheless, the medium- or long-time clinical behavior is still to be clarified. They present several advantages over traditional casted ones. Better biocompatibility, one-time surgery possibility, a reduction in total mass of the material used, optimization of arms and fixation screws dimensions and number (thus reducing costs and avoiding micromovements) or time of surgery reduction (ensuring a better fitting and avoiding time to re-adapt bone), can be cited among the improvements [13‐15]. New finite element method analysis on additively manufactured SI are desirable, to further enhance this advantages and also could be helpful to prevent overextending the implant.

In cases of extreme resorption, SI may be a feasible treatment option in the hands of experienced clinicians. However, in cases where residual bone available allows to room short root-shape implants or standard ones (even with the need for ancillary surgical procedures) the use of SI could be arguable as root-shape implants performance is further evidenced in the literature. Zygomatic implants are another alternative when the maxillary bone is completely or partially absent if the anatomy of the defect, the remnant bone and the maxillary sinus is favorable. However, zygomatic implants are also considered a complex treatment with significant surgical risk and potential for complications and the success of the treatment is highly dependent on the clinician experience. [69]

In sight of the results of the present study, the use of SI should be based on case selection such as severe atrophy and the impossibility (or unwillingness on the part of the patient) to conduct microvascular bone reconstruction or even patients that have a reduced expected lifespan.

A significant limitation of the present review is the absence of RCTs, prospective studies or other studies with a higher level of evidence in the available literature. A meta-analysis was not possible to obtain, due to this and the great heterogeneity between studies. On the other hand, no previous systematic review has been conducted on the topic to the best knowledge of the authors, and the results obtained could encourage to perform new well-designed studies to clarify the important lack of information in some key points for clinical practice. In sight of the results of the present systematic review, some treatment recommendations for former SI in older studies could be partially outdated.

Subperiosteal implants have been used for decades, but lost relevance among clinicians due former poor clinical performance. Improvements through new technologies development have brought them to a new scenario. Based on the available studies (observational), “modern” CAD designed, and additively manufactured SI presented a satisfactory survival in the short time. However, further studies are needed to ascertain the success rate and the clinical behavior in the medium- and long-term. It would also be desirable to conduct further studies on CAD designed SI manufactured with the most modern subtractive manufacturing methods in view of the limited available clinical information.

Partial exposure was the most common complication reported. Post-operative complications, soft-tissue infection and interim prosthesis fracture were other remarkable complications reported. New SI designs may be helpful to prevent complications, but there is a need to strengthen the evidence with new clinical studies.