The survival rate for the maxillary sinus floor augmentation crestal approach according to Del Fabbro et al. [
9] was 98.12% at 1 year, 97.40% at 2 years, 96.75% at 3 years, and 95.81% at 5 years. Similar to the lateral approach, it has a high predictability and is widely used in clinical settings. The advantage of the crestal approach is that the surgery is noninvasive and more rapid than the lateral approach. However, a disadvantage of this method is that as the surgery is performed blind, technical expertise is essential. Moreover, the amount of elevation is limited and any response to perforation of the sinus membrane is difficult.
Membrane thickness
In this study, patients with a history of maxillary sinusitis, those with suspected dental maxillary sinusitis such as apical lesions that cause mucosal thickness, those suspected of having allergies such as hay fever, those having findings considered to be mucocele, and those with a mucosal thickness ≥ 5 mm were excluded. Regarding the thickness of the maxillary sinus membrane, many reports state that the physiological thickness is 1–2 mm [
10]. The preoperative maxillary sinus membrane thickness in this study was 1.09 mm on average, which is equivalent to the value (average 1.17 mm: CT 1.33 mm: histology 0.48 mm) in the report on the physiological thickness of the maxillary sinus membrane using cadaver and CT [
10]. There is concern that increase in the thickness of the maxillary sinus membrane is associated with bacterial infection and a decrease in elimination function. However, there are also maxillary sinuses where even if the sinus membrane thickens, ventilation of the mucosal epithelium and the elimination function remain normal. In addition, there is no data or guideline that dentists and otolaryngologists alike can use regarding the relationship between mucous membrane thickness and decreased maxillary sinus function [
11]. Even when there is mild mucosal membrane thickening, we suggest that there will be no problem with ventilation and the elimination function provided there is no clinical symptom and the maxillary sinus ostium is patent. In a preoperative analysis by CBCT, Shanbhag et al. [
12] considered sinus membrane thickness ≥ 2 mm pathological, and when it was ≤ 2 mm, ostium patency was not classed as obstructed. Even when the thickness was 2–5 mm, impairment was seen in only 6.7% of cases while increased findings of ostium obstruction were reported in 24% and 35.3% of cases when the thickness was > 5 mm and > 10 mm, respectively. However, a mucosal thickness > 2 mm was seen in 60.6% of patients, and when a thickness of ≤ 2 mm was set as normal, there seemed to be an increase in patients being falsely diagnosed with maxillary sinusitis. Carmeli et al. [
13] examined thickness of the mucous membrane and ostium obstruction by CT. They reported that membrane thickness exceeding 5 mm is associated with the risk of ostium obstruction. Therefore, in this study, in addition to the medical history and the presence or absence of clinical symptoms such as nasal congestion, rhinorrhea, and headache, patients whose CBCT examination revealed mucosal membrane thickness below 5 mm, and the ostium was not affected by edema, were the subjects.
Length of the implant
Short implants were recently introduced owing to improvements in the implant surface. The European Association for Osseointegration (EAO) 2006 Consensus Conference defines short implants as those ≤ 8 mm in length. A systematic review based on meta-analyses through randomized comparison tests showed that the 5-year survival rate of implants ≤ 8 mm in length was not significantly different compared to longer implants [
14]. However, when using implants ≤ 8 mm in length, the superstructure was connected. In addition, Renouard and Niasand [
15] reported that although there is a higher tendency of failure in short implants, this trend decreases with a rough surface. Manzano et al. [
16] reported that as the maxillary molar region is fragile bone, implants < 10 mm in length have a higher risk of decreased survival. Since there is insufficient data on the long-term survival of non-connected implants < 10 mm in length in the maxillary molar region, in this study, we used implants 9.5 to 10 mm in length for single crown restorations, and implants 8 mm in length were connected to the superstructure. Implants < 8 mm in length were not used.
Residual bone height
The use of the crestal approach is limited to cases in which the vertical volume of the existing bone allows initial fixation of the implant. A prospective study by Zitzmann and Schärer [
17] recommended that the existing vertical bone dimension at the implant site should be at least 6 mm. Pjetursson et al. [
18] compared the prognosis of implants placed by the osteotome technique and those place by the conventional method and reported that the osteotome technique has a good prognosis if the vertical dimension of the existing bone is ≥ 5 mm. Călin et al. [
19] noted in a systematic review of the osteotome technique that if the vertical dimension of the existing bone is ≥ 4 mm, the success of the implant is not affected, but below that value, there is an impact. In this study, we obtained a good initial fixation in all cases with a mean bone vertical dimension of 6.35 ± 1.15 (4.2–8.5) mm, and there were no clinical problems in any implant at follow-up.
BH and SL
The BH was 3.17 ± 0.97 (1.20–5.50) mm immediately after surgery, and − 0.25 ± 1.19 (− 2.58–3.00) mm during the follow-up period of 81.9 ± 25.1 (36–112) months after surgery, with a decrease of 3.42 ± 1.43 (0.50–6.80) mm. A CBCT analysis of maxillary sinus floor augmentation using the lateral approach with β-TCP by Okada et al. [
7] revealed a decrease in volume at the implant site 6 months to 1 year after surgery, which stabilized after 3 years. Zijderveld et al. [
21] examined the prognosis of maxillary sinus floor augmentation using the lateral approach with autogenous bone and β-TCP with panoramic radiography and reported that in both cases, most absorption occurred during the first 7.5 months, with only small changes observed after 1.5 years. At the follow-up examination in this study, volumetric changes following grafting had already stabilized, and the BH in many implants at this point was approximately zero. This indicated that over time, the position of the maxillary sinus floor approached the implant apex.
BH loss at follow-up was negative in all of the implants, but we were able to obtain good long-term prognosis regardless of whether the BH values were positive or negative at follow-up. Okada et al. [
7] also reported that there were no clinical problems in maxillary sinus floor elevation using the lateral approach with β-TCP alone, regardless of the position of the maxillary sinus floor relative to the implant apex. Nedir et al. [
22] also reported that in graftless maxillary sinus floor augmentation using the crestal approach, long-term prognosis was good if the implant apex protruded into the maxillary sinus.
The crestal approach for maxillary sinus floor augmentation elevated the maxillary sinus floor less than the lateral approach. Zitzmann and Schärer [
17] reported that with the lateral approach, the volume after surgery was 10 mm when the implant was placed simultaneously and 12.7 mm when the implant was placed later, whereas it was 3.5 mm with the crestal approach. Pjetursson et al. [
23] reported that maxillary sinus floor augmentation with the crestal approach using deproteinized bovine bone mineral (DBBM) resulted in a mean elevation of 4.2 mm at an average of 3.2 years after surgery. In this study, the mean volume of maxillary sinus floor elevation immediately after surgery was 6.54 ± 1.51 (3.24–8.7) mm and 3.11 ± 1.37 (0.40–5.57) mm at follow-up.
The most common intraoperative complication of maxillary sinus floor augmentation with the crestal approach is perforation of the sinus membrane [
15]. It is difficult to treat, but Pjetursson et al. [
23] reported that it occurs in 10.4% of cases, whereas Călin et al. [
19] reported a mean of 6.28% (0–26%). While stable elevation of the maxillary sinus is desired for good implant prognosis, intraoperative elevation of the maxillary sinus floor must be minimized to avoid perforation of the sinus membrane. In this study, there was a strong correlation between the SL at follow-up and the IL; there was also a correlation with SL immediately after surgery. However, there was no correlation with BH immediately after surgery. As such, to secure long-term maxillary sinus floor elevation, the length of the implant protruding into the maxillary sinus is important. Si et al. [
24] performed maxillary sinus floor augmentation using the crestal approach and reported that the volume of elevation of the maxillary sinus floor is related to the length of the implant protruding into the sinus. In addition, there was a strong correlation between the decrease in the BH from immediately after surgery to follow-up and both the SL immediately after operation and the BH immediately after operation; there was also a correlation with the IL. Therefore, increased maxillary sinus floor elevation also increases absorption; the material above the implant apex tends to be absorbed.
It is not effective to elevate the maxillary sinus floor much beyond the implant apex considering absorption, and it should be kept within an appropriate range to avoid perforation of the sinus membrane. There are few studies of the appropriate elevation volume of the maxillary sinus floor in maxillary sinus floor augmentation with the crestal approach. Pjetursson et al. [
18] compared maxillary sinus floor elevation using the crestal approach with DBBM and without a graft. With a graft, there was a dome after surgery similar to that in the present study. This site gradually shrank, reducing the maxillary sinus floor elevation to 4.1 mm 3 years after surgery. However, cases without a graft showed slight structural changes between the implant apex and maxillary sinus floor membrane immediately after surgery, and the maxillary sinus floor elevation 3 years after surgery was 1.7 mm. As such, maxillary sinus floor elevation can be secured with higher predictability by using a graft.
Si et al. [
25] compared the use of a graft of a mixture of DBBM and autogenous bone to cases that did not receive a graft and reported that elevation was significantly higher 6 months after surgery in cases that received a graft, but decreased thereafter. Three years after surgery, the elevation in each case was approximately 3 mm. Thus, future studies should be conducted to clarify the optimal elevation volume of the maxillary sinus floor and selection of graft material in maxillary sinus floor augmentation with the crestal approach. Sonoda et al. [
26] surveyed the relationship between the volume of transplanted bone and the volume of maxillary sinus floor elevation using the crestal approach. In blind maxillary sinus floor augmentation using the crestal approach, if the membrane perforation risk is reduced by measuring the transplant material suited for the volume of maxillary sinus floor elevation needed, the success rate of the crestal approach may be increased.
This study has some limitations. First, the size of the study population is small. Further, other clinical parameters, such as sinus width and the type of defect (single or continuous) on which the crestal approach was applied, might affect the height of the augmented bone. In this study, these parameters were not taken into account. Therefore, these issues should be addressed in future studies.