Clinical Performance of Short Expandable Dental Implants for Oral Rehabilitation in Highly Atrophic Alveolar Bone: 3-year Results of a Prospective Single-Center Cohort Study

Background and Objectives: Oral health-related quality of life (OHRQOL) is compromised during the post-implant healing period, especially when vertical augmentation is required. A long-term trial sought to evaluate a short dental implant system with an apically expandable macro-design. Materials and Methods: Over 4.5 years, patients with limited vertical alveolar bone were consecutively recruited into this prospective cohort study. Implant success rate, OHRQOL (Oral Health Impact Profile (OHIP)-14), implant stability, and crestal bone changes were evaluated. Results: Data from 30 patients (mean age: 64.6 years, range 44–83) were analyzed, which related to 104 implants (53 in the maxilla, 51 in the mandible). Over the mean follow-up (42.6 ± 16.4 months), the implant success rate was 94.7% in the mandible (two implants lost) and 83.6% in the maxilla (four implants lost; p = 0.096), and the prosthetic success rate was 100%. The median OHIP-14 scores improved from 23 (interquartile range (IQR) 9–25.5) to 2 (IQR 0–5; p < 0.001). The mean implant stability quotient (ISQ) was 71.2 ± 10.6 for primary stability and 73.7 ± 13.3 (p = 0.213) for secondary stability, without significant maxilla-versus-mandible differences (p ≥ 0.066). Compared to the baseline, median crestal bone changes after loading were 1.0 mm (IQR 0–1.3) and 1.0 mm (IQR 0.2–1.2) in the maxilla and mandible (p = 0.508), respectively, at the end of the first year, 1.1 mm (IQR 0–1.3) and 1.0 mm (IQR 0.1–1.2) (p = 0.382), respectively, at the end of the second year, and 1.2 mm (IQR 0–1.9) and 1.1 mm (IQR 0.1–1.2) (p = 0.304), respectively, at the end of the third year. Conclusions: In patients with limited vertical bone height, short implants with optimized macro-design constitute a reliable method for functional rehabilitation, avoiding extensive alveolar bone augmentation.


Introduction
The treatment process is extensive for patients with limited vertical alveolar bone height because augmentation procedures are required prior to implanting standard dental implants [1]. However, the remodeling of augmented alveolar ridges occurs with volume loss, due to the remodeling or uncertain predictability of various developed procedures. Depending on general and local factors, up to 25% of the primary volume is resorbed [2]. For several years, short dental implants (<8 mm) The measures employed were implant success and survival rate, as well as OHRQOL. The implant success rate was calculated using known success criteria (functional implant, no sign of infection or pain, no mobility, no radiolucent area around the implant) [25,26], while implant survival was computed using the Kaplan-Meier method. For implant survival analysis, only patients with a known observation time were included. Implant survival was defined as the "implant being osseointegrated and prosthetically in function" at the time of the latest follow-up examination. Oral Health Impact Profile 14 (OHIP-G14) questionnaires were evaluated at baseline (prior to implantation) and 6 months after prosthetic rehabilitation. Each OHIP item elicited information about how frequently subjects had experienced a specific impact in the previous month. OHIP-G14 is a self-administered questionnaire that follows a standard ordinal format ("never" = 0, "hardly ever" = 1, "occasionally" = 2, "often" = 3, and "very often" = 4). Seven dimensions of preoperative and postoperative OHIP-14 scores, as well as cumulative scores, were determined for each patient [27,28]. Oral health-related quality of life (OHRQOL) was re-assessed 6 months after prosthetic rehabilitation.
Measurement of primary stability (resonance frequency analysis), primary wound closure 6.
Postoperative digital periapical radiogram Crestal bone changes were evaluated yearly using digital periapical radiographs with the rectangular technique (see Section 2.6).

Implants
This study employed a CE-certified short expandable titanium dental implant (PYRAMIDION dental implant, DenTack Implants Ltd., Kfar-Saba, Israel). The implant was submitted to all mechanical fatigue tests in vitro (ISO 14801:2016 standard and FDA guidance for root-form endosseous dental implants and endosseous dental implant abutments). Apical expansion was performed after implant insertion using a special expansion tool and ratchet torque, which resulted in a pyramid shape [24]. The implants had the following dimensions and special characteristics: 5 mm, 6 mm, and 7 mm in length; 3.75 mm and 4.1 mm in diameter; internal (7 mm length) or external (5 mm and 6 mm length) hexagon platform.
Sample size calculation on the implant level was performed using the open-source statistical program G*Power 3.1, based on the t-test for differences between two independent means and considering the stability values in both jaws of the pilot study [31]. The total calculated sample size was 112 and the statistical power was 0.90.

Surgical and Prosthetic Protocol
Implantological treatment planning followed anamnesis, clinical, and orthopantomographic examination. Chair-side examination assessed cranial nerves, skin in the head and neck region, temporomandibular joints, masticatory muscles, and cervical lymph nodes. Intraoral examination included assessment of dental, periodontal, and mucosal pathologies, as well as edentulous alveolar ridges (Cawood and Howell category), occlusion, and existing restorations. Concerning the position and number of implants, the recommended categories from the German consensus conference were employed [32].
Prior to being enrolled in this study, all patients signed a written informed consent form. Surgical treatment was performed under local anesthesia. A mid-crestal incision was made and the mucoperiosteal flap was elevated (after a single median buccal release incision for edentulous jaws). Based on preoperative orthopantomography and intra-operative tactile sensation, the bone quality at the implant sites was recorded by the first author, using the classification from Lekholm and Zarb (D1, D2, D3, and D4) [33]. Table 2 summarizes the drilling sequence, manual implant insertion, and expansion. All surgeries, except for one patient (supervision by the first author), and all follow-up examinations were performed by the same surgeon (W.R.), so as to reduce performance and inter-observer variations/bias. Implants were inserted 0.5 mm sub-crestally using a hand ratchet. Thereafter, the apical expansion was performed according to the manufacturer's recommendations and using the appropriate expansion tool and hand ratchet. After a cover screw was positioned, a periosteal incision was made and mucosal wounds were closed using absorbable sutures (Monocryl 5-0, ETHICON, Johnson and Johnson, New Brunswick, NJ, USA). In some cases, lateral augmentation procedures were included, which were combined with antibiotic therapy consisting of amoxycillin 1 g every 8 h for 7 days. All patients were postoperatively instructed to use an oral antiseptic agent for 7 days (0.2% chlorhexidine) and to temporarily take the non-steroidal anti-inflammatory drug ibuprofen (600 mg).
Participants were instructed not to wear their dentures for 1 week following surgery and to avoid brushing at the surgical site. Patients were followed up after 7 days. The conventional dentures were subsequently relined with a soft material (Visco gel, Densply, Salzburg, Austria). The following conventional periods of submerged healing were chosen for this study: 3 months in the mandible and 6 months in the maxilla. During re-entry surgery, a minimum of 2 mm keratinized peri-implant soft tissue mucosa was considered, where needed, using curvilinear crestal or palatal/lingual para-crestal oblique incisions [34].
All prosthetic treatments were provided by two experienced specialists (R.S. and J.H.) in the Department of Prosthetic Dentistry of a university school of dental medicine. Prosthetic treatment was initiated after a minimum of 2 weeks following surgical re-entry. This required three sessions for fixed dentures, four sessions for removable dentures with ball attachments, six sessions for combined fixed-removable dentures, and seven sessions for removable dentures with jaw bars [24]. The abutment screws were fixed with a torque of 15Ncm, according to the manufacturer's recommendation. Wherever possible, adjacent implants were primarily splinted (crowns, bar) and extra-axial loading during dynamic occlusion was avoided. However, in some cases, eccentric group guidance was achieved. In order to reduce overloading in the peri-implant bone and implant-abutment connection, the occlusal surface was designed to be smaller [13,14,18,35,36]. The first follow-up was scheduled for a maximum of 4 weeks later. Further follow-ups were scheduled quarterly. Patients that had attended regular check-ups for correct maintenance of prostheses were evaluated to determine their need for relining and changes to retention inserts.

Clinical and Radiological Follow-up
The first follow-up investigation was arranged for a maximum of 4 weeks after prosthetic rehabilitation. Further aftercare was arranged quarterly in the first year, and every 6 months thereafter. Patients were screened clinically and radiologically (yearly) for biological and technical complications. Marginal bone changes were evaluated yearly using digital periapical radiograms, ideally with the rectangular technique (Sidexis imaging software, Sirona, Bensheim, Germany). The distance between each implant shoulder and first bone-implant contact at the mesial and distal aspect was measured by the first author (W.R.), and the mean values per implant were calculated [37] 1, 2, and 3 years after loading. The known implant length served as a reference to verify the precise measurement of crestal bone changes. The authors applied Buser's implant success criteria [25], which incorporate the absence of persistent subjective complaints, recurrent peri-implant infection, mobility, and continuous radiolucency, as well as the possibility of prosthetic loading. Crestal bone changes were classified according to Linkevicius (2019) as: "zero bone loss" = 0 mm, "stable bone remodeling" ≤ 1.2 mm, and "progressive bone loss" > 1.2 mm [38].

Data Gathering and Statistics
All patients were pseudonymized. Parameters were added to a databank and analyzed statistically using statistics software (IBM SPSS statistics, Version 20, Chicago, IL, USA). Descriptive statistics presented the distribution of several occurrences and frequencies, as well as combinations of certain features. The distribution of continuous data was tested by the Shapiro-Wilk test. Analytical statistical tests were performed depending on the scale, with a chi-squared test for categorical parameters, paired and independent t-tests for differences in mean values (normally distributed data), and a Wilcoxon signed rank test (for paired samples) or Wilcoxon rank sum test (for independent samples) for differences in median values (non-normally distributed data). The Bonferroni correction was used to counteract the problem of multiple comparisons. Implant survival was analyzed by Kaplan-Meier analysis and log-rank test. The level of significance was set at 5%.

Study Population
Over a 43-month period (July 2014-January 2018), 34 patients (female n = 21, male n = 13) with an average age of 64.6 years (range 28-84 years) were enrolled in this study. The preoperative physical status classification, according to the American Society of Anesthesiologists, yielded: ASA Score 1 (healthy patient) in 15 patients, ASA Score 2 (patient with mild systemic disease: arterial hypertension, chronic bronchitis, diabetes mellitus, gastric ulcer, cured hepatitis C virus infection, hypothyroidism, osteoarthrosis, venous thrombosis) in 10 patients, and ASA Score 3 (patient with severe systemic disease: congestive heart failure, hemiplegia, history of a cured early stage oral squamous-cell carcinoma) in 5 patients. The anamnesis revealed that six patients were smokers. With regards to oral diseases, eight patients had a history of marginal periodontitis and two patients displayed chronic mucositis. Clinical examination demonstrated the following categories of edentulous alveolar process atrophy, according to Cawood and Howell: category III n = 29 and category IV n = 75. Indication categories related to the recommended implantological treatment [34] were distributed as follows: complete edentulous maxilla (IIIa) n = 5, complete edentulous mandible (IIIb) n = 9, partially edentulous areas confined to several teeth n = 2 (IIa) and n = 14 (IIb), and a partially edentulous area confined to a single tooth (Ib) n = 1. Figure 1 presents the flowchart for this study. In these patients, a total of 122 implants were inserted (maxilla n = 57, mandible n = 65). Over the total follow-up period, three patients dropped out due to non-compliance and one dropped out due to malignancy (total n = 13 implants).  [34] were distributed as follows: complete edentulous maxilla (IIIa) n = 5, complete edentulous mandible (IIIb) n = 9, partially edentulous areas confined to several teeth n = 2 (IIa) and n = 14 (IIb), and a partially edentulous area confined to a single tooth (Ib) n = 1. Figure 1 presents the flowchart for this study. In these patients, a total of 122 implants were inserted (maxilla n = 57, mandible n = 65). Over the total follow-up period, three patients dropped out due to non-compliance and one dropped out due to malignancy (total n = 13 implants). Based on the radiological findings (preoperative orthopantomography) for 30 patients, the bone quality at implant sites was as follows: D1 in n = 18 cases (maxilla n = 2, mandible n = 16), D2 in n = 24 (maxilla n = 4, mandible n = 20), D3 in n = 41 (maxilla n = 26, mandible n = 15), and D4 in n = 21 (maxilla n = 21). The implant dimensions that were used were as follows: implant lengths of 5 mm in n = 7, 6 mm in n = 2, and 7 mm in n = 95; implant diameters of 3.75 mm in n = 29 and 4.1 mm in n = 75. Implant positions, prosthetic treatments, and success rates are summarized in Table 3. This table also contains nine patients (indicated by "#") whose short-term results have previously been published [24]. In terms of validating the initial outcome, these individuals are included in the present analysis because of long-term follow-up. Four patients that dropped out are not included in Table 3. Based on the radiological findings (preoperative orthopantomography) for 30 patients, the bone quality at implant sites was as follows: D1 in n = 18 cases (maxilla n = 2, mandible n = 16), D2 in n = 24 (maxilla n = 4, mandible n = 20), D3 in n = 41 (maxilla n = 26, mandible n = 15), and D4 in n = 21 (maxilla n = 21). The implant dimensions that were used were as follows: implant lengths of 5 mm in n = 7, 6 mm in n = 2, and 7 mm in n = 95; implant diameters of 3.75 mm in n = 29 and 4.1 mm in n = 75. Implant positions, prosthetic treatments, and success rates are summarized in Table 3. This table also contains nine patients (indicated by "#") whose short-term results have previously been published [24]. In terms of validating the initial outcome, these individuals are included in the present analysis because of long-term follow-up. Four patients that dropped out are not included in Table 3.   Additional minor augmentation procedures were necessary in 12 patients at 23 implant sites. These included lateral augmentation with bone grafting (n = 4), bone spreading (n = 12), and internal sinus lift using an osteotome for condensing preparation after underdrilling and particulate bone grafting (n = 7). Patients were rehabilitated with fixed dentures in 11 cases and with removable dentures in 19.
In relation to bone quality at implant sites (Lekholm and Zarb), the difference in cumulative implant survival was not statistically significant: D1 bone 83%, D2 bone 100%, D3 bone 83%, and D4 bone 86% (log-rank test, p = 0.275). Otherwise, the status of the alveolar process atrophy (Cawood and Howell) significantly influenced implant survival rate: category III 100% and category IV 83% (log-rank test, p = 0.011). This also significantly influenced prosthetic rehabilitation: removable 93% and fixed 100% (log-rank test, p < 0.001). In three patients, five implants were lost before loading, and four implants were lost under loading for a further three patients. The affected patients displayed compromised bone quality. This included tumor patients (squamous-cell carcinoma on the floor of the mouth), patients with highly atrophic alveolar bone (Cawood and Howell IV) [20], and patients with D3-D4 trabecular bone structures (according to Lekholm and Zarb) [33] (Table 3). In these patients, the prosthetic restauration was successfully modified. In one case (Patient 12), an abutment screw loosening was diagnosed and successfully refastened. No other technical complications were observed in terms of implant fracture, abutment screw fracture, or retention system fracture.

Statistics (Wilcoxon Signed
Rank Test for Paired Samples)

Functional limitation
Have you had trouble pronouncing any words because of problems with your teeth, mouth, or dentures?
Have you felt that your sense of taste has worsened because of . . .

Biomechanical Implant Stability
Data were distributed normally for primary stability in the maxilla and mandible (Shapiro-Wilk test, p ≥ 0.132) and for secondary stability in the maxilla and mandible (p ≥ 0.145). RFA returned a mean ISQ of 71.2 ± 10.6 ISQ units for primary stability (Figure 3a,b) and 73.7 ± 13.3 ISQ units for secondary stability (Figure 4a,b) (t-test for paired samples, p = 0.213). There were no significant differences between the maxilla and mandible (t-test for independent samples, p ≥ 0.066).

Biomechanical Implant Stability
Data were distributed normally for primary stability in the maxilla and mandible (Shapiro-Wilk test, p ≥ 0.132) and for secondary stability in the maxilla and mandible (p ≥ 0.145). RFA returned a mean ISQ of 71.2 ± 10.6 ISQ units for primary stability (Figure 3a,b) and 73.7 ± 13.3 ISQ units for secondary stability (Figure 4a,b) (t-test for paired samples, p = 0.213). There were no significant differences between the maxilla and mandible (t-test for independent samples, p ≥ 0.066).
A representative "zero bone loss" case in a rehabilitated female patient is depicted in Figure 6a A representative "zero bone loss" case in a rehabilitated female patient is depicted in Figure 6a

Discussion
Our study sought to evaluate a new expandable dental implant system under difficult, local bony conditions. Only a few reports in the literature have addressed expandable dental implants, and no other studies report on a comparable expandable short implant. The advantages and potential limitations of short and expandable implants are discussed in our earlier paper [24].
Recent literature has shown that short implants are increasingly accepted in the field of oral implantology [6,7,39,40]. The survival rate for short dental implants has increased from 80% to >90% over time [39]. This was also confirmed in two recent studies. For short dental implants that support single crowns and fixed bridges, especially in the mandible, a 2-year success rate of 97% [41] and 5year outcome of 92.2% [7] were reported. Results from a 3-year randomized controlled trial indicated that short implants achieve similar outcomes, compared to longer implants that are inserted into augmented bone, and might be preferable compared to bone augmentation, especially in the posterior mandible [42]. According to another prospective 10-year cohort study, follow-up of single  Table 3). (a) Standard periapical radiogram implants i16 and i15; (b) Standard periapical radiogram implants i25 and i26.
A representative "zero bone loss" case in a rehabilitated female patient is depicted in Figure 6a

Discussion
Our study sought to evaluate a new expandable dental implant system under difficult, local bony conditions. Only a few reports in the literature have addressed expandable dental implants, and no other studies report on a comparable expandable short implant. The advantages and potential limitations of short and expandable implants are discussed in our earlier paper [24].
Recent literature has shown that short implants are increasingly accepted in the field of oral implantology [6,7,39,40]. The survival rate for short dental implants has increased from 80% to >90% over time [39]. This was also confirmed in two recent studies. For short dental implants that support single crowns and fixed bridges, especially in the mandible, a 2-year success rate of 97% [41] and 5year outcome of 92.2% [7] were reported. Results from a 3-year randomized controlled trial indicated that short implants achieve similar outcomes, compared to longer implants that are inserted into augmented bone, and might be preferable compared to bone augmentation, especially in the posterior mandible [42]. According to another prospective 10-year cohort study, follow-up of single

Discussion
Our study sought to evaluate a new expandable dental implant system under difficult, local bony conditions. Only a few reports in the literature have addressed expandable dental implants, and no other studies report on a comparable expandable short implant. The advantages and potential limitations of short and expandable implants are discussed in our earlier paper [24].
Recent literature has shown that short implants are increasingly accepted in the field of oral implantology [6,7,39,40]. The survival rate for short dental implants has increased from 80% to >90% over time [39]. This was also confirmed in two recent studies. For short dental implants that support single crowns and fixed bridges, especially in the mandible, a 2-year success rate of 97% [41] and 5-year outcome of 92.2% [7] were reported. Results from a 3-year randomized controlled trial indicated that short implants achieve similar outcomes, compared to longer implants that are inserted into augmented bone, and might be preferable compared to bone augmentation, especially in the posterior mandible [42]. According to another prospective 10-year cohort study, follow-up of single crowns supported by short implants in posterior regions and loaded after 6-7 weeks maintained long-term full function in 91.7% of cases, with low marginal bone loss (0.8 ± 0.7 mm) [43].
We achieved an overall implant success rate of 94.7% in the mandible and 83.6% in the maxilla for our elderly heterogeneous cohort (medical condition, indication category, implant site distribution, prosthodontic treatment), which is not fully comparable with results in the recent literature. The heterogeneity of the study cohort is therefore considered a weakness of the study design. We addressed this by presenting stratified results according to comorbidities and local conditions (anatomical region, bone quality, minor augmentations, prosthetic rehabilitation). There is a lack of directly comparable data from earlier clinical studies using this short implant design. In our study, very few implants were lost before loading and under loading (four vs. four, Table 3). The six patients that were affected presented compromised bone quality. This study returned a lower overall survival rate for implants in the maxilla than studies conducted by Slotte et al. (2015) [7] and Malmstrom et al. (2016) [41], which can be attributed to less optimal patients being included in our study. In a previous systematic review, 11 studies reported more implant failures before loading, while seven studies stated more implant failures under loading [39]. Furthermore, it is evident that there is a lower risk of complications when using short implants, compared to standard implants, in situations involving vertical bone loss that requires augmentation [2,4,5] or nerve lateralization [40].
Long-term success in implantology depends on a sufficient quantity and quality of peri-implant bone, as well as healthy mucosa. A wide range of modalities can be used to fully describe bone health, including biomarkers in serum and urine, imaging techniques, biomechanical testing, and histomorphometry [44]. Furthermore, studies on dental implant treatment present ambiguous bone tissue characteristics. In addition to the widely used Lekholm and Zarb classification, other authors describe bone quality using the Misch, Trisi, and Rao classification systems [45]. Structural jawbone quality relates to the amount of cortical and trabecular bone, while bone density relates to the amount of mineralization. The most precise methods of evaluating trabecular microstructure and bone density are multislice computed tomography (CT), cone beam CT, and micro-CT [46]. Triches et al. analyzed the relationship between insertion torque and the tactile, visual, and gray-value measures of bone quality with short implants, concluding that assessment methods are consistently related [47].

Oral Health-Related Quality of Life
In a representative normal population, OHIP scores typically tend to increase with age, along with a decreasing number of natural teeth, with scores ranging between 10 and 34 points [48][49][50]. According to a national population-based study, reference OHIP-14 values among 90% of subjects without dentures yielded ≤11 points, ≤17 points for subjects with a removable partial denture, and ≤25 points for subjects with a complete denture [27]. Based on the data collected in a national survey, John et al. (2004) concluded that specific OHIP scores can also provide a frame of reference for specific oral conditions, when OHRQOL is measured. Therefore, we believe that such evidence can be used as a model of comparison for our results. In the present study, the initial median OHIP score was 23 (IQR 9-25.5), which corresponds to subjects with removable dentures. The median decrease in OHIP cumulative scores was about 21 points to 2 (IQR 0-5), which corresponds to subjects without dentures. Our findings indicate a relevant increase in OHRQOL, suggesting successful patient-based oral rehabilitation, which aligns with results reported by John et al. [27] and Reißmann et al. [49].

Implant Stability under Difficult Conditions
Several investigators have analyzed the preferred indications of short dental implants in posterior mandibles and maxillae, and they have also outlined its cost efficiency compared to additional vertical augmentations. Our study applied a short implant in both jaws, with almost all possible indication categories represented, which confirms the broad versatility of the implant (Table 3).
Earlier biomechanical finite element studies have confirmed that apical expansion results in a favorable stress reduction of almost 10% in the crestal bone [55]. We can assume that, in addition to the microthread and platform-switching concept [56], peri-implant crestal bone strain could be reduced by apical expansion.
With regards to RFA, the values obtained were related to bone quality and quantity, as well as the exposed implant height above the alveolar crest, which depends on the type of implant and insertion technique [57]. Our results (mean primary stability of 66.1 ISQ units in the maxilla and 75.9 ISQ units in the mandible; mean secondary stability of 68.2 ISQ units in the maxilla and 80.1 ISQ units in the mandible) were comparable to those obtained with standard-length implants. Becker et al. actually collected similar data (standard-length implants): primary stability of 72.1 ISQ units and secondary stability of 72.6 ISQ units [58]. These values are marginally lower than those of short implants that are only inserted into the posterior mandible (79.0 ISQ units) [9]. Other authors measured 68.2 ISQ units in the posterior maxilla (6 mm implants) [42]. Overall, the special design of the current study yielded reliably high stability values, compared to standard implant dimensions [30].
Although lateral augmentations are only performed in a few patients, they cannot be avoided in cases of a narrow alveolar ridge. Nevertheless, benefit-risk evaluation reveals that patients benefit from employed procedures without undergoing extensive additional vertical augmentations. In the present study, the apical implant design was shown to influence implant stability and bone-to-implant contact, which reinforces the findings of Romanos et al. [16] and Gehrke et al. [59]. The expansion procedure presents an additional bicortical anchorage in the oro-vestibular direction (pyramid shape of the alveolar process), which is understood to optimize load resistance [60]. In hard bone, manufacturers' recommendations should be strongly considered. We found only one study reporting trans-crestal sinus lifting (average residual bone height of 4.7, using platelet concentrates) in association with 41 short implants [61], which is comparable to the case presented in Figures 6a and 7b. The authors concluded that a stable augmented height gain of 4.2 mm was found after a 3-year follow-up.
Examples of alternative methods that aim to achieve a high level of biomechanical stability are osseodensification [62] and special thread designs [16,63]. For the posterior mandible, a comparative study of immediate loading (single crowns) on short implants and conventional dental implants reported comparable 1-year results for implant survival, marginal bone level, and ISQ values [64].

Peri-Implant Crestal Bone Loss
In the current study, crestal bone changes under loading in the first year were lower than in the second year, which did not differ from those obtained in the third year. Moreover, there were marginal differences between the maxilla and mandible in the first year, which only partially agrees with previous observations that have been reported in the literature [4,56].
Microbiological conditions influence the maintenance of peri-implant bone and play a critical role already during the osseointegration period. Clinical and microbiological analyses demonstrated an increased severity of marginal bone loss around non-submerged implants in relation to the salivary microbiome, particularly for participants with increased proportions of periodontal pathogenic species [65]. In a recent review of the metagenomics and culturomics of a peri-implant microbiome, Martellacci et al. (2019) asserted that teeth and implants do not appear to share the same microbiome, which is more diverse in healthy teeth [66]. Peri-implant and subgingival microbiota can more precisely be characterized by adopting culturomics [67]. The authors isolated a large amount of "uncultivable" species, and of 48 species, only 30 had been previously identified by metagenomics.
Another factor that affects peri-implant bone conditions is abutment morphology [68,69]. Furthermore, specific abutment surface characteristics (anodization) seem to be associated with a better soft-tissue outcome (greater height of keratinized mucosa) [70].
In addition to microbiological conditions, there are several biomechanical aspects that influence maintenance of peri-implant crestal bone. Conical and parallel surfaces of the implant-abutment connection (internal hexagon of the employed implant) provide rotational stability and combine the advantages of reduced microgaps and micromovement [71]. The implant shoulder exhibits microthreads and the platform-switching concept to reduce peri-implant bone strain [56,60]. The innovative macro-design enables apical expansion, which thus increases stability, in addition to enabling bone-to-implant contact. Another essential factor is the thickness of the implant shoulder [71], which may be a weak point in the design of a short implant, due to potential elastic deformity under extra-axial loading. This feature may be the reason for non-inflammatory peri-implant crestal bone loss. We addressed this aspect by splinting adjacent implants wherever possible [72]. Brenner et al. [18] and Pommer et al. [53] have suggested that prosthodontic factors should be considered to avoid screw loosening, component fracture, loss of marginal bone, or even loss of osseointegration. Different attachment systems (e.g., locator vs. ball attachment) and their varying heights clearly impact stress distribution at the implant neck [73]. Another essential factor is the peri-implant soft tissue in severely resorbed alveolar bone. Due to a loss of fixed/keratinized mucosa with progredient vertical bone atrophy, soft tissue conditioning must be meticulously addressed [74][75][76]. Therefore, based on the authors' experience, implantological treatment using short implants requires adequate surgical skills.
Comparable studies have reported a crestal bone loss of 0.5-0.6 mm at 24 months [41]. Other authors reported a mean loss of 0.57 mm, 0.55 mm, and 0.53 mm in the mandible (without significant change after 1 year) [7]. Conversely, randomized controlled trials demonstrated peri-implant marginal bone loss of 0.7 mm at 1 year after loading [77] and 1.1 mm [4] in the maxilla, which is the same median value as that measured in the present heterogeneous study cohort over the 3-year period (at first year: , respectively). Long-term results from a randomized controlled trial, at 8 years after loading, showed that short implant patients lost significantly less crestal bone in the posterior mandible (an average of 1.6 mm vs. 2.5 mm in the augmented group) [78].
In comparison with conventional hollow-screw implants, concerns regarding expandable implants include the presence of gaps down to the apical region (potential microleakage that is comparable to distractible implants and endodontically treated teeth) and potential technical complications (the possibility that the implant cannot resist the load transmission). Both aspects were examined in vitro, which means the implant was submitted to microbiological and mechanical tests. Over the 3-year follow-up period, and in accordance with the study's initial results, we did not observe any inflammatory signs in the apical region, either clinically or radiologically (Figure 6a,b and Figure 7a,b), or technical complications in terms of fracture.

Limitations of the Study and External Validity
The main limits of the study are its single-arm study design, and thus the inherent lack of a control group, and the heterogeneous baseline characteristics of the study cohort. The authors regard systemic comorbidities and local conditions as confounding factors. Therefore, generalization of the present findings to other settings should be performed with caution. Nevertheless, in terms of validating the initial outcomes, this larger trial sought to reevaluate the long-term safety of a specific short dental implant system. Due to their broad indication categories, the results of this study should be replicable in other settings with comparable patient characteristics and practitioner experience. In terms of future research, a randomized clinical trial should compare the presented macro-design with other short implants (e.g., a root-shaped, progressive thread design) in order to reduce the risk of confounding.

Conclusions
Within the limitations of a single-arm study and the mean 3-year follow-up period, the results demonstrate a reliable improvement in functional oral rehabilitation, especially for elderly patients whose general and local conditions make implantation difficult. The status of the alveolar process atrophy, the need for minor augmentation procedures, and the type of prosthetic rehabilitation (removable vs. fixed) significantly influenced the implant survival rate. Measures of OHRQOL were considerably enhanced. The tested implant was useful in terms of all bone qualities, as it exhibited high initial and secondary biomechanical stability in the maxilla and mandible. Median 3-year crestal bone changes under implant loading demonstrated maintenance of the peri-implant alveolar bone (stabile bone remodeling) in both jaws.