1 INTRODUCTION

It is well known that tooth loss triggers progressive bone resorption and three-dimensional alveolar bone loss.¹ Indeed, the alveolar ridge is rapidly reabsorbed during the first 6 months after dental extraction and several factors such as the presence of periodontal disease, periapical pathology, or trauma, can increase resorption even before tooth loss.² The residual bone volume after extractions may not be sufficient for the placement of standard implants.^3-6

Commonly, an atrophic or severely deficient edentulous ridge will require bone augmentation either simultaneous to implant placement or in a staged approach.⁷ High predictable results, low complication rates, and an implant survival rate of 97%–100% have been reported for horizontal bone augmentation procedures.^{7, 8} However, vertical ridge augmentation (VRA) is a more technically and biologically challenging technique and has been associated with higher complication rates and less predictable results, due to its high sensitivity.^9-11

For this reason, other approaches, such as short dental implants, have been proposed.^{12, 13} However, on many occasions, the bone height is insufficient for the placement of short implants¹⁴ and they are not normally used in the anterior maxillary region because of the high aesthetic demands of this area.¹⁵

For these reasons, VRA techniques are used to provide adequate bone volume and to achieve an adequate aesthetic result in various clinical situations.¹⁶

Many systematic reviews about VRA reported data regarding vertical bone gain (VBG) and/or complication rates and/or implant survival using the different surgical approaches: Guided Bone Regeneration, inlay graft, onlay graft, distraction osteogenesis, or split bone block technique.^17-23

During the consensus report on Bone Regeneration in 2018, Jepsen et al. stated that VRA is a high technique-sensitive surgical intervention where the key elements are appropriate graft/barrier stabilization, the appropriate flap management, and soft tissue closure. The results of the systematic review showed a weighted mean clinical VBG of 4.16 mm (95% confidence interval [95% CI: 3.72–4.61 mm). The weighted mean complication rate was 16.9% (95% CI: 12.5–21), influenced by the type of procedure with a 47.3% complication rate for distraction osteogenesis, 12.1 % guided bone regeneration, and 23.9% for the use of blocks. The weighted mean bone loss of studies with at least 12 months follow-up, was 1.01 mm (distraction osteogenesis showed a bone loss of 1.4 mm, guided bone regeneration of 0.99 mm, and autologous bone blocks of 0.77 mm).

Although most of systematic reviews reported data related to peri-implant bone loss (PBL) of implants placed in augmented sites using different approaches, none has been done primarily about this issue, considering PBL as the primary outcome.¹¹

Since the patient, intervention, comparison, outcome (PICO) question and primary outcomes are strictly related together and they strongly influence the selection of the studies and the data collected from them, the lack of a dedicated search strategy on PBL is the rationale of the present systematic review and meta-analysis (MA).

Furthermore, a systematic review dedicated to PBL can draw more consistent data on implant survival and success rates, that are closely related to PBL and follow-up.

The first objective of this systematic review was to evaluate if there are significant differences between the different VRA techniques in relation to the PBL over time, considering results after three different follow-up periods (short term, medium term, and long term). The secondary objectives were to assess whether any differences exist between VBG, complication rates, and implant survival and success.

2 MATERIALS AND METHODS

This systematic review fully adhered to guidelines of Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA)²⁴ and was registered at the PROSPERO platform under code number CRD4201912958 (https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=129587).

The criteria for selecting studies to be included in the systematic review were based on PICO strategy. The focused question is: Which is the VRA technique having less PBL over the years?

Population: Patients over 18 years of age, showing good general health, with different vertical deficiencies. Intervention: Treatment of vertical deficiencies by means of different VRA techniques. Comparison: guided bone regeneration (GBR; nonresorbable, resorbable membranes and titanium mesh), Onlay graft (ONLAY), not covered and covered with a resorbable membrane), Inlay graft (INLAY), Distraction Osteogenesis (DISTR), and Split bone block technique (SBB). Outcomes of interest: the changes of crestal bone levels around implants placed in vertically augmented alveolar ridges.

3 RESULTS

3.1 Study selection

The electronic search identified 3881 publications from 1996 to 2021. Five more articles were added after the manual search. Overall, 433 articles were considered suitable for abstract screening, 324 of which were excluded. Hence, 109 articles were judged eligible for full-text assessment and then 66 articles were excluded because they did not meet the inclusion criteria. Finally, 43 studies were included and subjected to data extraction (Graph 1). The reasons for excluding the remaining articles during this phase are described in Table A5.

3.4 Pairwise MA—All studies

3.4.1 Primary outcome: PBL over time

Considering all 42 included studies, the weighted mean estimate (WME) of PBL values was 1.38 mm (CI: 1.10–1.66) after a mean follow-up of 41.0 ± 27.8 months (median 36 months; range: 12–120 months). Complete data for each technique were reported in Table 5.

TABLE 5. Meta-analysis for peri-implant bone loss at different follow-up periods.

Follow-up period	VRA techniques	No. studies	No. patients	No. implants	Weight mean esitmate (mm)	95% CI (mm)	p-Value	Heterogeneity (I2) (%)	p-Value (heterogeneity)
Overall	GBR	14	320	779	1.08	0.88–1.27	<0.001	92.0	<0.001
	INLAY	14	321	1408	1.78	1.09–2.47	<0.001	99.9	<0.001
	ONLAY	13	389	1177	1.31	0.87–1.75	<0.001	98.2	<0.001
	DISTR	4	75	263	1.81	1.31–2.31	<0.001	95.6	<0.001
	SBB	2	163	377	0.66	0.55–0.77	<0.001	34.8	0.215
	Overall	43	1268	4004	1.41	1.13–1.69	<0.001	99.6	<0.001
Short term	GBR	8	315	756	0.76	0.56–0.97	<0.001	90.6	<0.001
	INLAY	10	321	1206	1.08	0.60–1.55	<0.001	99.9	<0.001
	ONLAY	6	389	1156	0.99	0.52–1.46	<0.001	98.3	<0.001
	DISTR	3	75	258	1.11	0.72–1.50	<0.001	95.0	<0.001
	SBB	1	163	355	—	—	—	—	—
	Overall	27	1263	3731	0.95	0.69–1.22	<0.001	99.7	<0.001
Medium term	GBR	10	234	552	1.12	0.89–1.35	<0.001	87.8	<0.001
	INLAY	9	257	1058	1.50	0.75–2.25	<0.001	99.6	<0.001
	ONLAY	6	293	1035	1.38	0.38–2.17	<0.001	99.3	<0.001
	DISTR	2	65	191	1.21	0.85–1.58	<0.001	94.2	<0.001
	SBB	1	163	355	—	—	—	—	—
	Overall	26	1012	3191	1.27	0.83–1.72	<0.001	99.7	<0.001
Long term	GBR	6	152	233	1.13	0.71–1.55	<0.001	91.1	<0.001
	INLAY	9	216	758	2.13	1.62–2.64	<0.001	98.8	<0.001
	ONLAY	7	270	825	1.42	0.62–2.23	0.001	99.0	<0.001
	DISTR	1	14	40	2.60	2.13–3.07	—	—	—
	SBB	2	139	327	0.66	0.55–0.77	<0.001	34.8	0.215
	Overall	24	791	2183	1.62	1.13–2.10	<0.001	99.6	<0.001

• Abbreviation: VRA, vertical ridge augmentation.

The lowest pooled PBL value was found for SBB (WME: 0.66 mm, CI: 0.55–0.77, n studies = 2) over a mean period of 65.0 ± 11.4 months (median 65.04—range: 12–120 months). All other techniques showed similar values compared with the mean values, without significant statistical differences, except for GBR showing a lower pooled PBL value (WME: 1.06 mm, CI: 0.87–1.26) with respect to Distraction Osteogenesis (WME: 1.81 mm, CI: 1.31–2.31).

The Forest plot related to overall PBL is shown in Figure 1.

3.4.2 PBL at short-term follow-up

A total of 26 studies reported PBL values after a short-term follow-up with a WME of 0.95 mm (CI: 0.69–1.22). Among these studies, 10 evaluated PBL values around 1206 implants after inlay graft, 8 evaluated PBL values for 756 implants after GBR, 6 evaluated PBL values around 1156 implants after onlay graft, 3 evaluated PBL values around 258 implants after DISTR, and only one evaluated PBL values around 355 implants after SBB and. Complete data about the number of patients and implants for each technique are given in Table 5.

After a short-term follow-up, similar PBL values were observed for GBR (WME: 0.76 mm; Cl: 0.56–0.97), onlay graft (WME: 0.99 mm; CI: 0.52–1.46), inlay graft (WME: 1.08 mm; Cl: 0.60–1.55) and distraction osteogenesis (WME: 1.11 mm; CI: 0.72–1.50). Only one study reported results on SBB,²⁵ with a PBL value of 0.21 ± 0.18 mm, lower than all other techniques included in the MA. Forest plots are shown in Figure 2F.

3.4.3 PBL at medium-term follow-up

A total of 26 studies reported PBL values after a medium-term follow-up with a WME of PBL of 1.27 mm (CI: 0.83–1.72). Among these studies, 10 evaluated PBL values around 552 implants after GBR, 9 evaluated PBL values for 1058 implants after inlay graft, 6 evaluated PBL values around 1035 implants after onlay graft, 2 evaluated PBL values around 191 implants after DISTR, and only one evaluated PBL values around 355 implants after SBB. Complete data about the number of patients and implants for each technique were reported in Table 5.

At medium-term follow-up, similar PBL values were reported for the different techniques: GBR (WME: 1.12 mm; CI: 0.89–1.35), DISTR (WME: 1.21 mm; CI: 0.85–1.58), onlay graft (WME = 1.38 mm; CI: 0.58–2.17), and inlay graft (WME: 1.50 mm; CI: 0.75–2.25). No significant differences were observed, except for SBB. Similarly, to the results after a short follow-up, Khoury et al.²⁵ was the only study reporting data on SBB, with a value of PBL of 0.26 ± 0.21 mm, again lower than all other techniques. Forest plots were reported in Figure 3.

3.4.4 PBL at long-term follow-up

A total of 24 studies reported PBL values after a long-term follow-up with a pooled PBL of 1.62 mm (CI: 1.13–2.10). Among these studies, 9 evaluated PBL values around 758 implants after inlay graft, 7 evaluated PBL values around 825 implants after onlay graft, 6 evaluated PBL values around 233 implants after GBR, 2 evaluated PBL values around 327 implants after SBB, and only one evaluated PBL values around 40 implants after DISTR. Complete data about the number of patients and implants for each technique are given in Table 5.

As opposed to short- and medium-term follow-up, in the long-term follow-up similar pooled PBL values were observed between SBB (WME: 0.56 mm; CI: 0.30–0.82), GBR (WME: 1.11 mm; CI: 0.76–1.47) and onlay graft (WME: 1.59 mm; CI: 0.80–2.37). However, the PBL values of SBB and GBR were significantly lower than those for inlay graft (WME: 1.99 mm; Cl: 1.66–2.32); while no significant differences were observed between Inlay and Onlay (Graph 2). Only one study reported results on DISTR giving a mean PBL value of 2.6 ± 0.9 mm.⁵¹ Forest plots are shown in Figure 4.

3.4.5 Secondary outcomes

Of the 43 included studies reporting data about PBL over time, none provided all secondary outcomes analyzed in this systematic review (VBG, HCR, SCR, SURV, SUCC, PROMs, PPD, GI, or BI, Mucositis and Peri-implantitis). These parameters have not been assessed in each single study and data analysis was not appropriate for all the variables. The authors agreed to report only data related to VBG, HCR, SCR, SURV, and SUCC after MA. Details about all secondary outcomes are described in Tables 6 and 7.

TABLE 6. Meta-analysis of secondary outcomes.

Outcomes	VRA techniques	No. studies	Weight mean estimate (%)	95% CI	p-Value	Heterogeneity (I2) (%)	p-Value (heterogeneity)
Surgical complications	GBR	5	14.6	7.7–22.9	<0.01	15.6	0.32
	INLAY	9	41.6	31.2–52.3	<0.01	59.6	0.01
	ONLAY	8	5.1	2.0–9.4	<0.01	31.5	0.17
	DISTR	0	—	—	—	—	—
	SBB	1	—	—	—	—	—
	Overall	23	19.4	11.5–28.5	<0.01	87.7	<0.01
Healing complications	GBR	13	15.0	9.3–21.7	<0.01	48.2	0.02
	INLAY	11	13.8	8.1–20.6	<0.01	41.8	0.07
	ONLAY	10	13.6	7.4–21.0	<0.01	55.0	0.02
	DISTR	3	43.0	2.2–90.6	0.02	—	—
	SBB	2	6.8	3.3–11.1	<0.01	—	—
	Overall	36	15.9	11.7–20.4	<0.01	68.2	<0.01
Vertical bone gain	GBR	12	4.43	3.88–4.98	<0.01	89.4	<0.01
	INLAY	4	5.81	5.41–6.20	<0.01	77.2	0.02
	ONLAY	6	5.06	2.84–7.29	<0.01	99.0	<0.01
	DISTR	3	7.50	4.01–0.99	<0.01	96.0	<0.01
	SBB	2	6.64	4.78–8.50	<0.01	95.7	<0.01
	Overall	24	5.17	4.60–5.73	<0.01	97.6	<0.01
Implant survival	GBR	14	99.9	99.3–100.0	<0.01	95.6	0.76
	INLAY	14	94.6	92.2–96.6	<0.01	56.5	0.03
	ONLAY	13	97.1	92.9–99.7	<0.01	88.4	<0.01
	DISTR	4	99.2	95.6–100.0	<0.01	59.8	0.06
	SBB	2	99.5	98.2–100.0	<0.01	—	—
	Overall	43	97.9	96.6–99.0	<0.01	77.8	<0.01
Implant success	GBR	10	98.2	94.1–100.0	<0.01	81.7	<0.01
	INLAY	8	83.6	74.4–91.1	<0.01	90.4	<0.01
	ONLAY	8	89.1	81.6–94.9	<0.01	83.6	<0.01
	DISTR	2	94.1	90.0–97.3	<0.01	—	—
	SBB	0	—	—	—	—	—
	Overall	25	92.2	88.3–95.5	<0.01	89.0	<0.01

• Abbreviation: VRA, vertical ridge augmentation.

TABLE 7. Secondary outcomes of the included studies in meta-analysis.

Author and pubblication year	SCR (%)	HCR (%)	VBG (mm)	SURV (%)	SUCC (%)
Simion et al.30	NM	18.2	NM	100.0	100
	NM	18.8	NM	100.0	100
Nyström et al.31	NM	NM	NM	72.8	72.8
Chiapasco et al.32	NM	8.1	9.9 ± 3.4	100.0	94.2
Hallman et al.33	NM	NM	NM	87.0	70
	NM	NM	NM	94.5	52
Polo et al.34	NM	NM	4.2 ± 1.9	94.1	NM
Chiapasco et al.35	33.3	7.7	NM	94.5	82.9
Pieri et al.74	6.3	6.3	3.71 ± 1.24	100.0	93.2
Marchetti et al.37	NM	16.7	NM	89.4	67.3
Canullo et al.38	NM	10.0	5.3 ± 1.9	100.0	100
Urban et al.39	NM	2.8	5.5 ± 2.29	100.0	94.7
Corinaldesi et al.40	NM	23.1	5.45 ± 1.81	100.0	100
	NM	9.1	4.50 ± 1.16	100.0	91
Nyström et al.41	NM	NM	NM	85.0	NM
Felice et al.75	40.0	10.0	NM	100.0	90
	20.0	20.0	NM	100.0	86.9
Merli et al.43, 44	NM	36.4	2.16 ± 1.51	100.0	100
	NM	45.4	2.48 ± 1.13	100.0	100
Canullo et al.45	NM	5.0	5.61 ± 1.45	100.0	NM
Todisco46	NM	10.0	5.2 ± 1.5	100.0	NM
De Riu et al.47	13.3	6.6	3.77 ± 0.88	95.0	NM
Pieri et al.48	3.3	10.0	1.71 ± 0.75	100.0	100
Mertens et al.49	NM	13.3	NM	97.9	95.7
Pérez-Sayáns et al.50	NM	85.7	NM	100.0	NM
Kim et al.51	NM	42.9	8.4 ± 2.6	97.3	92.7
	NM	28.6	6.5 ± 2.3	94.1	90.2
Dottore et al.52	NM	NM	7.0 ± 2.6	95.5	90.9
	NM	NM	6.5 ± 1.6	95.5	90.9
Peñarrocha-Oltra et al.53	5.0	25.0	NM	95.6	91.1
Boven et al.54	8.3	NM	6.5 ± 2.4	98.7	NM
Fontana et al.55	NM	23.8	4.02 ± 1.53	93.6	71.9
Yu et al.56	NM	33.4	5.7 ± 1.09	100.0	NM
Barone et al.73	NM	20.0	6.0 ± 0.7	100.0	93.8
	NM	30.0	7.4 ± 0.8	100.0	82.4
Park et al.58	NM	NM	4.54 ± 2.48	100.0	NM
	NM	NM	3.90 ± 0.85	100.0	NM
Pieri et al.36	12.5	20.8	NM	95.5	NM
Rokn et al.59	NM	0.0	NM	100.0	100
	27.3	45.5	NM	100.0	100
Felice et al.60	6.9	0.0	NM	91.7	NM
	53.3	23.3	NM	95.1	NM
Bolle et al.61	5.0	0.0	NM	95.4	NM
	25.0	10.0	NM	97.9	NM
Chiapasco et al.62	1.4	8.3	NM	98.5	NM
Marconcini et al.63	35.0	8.7	5.6 ± 0.8	95.5	90.9
Khoury et al.25	2.1	5.4	7.6 ± 3.4	98.7	NM
Geng et al.64	28.6	30.2	5.4 ± 0.2	96.7	95.3
Felice et al.65	0.0	5.0	NM	95.1	NM
	35.0	15.0	NM	93.6	NM
Felice et al.66	40.0	13.3	NM	92.3	NM
	66.7	0.0	NM	96.7	NM
Esposito et al.67	40.0	5.0	NM	96.9	NM
	70.0	15.0	NM	93.5	NM
Chiapasco et al.68	4.0	4.0	NM	98.1	85.2
Pistilli et al.69	20.7	6.9	6.03 ± 1.09	98.4	NM
Cucchi et al.70	5.0	15.0	4.2 ± 1.0	100.0	100
	15.8	21.1	4.1 ± 1.0	100.0	100

• Abbreviations: HCR, healing complications rate; NM, parameter not measured. SCR, surgical complications rate; SUCC, success rate; SURV, survival rate; VBG, vertical bone gain.

The rate of surgical complications was evaluated in 23 studies with a pooled value of 19.4% (CI: 11.5%–28.5%). Onlay graft (WME: 5.1%; CI: 2.0–9.4) showed lowest SCR; while, inlay graft (WME: 41.6%; Cl: 31.2–52.3) had the highest SCR compared with both onlay graft and GBR (WME: 14.6%; CI: 7.7–22.9). No MA was carried out for DISTR (no data available) and SBB (data were available from only one article²⁵). Forest plots are shown in Figure 5.

The rate of healing complications was assessed in 35 studies with a pooled value of 15.9% (CI: 11.7–20.4). The analysis showed similar values for all the techniques investigated, though SBB tended to have the lowest rate of complications (WME: 6.80%; CI: 3.3–11.1). Moreover, seven studies^{31, 33, 34, 42, 52, 54, 58, 76} did not report complications during the healing period. The Forest plot is shown in Figure 6.

Regarding VBG, the values reported in 23 studies provided a pooled value of 5.26 mm (CI: 4.73–5.79). The analysis of VBG showed similar values for DISTR (WME: 7.50 mm; Cl: 4.01–10.99), SBB (WME: 6.64 mm; CI: 4.78–8.50), inlay graft (WME: 5.81 mm; 95% Cl: 5.41–6.20), onlay graft (WME: 5.06 mm; CI: 2.84–7.29), and GBR (WME: 4.43 mm; CI: 3.88–4.98). The Forest plot is shown in Figure 7.

Considering the overall values of the PBL/VBG ratio, SBB reported the lowest coefficient of 0.09 (n = 2); inlay had a coefficient of 0.15 (n = 5), GBR and Onlay showed a similar value of 0.22 but a different number of studies (n = 14 and n = 6, respectively), and finally DISTR reported the highest coefficient of 0.35 (n = 3). All data related to the overall PBL/VBG ratio and respective values in each follow-up period are listed in Table 8.

TABLE 8. Ratio PBL-VBG.

Group	No. studies	Mean	SD	95% CI
Overall
DISTR	3	0.35	0.25	−0.29; 0.98
GBR	14	0.23	0.07	0.18; 0.27
INLAY	5	0.15	0.03	0.11; 0.18
ONLAY	6	0.22	0.12	0.10; 0.35
SBB	2	0.09	0.06	−0.47; 0.65
Total	30	0.22	0.12	0.17; 0.26
Short term
DISTR	2	0.35	0.38	−8.01; 16.76
GBR	8	0.21	0.08	1.77; 2.95
INLAY	4	0.13	0.02	2.98; 4.10
ONLAY	3	0.24	0.09	−1.55; 6.86
SBB	1	0.03	—	—
Total	18	0.20	0.13	2.43; 3.53
Medium term
DISTR	1	0.12	—	—
GBR	9	0.23	0.07	0.18; 0.28
INLAY	2	0.16	0.02	−0.07; 0.38
ONLAY	3	0.24	0.15	−0.12; 0.60
SBB	1	0.03	—	—
Total	16	0.20	0.09	0.15; 0.25
Long term
DISTR	1	0.31	—	—
GBR	7	0.25	0.07	0.19; 0.32
INLAY	1	0.21	—	—
ONLAY	3	0.27	0.15	−0.11; 0.64
SBB	2	0.10	0.05	−0.36; 0.56
Total	14	0.23	0.10	0.18; 0.29

• Note: Ratio between PBL and VBG values of each technique analyzed per follow up period.
• Abbreviations: PBL, peri-implant bone loss; SD, standard deviation; VBG, vertical bone gain; VRA, vertical ridge augmentation.

The implant survival was calculated in all the studies (WME: 97.9%; CI: 96.6–99.0). The analysis revealed similar values for GBR (WME: 99.9%; CI: 99.3–100.0), SBB (WME: 99.5%; CI: 98.2–100.0), DISTR (WME: 99.2%; CI: 95.6–100.0), onlay graft (WME: 97.1%; CI: 92.9–99.7), and inlay graft (WME: 94.6%; CI: 92.2–96.6). Significant differences were observed between GBR and inlay grafts in these studies. The Forest plot is shown in Figure 8.

Finally, the implant success rates were evaluated in 24 studies with a pooled proportion of 92.2% (CI: 88.3–95.5). As with the survival rates, the analysis revealed similar values for GBR (WME: 98.2%; CI: 94.1–100.0), DISTR (WME: 94.1%; CI: 90.0–97.3), onlay graft (WME: 89.1%; CI: 81.6–94.9), and inlay graft (WME: 83.6%; CI: 74.5–91.1). Significant differences were observed between GBR and inlay grafts in these studies. No mean value could be established for SBB, because no articles reported implant success rates. The Forest plot is shown in Figure 9.

3.4.6 Other outcomes

Due to the insufficient amount of data for each technique, no MAs were conducted on PPD, GI, PROM's, and peri-implantitis data.

The PPD was provided in seven studies with a mean value of 2.36 ± 0.70 mm. Of these seven publications, four of them were about GBR^{30, 39, 70, 74} two studies about onlay graft,^{48, 54} and one about inlay graft.³⁵ Moreover, the GI was assessed in two studies^{54, 70} with an interval between 0 and 1 using the modified gingival index; while the incidence of mucositis and peri-implantitis was reported in six studies,^{36, 39, 44, 54, 65, 68} with an average rate of 7.42%. Finally, the PROMs were investigated in seven publications, showing that patient's perception of the aesthetics and functionality is optimal, on average, for 93.04% (range: 85%–100%) of patients.

Complete detailed results of MA about secondary outcomes are given in Table 9.

TABLE 9. Secondary outcomes not included in meta-analysis.

Author and pubblication year	PPD (mm)	GI	MPR (%)	PROM's (%)
Simion et al.30	2.6 ± 0.7	NM	NM	NM
	2.0 ± 0.8	NM	NM	NM
Chiapasco et al.35	2.6 ± 1.2	NM	NM	92.3
Pieri et al.74	3.0 ± 0.7	NM	NM	NM
Urban et al.39	3.0 ± 0.6	NM	3.7	NM
Merli et al.43, 44	NM	NM	T: 0	NM
	NM	NM	C: 0	NM
De Riu et al.47	NM	NM	NM	94.0
Pieri et al.48	4.3 ± 0.7	NM	NM	93.0
Mertens et al.49	NM	NM	NM	100
Boven et al.54	1.4 ± 0.6	0	0.0	100
Pieri et al.36	NM	NM	31.4	NM
Chiapasco et al.62	NM	NM	NM	90.0
Marconcini et al.63	NM	NM	NM	85.0
Khoury et al.25	NM	NM	0.98	NM
Felice et al.65	NM	NM	NM	100
	NM	NM	13.6	NM
Cucchi et al.70	1.8 ± 0.5	1	NM	NM
	1.8 ± 0.8	1	NM	NM
Chiapasco et al.68	NM	NM	5.5	NM

• Abbreviations: C, control group; GI, Gingival index; NM, parameter not measured; PPD, probing pocket depth; PROM, patient-reported outcome measures; T, test group.

3.5 Pairwise MA—Only prospective studies (cohort studies and RCTs)

Considering only study with prospective designs, including prospective cohort studies, nonrandomized clinical trial, and randomized clinical trials, a total of 26 articles were analyzed in the second MA.

The overall results provided data for a total of 1663 implants with a mean follow-up of 28.9 months, giving a WME for PBL of 1.33 mm (CI: 0.94–1.72).

SBB (WME: 0.66 mm, CI: 0.55–0.77) and GBR (WME: 0.79 mm, CI: 0.58–1.00) had the lowest PBL values with no significant differences between them. Inlay graft, Onlay graft, and DISTR had values slightly higher and similar between them, that were 1.65 mm (0.76–2.53), 1.59 mm (0.77–2.52), and 1.53 mm (1.08–1.97), respectively. Complete data were reported in Table 10.

TABLE 10. Meta-analysis for peri-implant bone loss for RCTs and prospective studies.

VRA Techniques	No. studies	No. patients	No. implants	Medium follow-up (months)	Weight mean esitmate (mm)	95% CI (mm)	p-Value	Heterogeneity (I2) (%)	p-Value (heterogeneity)
GBR	7	118	351	20.5	0.79	0.58–1.00	<0.01	90.7	<0.01
INLAY	10	181	547	30.3	1.65	0.76–2.53	<0.01	99.9	<0.01
ONLAY	5	92	316	27.9	1.59	0.77–2.52	<0.01	99.3	<0.01
DISTR	3	60	122	23.4	1.53	1.08–1.97	<0.01	94.6	<0.01
SBB	2	139	327	65.0	0.66	0.55–0.77	<0.01	34.8	0.22
Overall	27	590	1663	28.9	1.33	0.94–1.72	<0.01	99.8	<0.01

• Abbreviations: RCTs, randomized clinical trial; VRA, vertical ridge augmentation.

Statistical analysis for PBL confirmed no statistical differences for all surgical techniques except for SBB and GBR compared with DISTR, as referred in the forest plot Figure 10.

Morevoer, the above-mentioned secondary outcomes were entirely described in Table 11 and statistical analyses for SCR, HCR, VBG, SURV, and SUCC were reported in Figures 11-15.

TABLE 11. Meta-analysis of secondary outcomes for RCTs and prospective studies.

Outcomes	VRA techniques	No. studies	weight mean estimate	95% CI	p-Value	Heterogeneity (I2) (%)	p-Value (heterogeneity)
Surgical complications	GBR	3	11.17	3.62–21.32	<0.01	10.01	0.34
	INLAY	8	44.11	32.82–55.69	<0.01	54.50	0.03
	ONLAY	3	8.67	0.68–21.74	0.01	—	—
	DISTR	—	—	—	—	—	—
	SBB	1	2.11	0.78–5.61	<0.01	—	—
	Overall	15	25.21	12.87–39.71	<0.01	89.06	<0.01
Healing complications	GBR	7	19.26	9.41–31.15	<0.01	53.95	0.03
	INLAY	8	10.65	5.91–16.33	<0.01	7.1	0.38
	ONLAY	4	10.78	3.81–19.97	<0.01	0.00	0.78
	DISTR	2	25.58	14.06–38.92	<0.01	—	—
	SBB	2	6.78	3.33–11.14	<0.01	—	—
	Overall	23	15.69	9.88–22.37	<0.01	71.51	<0.01
Vertical bone gain	GBR	6	3.94	3.11–4.78	<0.01	92.00	<0.01
	INLAY	2	6.16	5.31–7.01	<0.01	64.60	0.06
	ONLAY	2	2.73	0.75–4.75	<0.01	98.30	<0.01
	DISTR	2	7.05	1.47–12.64	<0.01	97.90	<0.01
	SBB	2	6.64	4.78–8.50	<0.01	95.70	<0.01
	Overall	14	4.93	3.94–5.93	<0.01	98.0	<0.01
Implant survival	GBR	7	100.00	99.39–100.00	<0.01	0.00	1.0
	INLAY	10	94.66	91.93–96.92	<0.01	45.80	0.05
	ONLAY	5	95.42	81.66–100.00	<0.01	93.10	<0.01
	DISTR	3	99.50	95.35–100.00	<0.01	—	—
	SBB	2	99.54	98.20–100.00	<0.01	—	—
	Overall	27	97.88	95.67–99.43	<0.01	83.81	<0.01
Implant success	GBR	5	99.67	97.77–100.00	<0.01	19.23	0.29
	INLAY	4	88.33	83.45–92.52	<0.01	37.90	0.17
	ONLAY	4	91.13	73.99–99.88	<0.01	92.53	<0.01
	DISTR	1	94.20	88.97–97.03	<0.01	—	—
	SBB	—	—	—	—	—	—
	Overall	14	94.90	90.30–98.25	<0.01	86.66	<0.01

• Abbreviations: RCTs, randomized clinical trial; VRA, vertical ridge augmentation.

The overall values for surgical and healing complications were respectively 15.69% (CI: 9.88–22.37) and 25.21 % (12.87, 39.71); the WME VBG was 4.93 mm (CI: 3.94–5.93); and implant survival and rates were respectively 97.88 % (CI: 95.67–99.43) and 94.90 % (CI: 90.30–98.25).

3.6 Pairwise MA—Only RCTs

Considering only study randomized clinical trials, a total of 11 articles were analyzed in the third MA, having the highest level of evidence. Four articles were about GBR,^{43, 44, 59, 70} six articles about inlay,^{52, 60, 61, 65-67} and one article compared inlay versus onlay.⁴² SBB and DISTR had no randomized studies to be included in this analysis.

The overall results provided data for a total of 491 implants with a mean follow-up of 25.1 months, providing a WME of PBL values 1.21 mm (CI: 0.93–1.48).

GBR showed the lowest value of PBL, giving a WME of 0.61 mm (CI: 0.51–0.70); this value is statistically lower than Inlay having a WME of 1.47 (CI: 0.89–2.04), that is in turn statistically lower than Onlay (WME: 2.88; CI: 2.18–3.58). Complete data were reported in Table 12.

TABLE 12. Meta-analysis of peri-implant bone loss only for RCTs.

Follow-up period	VRA techniques	No. studies	No. patients	No. implants	Medium follow-up (months)	Weight mean esitmate (mm)	95% CI (mm)	p-Value	Heterogeneity (I2) (%)	p-Value (heterogeneity)
Overall	GBR	4	62	201	24.0	0.61	0.51–0.70	<0.01	0.0	0.77
	INLAY	7	105	267	26.6	1.47	0.89–2.04	<0.01	97.2	<0.01
	ONLAY	1	10	23	18.0	2.88	2.18–3.58	<0.01	—	—
	DISTR	—	—	—	—	—	—	—	—	—
	SBB	—	—	—	—	—	—	—	—	—
	Overall	12	177	491	25.1	1.21	0.93–1.48	<0.01	95.70	<0.01

• Abbreviations: RCTs, randomized clinical trial; VRA, vertical ridge augmentation.

Statistical analysis for PBL confirmed statistical differences between GBR and Inlay, GBR and Onlay, and Inlay and Onlay, as referred in the forest plot Figure 16.

The secondary outcomes reported in RCTs were entirely described in Table 13 and statistical analyses for SCR, HCR, VBG, SURV, and SUCC were reported in five different forest plots (Figures 17-21).

TABLE 13. Meta-analysis of secondary outcomes for only RCTs.

Outcomes	VRA techniques	No. studies	Weight mean estimate	95% CI	p-Value	Heterogeneity (I2) (%)	p-Value (heterogeneity)
Surgical complications	GBR	2	13.30	3.13–27.64	<0.01	—	—
	INLAY	6	48.42	33.72–63.25	<0.01	58.00%	0.04
	ONLAY	1	20.00	5.67–50.98	0.03	—	—
	DISTR	—	—	—	—	—	—
	SBB	—	—	–	—	—	—
	Overall	8	34.76	20.57–50.30	<0.01	75.02%	<0.01
Healing complications	GBR	4	29.18	16.92–42.95	<0.01	26.27%	0.25
	INLAY	5	12.03	5.41–20.35	<0.01	22.60%	0.26
	ONLAY	1	20.00	5.67–50.98	0.03	—	—
	DISTR	—	—	–	—	—	—
	SBB	—	—	–	—	—	—
	Overall	10	18.43	11.30–26.62	<0.01	39.88%	0.07
Vertical bone gain	GBR	3	3.28	2.30–4.26	<0.01	89.80%	<0.01
	INLAY	1	6.64	5.83–7.44	<0.01	0.00%	0.59
	ONLAY	—	—	–	—	—	—
	DISTR	—	—	–	—	—	—
	SBB	—	—	–	—	—	—
	Overall	4	4.32	3.13–5.51	<0.01	93.60%	<0.01
Implant survival	GBR	4	100.00	98.98–100.00	<0.01	0.00%	1.00
	INLAY	7	96.17	93.24–98.42	<0.01	0.00%	0.92
	ONLAY	1	100.00	85.69–100.00	<0.01	—	—
	DISTR	—	—	—	—	—	—
	SBB	—	—	—	—	—	—
	Overall	11	98.64	97.03–99.73	<0.01	7.63%	0.37
Implant success	GBR	4	100.00	98.98–100.00	<0.01	0.00%	1.00
	INLAY	2	90.63	81.61–97.16	<0.01	—	—
	ONLAY	1	86.90	67.81–95.43	<0.01	—	—
	DISTR	—	—	—	—	—	—
	SBB	—	—	—	—	—	—
	Overall	7	98.06	94.02–100.00	<0.01	94.25%	0.01

• Abbreviations: RCTs, randomized clinical trial; VRA, vertical ridge augmentation.

GRADE assessment was used to define the level of evidence of the outcomes. In this systematic review, GRADE assessment was considered very low due to a high risk of bias, imprecision and inconsistency (Table 14).

TABLE 14. Summary of findings for GRADE statement of included studies.

The influence of vertical ridge augmentation techniques on peri-implant bone loss
Patient	Patients with different vertical deficiencies
Intervention	Vertical ridge augmentation techniques
Comparison	Guided bone regeneration (GBR; nonresorbable, resorbable membranes, and titanium mesh), Onlay graft (ONLAY), not covered and covered with a resorbable membrane), Inlay graft (INLAY), Distraction Osteogenesis (DISTR), and Split bone block technique (SBB)

Outcomes	Number of participants (Studies)	Relative effects (95% CI)	Certainty of evidence (GRADE)	Key messages in simple terms
Peri-implant bone loss	1127 (43)
PTFE vs. MESH		−0.03 (−1.31; 1.24)	⊕○○○a,b,c Very low	It is not possible to clearly identify the VRA technique having less peri-implant bone loss over the years
PTFE vs. COLLAG		−0.120 (−1.44; 1.20)
MESH vs. ONLAY		−0.079 (−1.39; 1.24)
COLLAG vs. SHORT		0.142 (−1.16; 1.44)
INLAY vs. ONLAY		−0.581 (−1.62; 0.46)
INLAY vs. SHORT		0.557 (−0.08; 1.19)
ONLAY vs. DISTR		−0.300 (−1.89; 1.29)

• Note: Study: RCTs, CCTs, prospective cohort studies, and retrospective cohort studies. 95% CI: 95% confidence interval. Domains that lower the level of evidence.
• Abbreviations: CCTs, controlled clinical trials; GRADE: Evidence grades. Grading of Recommendations Assessment, Development, and Evaluation; RCTs, randomized clinical trials.
• ^a Certainity of the evidence downgraded 1 level due to high risk of bias.
• ^b Certainity of the evidence downgraded 1 level due to inconsistency (substantial heterogeneity).
• ^c Certainity of the evidence downgraded 1 level due to Imprecision (wide confidence intervals).

4 DISCUSSION

VRA is one of the most challenging targets for bone regeneration in implant dentistry. It is a highly technique- and operator-sensitive procedure characterized by numerous intra- and postoperative complications.^{18, 77} However, the most recent reviews agreed that all surgical techniques significantly augment the clinical horizontal and vertical ridge dimensions.^{11, 17}

The terms of VRA have been used first in the half of 90s after some interesting studies about the GBR with nonresorbable membranes^78-80 then, the following researches have been focused on the predictability of implants placed in vertically agumented bone and the stability of their crestal bone levels over time.⁸¹

The primary purpose of this systematic review, conducted on 42 publications, was to analyze the changes in PBL over time of implants placed after different VRA techniques namely: GBR with nonresorbable barriers, resorbable barriers, or titanium mesh; Inlay Graft; Onlay Graft; Distraction Osteogenesis; or Split Bone Block technique.

The search strategy included studies published in the last 25 years, starting from 1996 because in the authors’ knowledges no studies about implant success and related peri-implant bone levels after VRA were published before 1996. A post hoc research was carried out using the above-mentioned search string including MeSH and related keywords and none of the studies published before 1996 matched the inclusion criteria.

Since the PBL is closely related to the follow-up time, the results were divided and analyzed for three different intervals: short-term, medium-term, and long-term follow-up.

While GBR, Inlay graft, and Onlay graft techniques had more than 12 studies each, the numbers of studies about SBBT and DISTR were limited. The small number of articles in the literature concerning SBB and DISTR related to PBL values did not make it possible to highlight statistically significant outcomes over all time periods. Indeed, to reach statistical significance in a MA, two articles are required at least for each outcome⁷² but in the short- and medium-follow-up there is only one study about SBB²⁵; and in the long-term follow-up, only one study about DISTR.⁵¹ Consequently, only GBR, Inlay graft, and Onlay graft have a meaningful MA for all follow-up periods.

Three different MAs were performed to completely investigate the primary outcomes of this review. The first one is MA including all studies, as retrospective cohort studies, prospective cohort studies, nonrandomized controlled studies, and randomized controlled studies, and it gives a comprehensive description of the PBL in an overall follow-up as well as in short-, medium-, and long- follow-up. The second one including all studies with prospective design (cohort studies, controlled clinical trials [CCTs], and RCTs) to reduce the risk of missing data and the risk of bias. A third MA including only RCTs allows to have the highest level of evidence and to understand the surgical techniques having stronger and solider data.

Considering the study design, only GBR, Inlay, and Onlay were evaluated in randomized or CCTs; SBB was evaluated only in prospective cohort studies and DISTR only in retrospective cohort studies.

In relation to data from RCTs, implants placed in bone vertically augmented using GBR seems to have less PBL (about 0.61 mm) than those placed in sites augmented by Inlay or Onlay bone grafts (about 1.47 and 2.88 mm, respectively). However, the number of studies is very limited as well as the direct comparison between different surgical techniques in the same study.

It is important to state that the different follow-up periods could influence the different values of PBL for each technique; for example, inlay grafts reported data after a follow-up of 60 months or longer,^{60, 65-67} while GBR had only RCTs with data after short or medium follow-up.^{43, 44, 59, 70}

For the short-term period, the overall MA indicated that all VRA techniques presented PBL data that were comparable to the PBL in native bone, reporting a mean values <1.0 mm. Up to now, a PBL threshold of 1.5 mm in the first year have been widely accepted and adopted in the most of the studies, as criteria for implant success^{27, 28}; however, this threshold was not based on scientific data and it is no longer acceptable in comtemporatry implant dentistry.⁸² the crestal bone loss were observed to be consistent with a good peri-implant health, that is <1.0 mm (0.68 ± 0.40 mm), over a short- and medium-term follow-up. Thus, the marginal bone loss that occurs in the early phase seemingly varies between brands and seems to be related to the implant design and implant abutment connection, but after this phase the most of implants experience no additional bone loss and exhibit PBL <1.0 mm associated with an apparently healthy peri-implant mucosa.^{83, 84}

The highest PBL value was reported by Polo et al.³⁴ for DISTR in the posterior mandibles (2.60 mm); otherwise, the lowest values (0.33 mm) were measured by Pieri et al.⁴⁸ after small VRA in the anterior maxilla, treated with mandibular block grafts covered with hydroxyapatite and resorbable membrane.

For the medium-term period, the overall MA indicated that GBR and DISTR showed slightly better values of PBL compared with inlay grafts; data on onlay grafts were highly variabile.

The major PBL values (3.17 mm) after 5 years of follow-up were reported by Nyström et al.,^{41, 76} who evaluated implants after Le fort 1 osteotomy and inlay grafts, in total edentulous atrophic maxillae, harvesting cortico-cancellous bone block from the iliac crest. The lower values (0.38 mm) were reported by Park et al.,⁵⁸ that only evaluated 26 implants in posterior mandibles augmented with allogeneic onlay bone blocks, after 2 years of follow up, using panoramic radiographs.

For the long-term period, the results of the MA showed significantly better results for SBB and GBR compared with inlay graft and DISTR. Once again, onlay graft showed a high variability and a mean value set between GBR and inlay graft values. The highest PBL was reported again by Nyström et al.^{41, 76} after Le fort 1 osteotomy and inlay grafts, in total edentulous atrophic maxillae at 5 and 10 years of follow-up (3.13 mm). The lowest PBL value was reported by Merli et al.,⁴⁴ after 6 years of follow-up, using two different GBR approaches: titanium-reinforced PTFE membranes (1.00 mm) and collagen membranes supported by titanium microplates (1.33 mm).

Up to now, no other recent review has evaluated studies in relation to PBL over time as primary outcome; however, some data related to changes in marginal bone levels after VRA with at least 12 months follow-up were described by Urban et al.,¹¹ that observed a significant bone loss over time with a mean value of 1.01 mm. The authors stated that marginal bone loss, that ranged from 1.40 mm for DISTR to 0.58 mm for GBR with resorbable membranes, was influenced by the type of procedure. The present review is primarily aimed to investigate the influence of the different VRA procedures on peri-implant bone levels after the functional loading.

Among the secondary outcomes, the report of surgical complications showed that VRA techniques had a mean value of about 19%, although most of them did not affect the final results of these procedures. In this regard, inlay graft presented the highest values, mainly due to the incidence of temporary paresthesia, usually resolving within a few days. It is important to note that in their randomized clinical trials about short implants versus inlay grafts in posterior mandibles, Felice et al.^{60, 66} and Esposito et al.,⁶⁷ accurately reported all kinds of neurological alterations from the day after surgery, recording about 55%–70% of temporary paresthesia, in comparison to other authors who included only neurological alterations lasting more than 1–3 months after surgery. The lowest values of surgical complications (1.4%) were reported by Chiapasco et al.⁶² in a large cohort study both in mandibles and maxillae with onlay grafts covered by resorbable membranes, although a retrospective design is not very reliable for obtaining data about surgical complications.

Concerning the healing complications, all VRA techniques showed similar percentage values, giving a mean value of about 16%. In contrast to surgical ones, this kind of complication usually affected the final results of bone augmentation, due to the partial or complete loss of bone graft.⁸⁵ The highest percentage was reported by Pérez-Sayáns et al.⁵⁰ in a prospective cohort study about DISTR using intrabony distractor (85.7%), where the most common problems were the deviation of the distraction axis, that can result in a partial reduction of the VBG. The lowest rate was reported by Urban et al.³⁹ in a retrospective cohort study on 35 patients treated with GBR using Ti-reinforced PTFE membranes (2.78%), where only one patient developed a fistula on top of the membrane area, two weeks after bone grafting.

With regard to VBG, the values reported in 23 studies provided a mean value of about 5.2 mm, showing similar values for the different techniques, with DISTR showing the highest value (9.9 mm) reported by Chiapasco et al.⁸⁶ As previously described, the lowest value (1.7 mm) was observed after mandibular block grafts by Pieri et al.⁴⁸ The observed data could be different compared with those reported in other systematic review focused on VBG as primary outcomes, because the present review aimed at understanding which techniques showed lowest PBL over time and consequently only included studies reporting data about PBL, and not all studies reporting the VBG values.^{11, 17, 20}

In regard to VBG, two interesting studies have been published recently but not included in this review because they did not report PBL even at 1 year of follow-up: Urban et al.⁸⁷ evaluated the use of reinforced PTFE meshes, reporting a mean VBG of 5.2 mm, that is equal to the value observed in this review; then, Chiapasco et al.,⁸⁸ using the customized CAD/CAM titanium meshes, had a mean value 4.8 mm.

In the present review the PBL/VBG ratio, that expresses the stability of augmented bone in relation to the entity of vertical augmentation, was measured for all type of procedures. Since some authors suggested the amount of vertical bone was directly related to the amount of observed bone resorption, a comparison of different PBL/VBG ratios was performed to understand which technique showed the more stable peri-implant levels independently of the amount of bone augmentation. This analysis found that SBBT and inlay graft are the two techniques in which the VBG had the lower influence on the maintenance of the peri-implant bone levels, although these data derived from fewer studies (n = 5) in comparison to those relative to GBR (n = 14).

The reasonable minimum vertical bone defect for performing a VRA could be a topic of debate, considering that the mean PBL was 1.4 mm after a mean follow-up longer than 3 years.

With regard to implant survival and implant success, the mean values were 97.9% and 92.2%, respectively. The type of procedure also influenced the outcomes: the SURV ranged from 94.6% for inlay graft to 99.9% for GBR; while the SUCC ranged from 83.6% for Inlay graft to 98.2% for GBR.

These mean rates were comparable to previous studies about implants placed in pristine bone after a mean observation period of at least 5 years.⁸⁹ Also, a recent systematic review confirmed that no statistically significant differences were observed between implants placed in pristine versus augmented sites for any outcome variables both at patient and at implant levels, respectively.¹⁵

In the present review, SBB have been distinguished from any other techniques due to its miscellaneous particularities, such as the autogenous bone lamina used as barriers similar to GBR, the apical space between the native bone and the lamina-like inlay graft, and the apposition of the autogenous bone above alveolar ridges like the onlay graft.⁹⁰ Further reviews could include SBB in GBR as well as in inlay or onlay grafts.

A matter of discussion could be the method used to express PBL values over time. The authors choose to divide the analysis of PBL over 3 different periods (short-term <1 year; medium-term 2–4 years; and long-term >5 years). There are mainly two reasons: (i) PBL is a variable strongly correlated with the length of follow-up and (ii) there are insufficient data related to each year of follow-up. This method made it possible to have a better comparability of PBL values in the first year compared with the following years, eliminating the effect of follow-up time on crestal bone levels.

The present systematic review and MA revealed that it is difficult to draw strong conclusions about the VRA technique having less PBL over the years due to the presence of some limitations.

The major limitation is the radiographic method to assess the PBL: the intraoral radiographs are surely more suitable and efficient for measuring the crestal bone levels around implants than orthopantomography.⁹¹ In this review, studies using both intraoral radiographs and/or orthopantomography have been included in order to not overlook some important studies about VRA. In this regard, the cone-beam computed tomography systems (CBCT) is more accurate, as it allows a three-dimensional measurement of peri-implant bone levels^{92, 93} and the scientific literature also validates the CBCT, instead of the bidimensional radiographs, for assessing the peri-implant bone levels^94-96 but the biological and financial costs for the patients limit its use.

Further limitations of this review are the number of included studies: there are several randomized controlled trials about VRA reporting complications, HBG, VBG, and/or implant survival, but few trials report PBL over time, which is the main inclusion criteria of this systematic review. For this reason, the authors incorporated all type of studies, including observational studies or trials comparing short versus long implant in augmented sites.^{59-61, 65-67} Moreover, we must note that the detection of small-study effect could limit the generalizability of this review's findings. However, the quality of analyzed studies was not weak because they included seven RCTs with moderate risk of bias and three with low risk of bias, without any RCT with high risk of bias; and more than half of nonrandomized trials had a moderate quality, confirming the presence of some significant studies about PBL after VRA.

Finally, this review could not describe separately the PBL values separately for mandible and maxilla because many authors did not break them down for maxilla and mandible, giving a mean result for both jaws. The scientific literature highlights how this outcome should be differentiated, since the mandible and the maxilla have a different bone density and different marginal bone loss.⁹⁷ Also, the type of implant–abutment connection, collar features, and surface characteristics can influence the changes in marginal bone levels as observed in different systematic reviews.^98-100

The strength of the present review is that it represents the only review evaluating PBL around implants in vertically augmented bone as the primary outcome. In order to give detailed information and better description about peri-implant bone levels over time, the data were analyzed in different follow-up intervals. The review included all types of VRA techniques: GBR including nonresorbable membranes, resorbable membranes, and titanium meshes; Onlay grafts; Inlay Grafts; Osteogenenic Distraction; and Split Bone Block technique. Then, it is important to highlight that two different MA approaches were used, offering more information about the primary outcome and related data. Finally, the most significant secondary outcomes were accurately reported and analyzed to provide a better understanding of the behavior of implants placed in regenerated bone.

5 FUTURE RESEARCH AND RECOMMENDATIONS

To have a high level of evidence and confirm the trends observed in the present review, the authors suggest the need for more CCTs and/or RCTs comparing two or more surgical techniques with standardized methods for radiographic measurements, such as periapical radiographs; more long-term studies reporting complete data after longer follow-up; and more multicenter cohort studies with a larger sample of patients.

Finally, the authors encourage continuing the long-term follow-up of treated patients in already published studies, to update the data over time.

6 CONCLUSIONS

The studies included in the present review directly addressed the focused question, reporting the changes of peri-implant bone levels during follow-up, finding a mean pooled value of PBL after a follow-up of more than 3 years that was <1.5 mm.

However, the outcomes of the MA should be interpreted with caution due to high variability with respect to number of studies for each technique, PBL measurements, duration of follow-ups, and eligibility criteria. While GBR, onlay graft, and inlay graft techniques are supported by data derived from a significant number of studies, there was a lack of studies about other surgical techniques (distraction and SBB).

Nevertheless, within the limitations of the present review, the primary findings of the MA, based on the changes between final and baseline values, have shown that the PBL could be influenced by the type of VRA procedure but there is a need to evaluate in randomized clinical trials the behavior of the peri-implant bone levels after long-term follow-up for all techniques in order to establish which is the VRA technique having less PBL over the years.

AUTHOR CONTRIBUTIONS

Alessandro Cucchi: Concept/design; critical revision of the article. Francesco Maiani: Statistic and data collection. Debora Franceschi: Drafting and reviewing article. Michele Sassano: Data collection and drafting. Antonino Fiorino: Statistic and data collection (responsible for the data analysis). Istvan A. Urban: Critical revision of the article. Giuseppe Corinaldesi: Concept/design.

CONFLICT OF INTEREST STATEMENT

The authors report no conflicts of interest related to this study.