Antibiotic Use in Dental Implant Procedures: A Systematic Review and Meta-Analysis

Background and Objectives: This project was developed from anecdotal evidence of varied practices around antibiotic prescribing in dental procedures. The aim of the study was to ascertain if there is evidence to support whether antibiotic (AB) use can effectively reduce postoperative infections after dental implant placements (DIPs). Materials and Methods: Following PRISMA-P© methodology, a systematic review of randomised controlled clinical trials was designed and registered on the PROSPERO© database. Searches were performed using PubMed®, Science Direct® and the Cochrane© Database, plus the bibliographies of studies identified. The efficacy of prophylactic antibiotics, independent of the regimen used, versus a placebo, control or no therapy based on implant failure due to infection was the primary measured outcome. Secondary outcomes were other post-surgical complications due to infection and AB adverse events. Results: Twelve RCTs were identified and analysed. Antibiotic use was reported to be statistically significant in preventing infection (p < 001). The prevention of complications was not statistically significant (p = 0.96), and the NNT was >5 (14 and 2523 respectively), which indicates that the intervention was not sufficiently effective to justify its use. The occurrence of side effects was not statistically significant (p = 0.63). NNH was 528 indicating that possible harm caused by the use of ABs is very small and does not negate the AB use when indicated. Conclusion: The routine use of prophylactic antibiotics to prevent infection in dental implant placement was found to be not sufficiently effective to justify routine use. Clear clinical assessment pathways, such as those used for medical conditions, based on the patients’ age, dental risk factors, such as oral health and bone health, physical risk factors, such as chronic or long-term conditions and modifiable health determinants, such as smoking, are required to prevent the unnecessary use of antibiotics.


Introduction
A dental implant (DI) placement (DIP) procedure, with artificial root placement, is the substitution of one or more missing or defective natural teeth. It has become the definitive therapy for the restoration of entire dental arches or for partially and fully edentulous patients [1]. DI surgery can be performed in an ambulatory (outpatient) setting using local or general anaesthesia depending upon the individual preferences, health statutes and case complexity [2]. The most common fabrication materials for DIs are metals and metal alloys, polymers and ceramics [3]. Metals and metal alloys have been widely used due to advantages over other materials, including good mechanical properties, biological compatibility and excellent corrosion resistance [3][4][5].
To function successfully, the implant must be surrounded by healthy and stable tissues [4][5][6]. DIP has become a routine treatment option due to its high survival and success rates, offering a predictable solution for tooth replacement [7,8]. Large-scale studies reported survival rates between 97% and 75% over 10 years and 20 years, respectively [9,10]. Although its success rate is very high, failures requiring removal occasionally occur [11,12].

Materials and Methods
This systematic review aims to assess the efficacy and safety of ABs among patients undergoing DI placement procedures. The review objectives included the identification of the type of ABs used, understanding how effective ABs are in reducing or preventing infection in DIPs and what is the incidence rate of AB adverse effects.
A systematic review was designed and implemented to investigate the efficacy and safety of ABs (independent of the type of ABs used, dose administered and timing of course of administration (preoperatively, postoperatively or both), as part of treatment, with or without AB therapy (no treatment or placebo) on DI failure and postoperative complications during DIP procedures.
The study protocol followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA-P©, Ottawa Hospital Research Institute and University of Ottawa, Ottawa, Canada) statement and the Cochrane Handbook for Systematic Reviewers' methodological guidelines (Cochrane collaboration, Oxford, UK) [20]. This systematic review was registered in the PROSPERO© (National Institute of Health, York, UK) database CRD42021269522. All statistical analyses were performed using Review Manager (RevMan ® ) statistical software version 5.4.1 (Cochrane collaboration, Oxford, UK). Risk ratios (RRs) with a 95% confidence interval (CI) were reported for dichotomous outcomes. To calculate the pooled RRs of the different studies, the Cochran-Mantel-Haenszel (M-H2) and 95% CI for effect measurement in meta-analysis was employed. Meta-analysis was used to combine outcome data from different studies using random effects or a fixed effect model [21].
The results were aggregated using the Cochran-Mantel-Haenszel (M-H2). The efficacy of the AB was assessed using the risk ratio (RR). The pooled RR with the corresponding 95% confidence interval (95% CI) was chosen as the effect size.
The effectiveness of the treatment with ABs to prevent infection, which can lead to implant failure or implant complications, was assessed using the number needed to treat (NNT), and safety was assessed by calculating the number needed to treat to cause harm (NNH), where an NNT of 1-5 is desirable [22,23]. All decimals in the NNT have been rounded up to the nearest whole number.
The statistical unit for "implant failure by patients", "post-operative infection" and AB adverse events was the number of patients. The statistical unit for implant failure by implant was the number of implants. No restriction criteria were applied for the definition of each outcome reported in this review.

Search Strategy
Search strategies utilised a combination of keywords using Medical Subject Headings (MeSH). No language limits were applied (date of last search: 17 January 2022). Publications included within PubMed© were searched since January 2000 until 17 January 2022.
The following term combination was used in the PubMed©, Science Direct ® and Cochrane© databases: ((Dental Implant insertion OR dental implants OR dental implant OR dental implant surgery OR Dental surgery OR dental implant placement OR oral implant surgery OR implant surgery [MeSH Terms]) AND (Implant failure OR implant loss OR Implant survival OR implant survival)) AND (Antimicrobial or antibacterial or antibacterial or antibiotics or antibiotic)) AND (randomised controlled trials OR controlled clinical trial OR randomised controlled trials OR random allocation OR single-blind method OR single-blind method OR double-blind method OR clinical trial OR clinical trials OR placebos OR placebo-controlled randomised clinical trial) AND (efficacy OR Effect OR effectiveness)) AND ("Antimicrobial" or "antibiotics", "prophylactic" OR "prophylaxis" OR "pre-operative" OR "preoperative" OR "peri operative" OR "postoperative" OR "preventative"). Issues of interest, according to the study population, intervention, comparative group, outcome measured and study design (PICOS [24]), were as follows: (

Selected Studies Summary
The initial search yielded 103 studies. An additional five RCTs were identified through a manual search of the similar articles and reference lists of included studies. After an investigation of titles and then abstracts, 74 articles were considered to be ineligible for inclusion. The full texts of 34 eligible articles were then reviewed, and 22 were excluded, leaving 12 studies that accurately fit the inclusion criteria. Selected RCTs were read in full text, by two independent reviewers. A PRISMA flowchart of the screening process for the systematic review and meta-analysis was developed (Figure 1). Following a review of these articles, 12 RCTs met the criteria for inclusion in this systematic review and meta-analysis.

Results
Five corresponding authors were contacted for clarification of the study design via e-mail inquiries [25][26][27][28][29], and two authors were contacted for clarification on the number of patients with implant failure, who provided the missing data; however, three authors did not respond [26][27][28]. Tables 1 and 2 describe the characteristics of the included studies and the reviewer comments. Six studies identified DI failure by number of patients, eight studies investigated the number of implants failed by the DI, eight assessed comparative data on postoperative complications and five reported data on the total adverse events that occurred between groups during the study period. Due to the small number of studies included in this meta-analysis, a funnel plot was not assessed [30].

Appraisal Summary
Caiazzo et al., 2011 [25] The RCTs concluded that single preoperative, combined preand postoperative antibiotic and postoperative antibiotic coverage versus no antibiotic treatment did not reveal any significant differences regarding implant failure in healthy patients under normal circumstances.
This study included healthy patients. The implant type places were titanium screw-type external hex, but the implant system was not reported. The overall reported success rate was 98.65 with two failures in the NOAB group. Study limitations: The mean age included in the metanalysis was between 42-and 52-years-old, which makes the results not applicable to the global population. There was no clear explanation of the ethics approval, surgical protocol, inclusion criteria or exclusion criteria. The authors were contacted by the reviewer who requested the missing data, and they responded that ''the data for clinical diagnosis for the infected surgical wound was not collected", which can be considered as reporting bias, especially, insufficient details of the assessors blinding were also prevalent. The planned sample size could not be obtained, and reasons were not explained.
Esposito et al., 2010 [31] Single administration of 2 g amoxicillin initiated one hour prior to implant placement could decrease early implant failure in patients not requiring a bone augmentation procedure. However, there were no significant differences for post-operative complications observed.
Implant procedures were conducted in 10 Italian private dental practices. The logistic regression model was fitted to determine confounding factors and the influence of the site effect on implant failure. No effect was observed between centres, and all had similar results (p = 0.97). The mean age of 49 and age range of 18-85 make the results applicable to the global population. Study limitations: No explanation provided on ethics approval or the preoperative surgical protocol. The planned sample size was 9 and could not be achieved and was insufficient to implicate a statistically significant difference.

Anitua et al., 2009 [26]
The finding does not support the use of 2 g of preoperative amoxicillin versus 2 g of placebo one hour before the placement of single dental implants in bone type II and III, and the results failed to show any advantages of preoperative antibiotics in the probability of having postoperative infection, characteristics of the saprophytic flora and incidence of antibiotic adverse events.
Single implants were placed in the maxilla or mandible. All patients received Biotechnology Institute (BTI) dental implants. Implant preparation rich in growth factor (PRGF) was applied in the operative protocol. Both researchers and patients remained blinded to the group receiving treatment. A logistic regression method was applied to determine confounding factors and the influence of different variables, including the centre, antibiotic or placebo, duration of the intervention, age and smoking habits. Study limitations: Participants had bone quality II or III; thus, the inclusion of patients with bone type I and IV could have altered the results. The planned sample size could not be achieved for a statistically significant difference. A larger sample size would be required to rule out the possibility of a difference between groups.
Abu Ta'a et al., 2007 [32] Single dose of preoperative 1 g amoxicillin and postoperative regimen of 500 mg four times per day for 2 days does not provide any advantages concerning peri-oral aerobic and anaerobic flora on nasal aerobic and anaerobic bacteria and also failed to have a noticeable impact on post-operative infection and implant failure, while strict asepsis applies during periodontal surgery.
High-end policy of asepsis was followed for asepsis prevention from the oral and nasal cavity. Both the surgical team and the patients were blinded to the groups. A team of several peri odontologist surgeons, with different levels of experience, and nurses performed the implant surgery. Postoperative infection and implant failure were assessed at the follow-up visits for up to 5 months after the implant procedures.
The implant system was not described in this study. Study limitations: Details of the enrolment period, explanation of how consent was obtained and details of the assessors blinding were missing. There were reports of patient-related variables with different conditions, such as blood-clotting problems, heavy smokers who consumed >40 cigarettes a day and patients with parafunctions included in the study, which may have affected the results. The study favoured the preoperative use of 3 g amoxicillin to improve the survival of dental implants, to effectively reduce postoperative pain, and improve interference with daily activity, considering that the duration of surgery and the number of implants had an effect on implant survival.
Only patients with the presence of a partial edentulous or edentulous alveolar ridge and presence of a tooth or several teeth regarded as non-restorable with the intention of immediate implant placement were included in the study. The number of implants in each of the study groups was not described. Multiple implant systems were used; however, 65% of implants placed were Biomet 3i. A higher level of postoperative pain following implant surgery was reported in patients with implant failure after 2 days and after 7 days. Study limitations: There was some proportion of smoker patients in antibiotic and placebo groups. The sample size was low, with only patients over the age of 40 years; therefore, the results may not be generalisable to the global population. In this study, factors related to osseointegration of the implants, such as the site of the implant and previous bone augmentation, were not described.
Tan et al., 2014 [27] The study demonstrated that the administration of pre-, perior postsurgical prophylactic antibiotics did not provide beneficial effects on patient-reported outcome measures (PROMs), such as pain, swelling, bruising and bleeding, and also did not influence clinical parameters of postsurgical complications, such as flap closures, suppuration, swelling, implant stability and pain.
The study examined three antibiotic regimens and one control group using randomisation tables allocating the patient a number with a corresponding envelope. In addition, all patients were treated periodontally for one minute prior to the implant installation. The pre-operative dosage of amoxicillin applied was based on the antibiotic dosage recommendation by the AHA for the prevention of infective endocarditis. The patients were not blinded to the allocated treatment, with panned single-tooth edentulous space in the maxilla or mandible placement with adequate pristine bone for a standard oral implant placement included in this study. The reason for the implant was a single-tooth edentulous space in the maxilla or mandible with adequate pristine bone for a standard oral implant placement without the need of simultaneous bone augmentation. The implant surgery was performed by oral and maxillofacial surgeons. Only one oral implant system (Straumann SLA) was used. The oral implants were either Standard Plus or Bone Level implants and placed into pristine bone, without any simultaneous bone augmentation. Moreover, a one-stage implant installation protocol was employed. Study limitations: Light smokers who consumed fewer than 20 cigarette a day were included in the study. The lack of patient blinding, as well as descriptions of the results for postoperative complications and missing the number of implant failures, resulted in detecting reporting bias.
Kashani et al., 2019 [34] The administration of single dose of prophylactic antibiotics in conjunction with implant placement surgery had a statistically significant lower early implant failure rate based on the early implant rate on both implant and patient levels in healthy individuals.
No preoperative oral hygiene protocol was applied in this study. Both patients and the surgeon remained blinded in the present trial. Patients were healthy and were divided into subgroups to identify potential confounding variables. Regarding the implant system, four implant brands were used: 611 implants from Nobel Biocare, 236 implants from Astra Tech, 112 implants from Straumann and implants from Ankylos. The jaw type, bone graft and number of implants placed per patient were sufficiently described. Study limitations: Details of the enrolment period and surgical protocol were not described.

Citation Main Results Appraisal Summary
Payer et al., 2020 [28] Systemic antibiotics in comparison with a placebo group were seen to provide no improvement in patient-reported outcome measures (PROMs) and no significant reduction in post-surgical complications in healthy patients undergoing advanced oral implant surgery and simultaneous GBR.
Only periodontally and medically healthy patients with a score of 1 or 2 according to the physical status classification of the American Society of Anaesthesiologists (ASA) with defects allowing for simultaneous fixture implant placement and GBR were included. The examiners and patients remained blinded. One implant system of the Straumann Standard Plus or Straumann Bone level was applied. The number of implants inserted in each group was not described. Up to four surgeons were included to perform the clinical procedure in each centre. Payer et al. (2020) were contacted to supply the missing information through e-mail inquiries; however, the corresponding author did not respond. Study limitations: The planned sample size could not be achieved, and it was supported by a grant from the international team for implantology (ITI Foundation).
Durand et al., 2021 [29] A single preoperative dose of 2 gr of amoxicillin one hour prior to implant placement was seen to be more beneficial than an additional postoperative antibiotic regimen for the prevention of implant complications after placing straightforward platform-switched implants for uncomplicated implant surgery in healthy patients, and the addition of amoxicillin 500 mg (three times daily) 7-day postoperatively was of no advantage or benefit on peri-implant crestal bone remodelling, post-operative morbidities and implant survival after placing straightforward platform-switched implants in healthy patients after a 1-year follow-up.
Only periodontally healthy patients or those presenting with gingivitis with adequate oral hygiene undergoing straightforward platform-switched implant placement from multiple ethnic backgrounds were included in the study. The groups were homogenous in age, sex, ethnicity, education and, smoking status. In addition, the groups were similar in surgical parameters (insertion torque, incision length, bone quality, implant location [maxilla vs. mandible]) and implant system, except two parameters: unequal distribution of implant insertion in each group, despite using random allocation, and the mean surgery duration was significantly longer in the control group. All implant surgeries were conducted in one treatment centre. The surgeons, participants and examiners were all unaware of subject allocation throughout the study. The measured outcomes were mesial and distal peri-implant crestal bone levels, postoperative pain severity and postoperative morbidity. The author was contacted to provide the missing information through e-mail inquiries; however, the corresponding authors did not respond. Study limitations: Since the sample size was low and study population was healthy, the results of this study prevent any generalisation for global populations, with comorbidities, smokers, more complex surgeries involving additional bone grafting procedures and when implant surgeries were executed by inexperienced surgeons. The duration of the enrolment period was not described. There was no clear explanation of ethical approval in this study. The decision whether to prescribe perioperative antibiotics was left to the surgeon, and this lack of randomisation could contribute to selection bias. In addition, adverse events were self-reported at the follow-up appointment and not collected on a daily basis, possibly introducing a recall bias. The sample size could not achieve statistical power. Systemic amoxicillin may have a modest effect on clinical parameters during the first postoperative week and may have a limited effect on biomarkers.
Only systemically and periodontally healthy individuals or those presenting with mild gingivitis for a single tooth implant were included in this study. The two implant systems from "Astra Tech" and "Zimmer Dental Implant" were used in this study. These two implant systems have different thread pattern designs and require different osteotomy procedures, and both systems require implant placement flush with the alveolar crest. All implant surgeries were conducted in one treatment centre. Screw-type, root-form, two-piece dental implants were inserted. The majority of implants (81%) were located in the posterior sextant, and the distribution between the maxilla and mandible was similar. The short-term evaluation of one week was chosen to determine early events occurring at the soft tissue level within the time limitation of antibiotic prescription. Study limitations: This study lacked relevant outcomes measured in this review. No specification given to who performed the surgery.
Esposito et al., 2008 [36] 2 g of preoperative amoxicillin versus 2 g of placebo one hour prior to the placement of dental implants resulted in no statistically significant differences in implant failures and postoperative complications in patients not requiring bone augmentation procedures.
Randomisation was performed using sequentially numbered, identical, opaque, sealed envelopes for concealment of allocation through the application of computer-generated restricted randomisation lists. Patients were grouped into three groups, non-smokers, light smokers (up to 10 cigarettes per day) and heavy smokers (more than 10 cigarettes per day). Dentists with extensive experience in implant treatment performed the implant surgery in each centre. Various implant systems were used. The majority of implants were manufactured by "Zimmer". There was no apparent relevant baseline imbalance between the two groups in terms of patient characteristics (gender, age, non-smoker, duration of intervention in minutes, total number of inserted implants, patients who took postoperative antibiotics and incidence of intraoperative complications). Patients and investigators remained blinded for the entire duration of the trial. Study limitations: There was no clear explanation of ethical approval. The sample size was underpowered to detect a statistically significant difference. This study was not sponsored; however, the placebo and antibiotic used in this study were donated by a drug company manufacturing generic drugs.
Laskin et al., 2000 [37] The study showed that the use of preoperative antibiotics significantly increased dental implant survival according to the incision type, mobility at implant placement and implant type.
The study did not compromise patients with mild and severe systematic disease. One implant system of the Spectra system was used in this study. Oral hygiene was prescribed for two weeks post-operatively. Patients were followed up from the time of placement to 36 months. Study limitations: The study did not define the inclusion and exclusion criteria and has been funded by USA Government-supported research. It was not clear who performed the implant surgery. The number of participants for whom implant failure was measured in each intervention group was missing and the method was unclear, with respect to the description of the randomisation procedure, method of sequence generation and method of allocation concealment, with inadequate reports of outcomes, causing performance bias, reporting bias and attrition bias.
Only four studies evaluated the number of supplemental analgesics between the intervention and control groups over the post-operative time points. The number of analgesics consumed in patients who were not given AB was statistically significant in the study by Nolan et al. [33]. However, the number of supplemental analgesics taken by the participants in each group was not significant in Tan et al. [27], Payer et al. [28] and Durand et al. [29]. The remaining studies did not evaluate the pain medication use [25,26,31,32,[34][35][36][37].
Five studies reported that surgery was performed by oral and maxillofacial surgeons [25][26][27][28]34]. Implantologists performed the procedures in Durand et al. [29] and odontologists in the Abu Ta'a et al. [32] study. Dentists were reported in two other studies [31,36]. However, in the Nolan et al. [33] study, the procedure was performed by two postgraduate students supervised by an experienced periodontist. Laskin et al. [37] and Khoury et al. [35] did not report on who performed the procedure. Five studies implemented a one-stage implant installation protocol without the need for a second-stage procedure to expose the implant [27,29,31,32,35]. Six studies implemented two-stage surgery [25,26,28,33,36,37]. The study by Kashani et al. [34] employed both one-and two-stage procedures.
The trial by Payer et al. [28] was supported by a grant from the Foundation of International Team for Implantology (ITI Foundation). Funding for the trial by Anitua et al. [26] was provided by the Biotechnology Institute, Vitoria, Spain. The study by Laskin et al. [37] was USA government-supported research, and there were no restrictions on study design. The studies of Esposito et al. [36] and Esposito et al. [31] did not declare a funding source; however, the placebo and ABs used were donated by a drug company manufacturing generic preparations. These companies were not involved in the design of the study, in the data evaluation or in commenting on the manuscript. Seven studies did not report external funding [25,27,29,[32][33][34][35].
No sample size calculation was reported in four studies [32,33,35,37]. Six described their sample size calculation but stated that their planned sample was not achieved leaving their studies insufficiently powered to confirm significance [25,26,28,29,31,36]. Two studies did not indicate whether their calculated sample was achieved [27,34]. All studies reported on the AB type and regimens. No study reported on the agent used as a placebo.
Six studies assessed implant failure as a main outcome in their trial. Each study reporting different definitions of implant failure. Only Nolan et al. [33] did not define "implant failure" in their study.
Caiazzo et al. [25]: implant failure is mechanical implant removal due to the lack of osteointegration.
Esposito et al. [31]: implant failure is an implant mobility measured manually and/or any infection dictating implant removal. The study also reported the number of patients with implant-supported prostheses, and the number of events for prosthetic failure was added to the total number of implant failure events. Prostheses failure was defined as prostheses that could not be placed or prosthesis failure (if secondary to implant failures).
Abu Ta'a et al. [32]: defined a failed implant as the presence of signs of infection and/or radiographic peri-implant radiolucencies that could not respond to a course of ABs and/or judged a failure after performing an explorative flap surgery by an experienced periodontologist.
Kashani et al. [34]: defined implant failure as the removal of an implant for any reason, from implant placement to abutment connection or prosthetic treatment.
Esposito et al. [36]: defined implant failure as implant mobility of each implant, measured manually, and/or any infection dictating implant removal. The study also reported the number of patients with implant-supported prostheses, and the number of events for prosthetic failure was added to the total number of implant failure events. The prostheses failure was defined as prostheses that could not be placed or prosthesis failure if secondary to implant failures.
Four studies did not define postoperative complications [25,34,35,37], whilst eight evaluated the postoperative complications as a main outcome. However, in three studies, the number of events in each group was not reported and only the p value was stated [27,28,33].
Anitua et al. [26]: the diagnosis of the presence of postoperative infections or postimplant infection been carried out using clinical criteria, which include inflammation, pain, heat, fever and discharge.
Abu Ta'a et al. [32]: postoperative infection was defined as the presence of purulent drainage (pus) or fistula in the operated region, together with pain or tenderness, localised swelling, redness and heat or fever of more than 38 • C.
Nolan et al. [33]: different variables used to identify post-operative infection and postoperative mortality were defined as swelling, bruising, wound dehiscence and suppuration.

Quality Assessment of the Selected Studies
The quality assessment tool of the Effective Public Health Practice Project (EPHPP, McMaster University 1280 Main Street West, Hamilton, ON, Canada, L8S 4K1) [38] was used to examine each study against six components. The study quality was assessed independently by two reviewers. Table 3 shows the global quality rating. All studies were classed as moderate-strong.  (2) global mean score were deemed suitable for the systematic review. Selection bias: strong = participants very likely to be representative of the target population of greater 80% participation/moderate = participants very likely to be representative of the target population of 60-79% participation/weak = participants very likely to be representative of the target population of less than 60% participation. Study design: strong = randomised controlled trials (RCTs) or controlled clinical trials (CCTs)/moderate = cohort analytic study, a case control study, a cohort design or an interrupted time series/weak = any other method or did not state the method used. Confounders: strong = controlled for at least 80% of relevant confounders/moderate= controlled for 60-79% of relevant confounders/weak = controlled for less than 60% of relevant confounders. Blinding: strong = outcome assessor not aware of intervention status of participants AND participants not aware of research question/moderate = outcome assessor is not aware of the intervention status of participants OR the study participants are not aware of the research question/weak = outcome assessor AND participants both aware of the above or blinding is not described. Data collection methods: strong = data collection tools shown to be valid AND reliable/moderate= tools are valid, but reliability not described/weak = data collection tools not shown to be valid. Withdrawals and dropouts: strong = follow-up rate more than 80% or greater/moderate = follow-up rate of 60-79% of participants/weak = follow-up rate of less than 60% participants or withdrawal and dropouts were not described.

Implant Failure Not Prevented by the Use of ABs, by Number of Implants
Eight studies evaluated the efficacy of AB treatment. Eight studies analysed the number of implants failed. Out of 1930 implants placed in the AB group, 23 failures were reported against 68 failures from 1566 implants inserted in the placebo group.
As the heterogeneity estimate was low (I 2 = 0%), the fixed effect model was used. The pooled estimate RR of 0.31 (95% CI, 0.20-0.48) was significant in the fixed effect model (p < 0.001) (Figure 3).

Implant Failure Not Prevented by the Use of ABs, by Patient
Six studies assessed implant failure by the number of patients in their trial, each reporting different definitions of implant failure outcomes. Only the trial conducted by Nolan et al. [33] did not define the implant failure outcome in their study. The fixed effect model was used because I 2 < 25%, and the pooled estimate RR of 0.341 (95% CI, 0.22-0.52) was significant (p < 0.001) (Figure 4).

Postoperative Complications Analysis
Eight clinical trials provided comparative data on the efficacy of AB treatment compared with the placebo, control or no AB treatment in patients undergoing implant placement based on postoperative complications. However, three studies did not report on the number of events in each group, with only the p value being stated [27,28,33]. Postoperative complications were reported in only five studies. The overall results show that there was no statistical significance difference for postoperative complications between the groups (p = 0.89). The fixed effect model was used because I 2 < 25% giving a pooled estimate RR of 1.01 (95% CI, 0.61-1.70), with p = 0.96-not significant ( Figure 5).

Postoperative Antibiotic Adverse Events Analysis
Five studies reported data on the efficacy of AB treatment compared with placebo, control or no AB treatment based on total adverse events. Three studies reported zero adverse events related to the use of ABs [26,31,32]. Three did not define the postoperative AB adverse events [26,29,32]. In three studies [26,31,32], AB adverse events were appropriately assessed, but no events occurred in both groups ( Figure 6). The 95% CI (0.08-5.04) is very wide, indicating a lack of confidence that AB use would affect adverse events (p = 0.63). Figure 6. Forest plot of comparison between AB and placebo, which gives the summary estimate (centre of diamond) and its 95% confidence interval CI (width of diamond) based on postoperative adverse events. Statistical method: Mantel-Haenszel with fixed effect model [26,29,31,32,36].
The terms implant success or implant survival were used interchangeably, with no specific criteria or definition given in most studies. In the study by Laskin et al. [37], the implant survival status was assessed from the time of placement to 36 months according to previous implant experience of the surgeon, implant coating, bone density; patient age, race and gender, incision type, mobility of the implant at placement and use of chlorhexidine; and health status. When implant survival rates were compared according to the patient age, it was observed that patients in all age groups who were provided preoperative AB treatment reported higher implant survival, except in the under 30 group. When implant survival rates were compared according to surgeon's' previous implant surgery experience, those surgeons with greater than 50 implant placements prior to the study had a slightly higher implant survival rate of 2.9% when preoperative ABs were used. In addition, a less experienced surgeon (<50 previous implant placements) had an even greater increase in the survival rate of 7.3% when preoperative ABs were used. When implant survival rates were compared according to the use of chlorhexidine in conjunction with preoperative ABs, its use improved implant survival by only 0.9% when preoperative ABs were provided, whereas without preoperative AB coverage, implant survival improved by 5.8%. Comparing both preoperative ABs and chlorhexidine to not using either, there was a substantial increase of 7.8% in implant survival [25][26][27][31][32][33][34]37]. Five studies did not specify the success rate [28,29,35,36].
The clinical diagnosis of an infected surgical wound was specified by Caiazzo et al. [25] as internal and external oedema and internal and external erythema, pain, heat and exudate. In Anitua et al. [26], the presence of infection was recorded at 3 days, 10 days, 1 month and 3 months after the implant placement. Six patients in AB and six patients in placebo groups experienced post-operative infections (not significant). Abu-Ta'a et al. [32] reported that patients self-reported infection using a mirror. One patient from the AB group and four patients in the NOAB groups developed post-operative infections, but significance was not reported. In Nolan et al. [33], post-operative morbidity (swelling, bruising, wound dehiscence and suppuration) was recorded on days 2 and 7 by the same examiner using Boolean variables: post-operative swelling and grading its severity, recorded by two independent examiners, where 0 was no swelling, 1 means mild swelling, 2 means moderate swelling and 3 means severe swelling. Post-operative bruising was significantly higher in the placebo group after 2 days, but not by day 7 post-operatively (p = 0.99). Two patients presented with suppuration, both of whom received a placebo preoperatively (not significant p = 0.49).
Esposito et al. [36] reported complications, such as wound dehiscence, suppuration, fistula, abscess and osteomyelitis, which were measured after one week, two weeks and four months after implant placement, in which they found 11 patients in the AB group had postoperative complications vs. 13 patients in placebo group. Two flap dehiscence events occurred in the placebo group at one week, and four events were observed in the AB group (not significant p = 0.684). Two weeks after implant placement, one peri-implant mucositis was reported in the placebo group and one flap dehiscence in the AB group (not significant p = 0.1). Four months after implant placement, two cases were reported in the placebo group and one mobile implant (provoking severe pain) occurred in the AB group (not significant p = 0.623). Post-surgical complications reported by Tan et al. [27] include flap closure, pain, swelling, suppuration, and implant stability assessed by trained standardised examiners. Postsurgical complications were reported at one week, two weeks and one month after implant installation. Concerning the flap closure, 5% of patients in the placebo group did not achieve complete wound closure compared to 0% for the three other groups. There was no significant difference reported in all other follow-up points by all groups. Of all patients, 17% experienced postoperative pain at one week following implant placement, indicating no statistically significant difference among all AB groups (p > 0.05). The presence of swelling was reported in 21.4% of all subjects at week one (27% in ABs group and 17.5% in placebo). This resulted in no statistically significant difference among all groups at any time (p > 0.05). Payer et al. [28] performed clinical measurements at 1, 2, 4 and 12 weeks after surgery and found no significant differences in flap closure, pain, swelling, pus and implant stability of the operation site between the two treatment groups at any time over the post-operative observation period. The clinical diagnosis of postoperative surgery-associated morbidities was evaluated by Durand et al. [29] after 1, 3 and 16 weeks and 1 year. Postoperative swelling was measured using a form graded as follows: 0, no swelling; 1, mild swelling; 2, moderate swelling; and 3, severe swelling.
Bruising, suppuration and wound dehiscence were evaluated dichotomously. The authors were contacted for the missing information about swelling and bruising in intervention and placebo groups, and they provided all data requested. Regarding suppuration, two participants in the AB group had suppuration at the one-week examination and were asked to rinse with 0.12% chlorhexidine gluconate twice daily for two weeks but no ABs were used. No more suppuration was reported at the three-week follow-up. There were no significant differences in postoperative morbidities between the intervention and placebo groups (p = 0.230).
In the study by Anitua et al. [26], patient records of smoking habits showed that 10 patients in the AB and 8 patients in the placebo group were smokers. However, the smoking habits variable was not associated with a higher risk of postoperative infection. Fifty percent of studies did not report on complications, and the other 50% reported no significant statistical differences between groups for all types of measured complications (swelling, bruising, wound dehiscence, pain and suppuration) [25][26][27][28][29][33][34][35][36][37].
No AB-related adverse events were reported in three studies [26,31,32]. In the study by Durand et al. [29], no definition of adverse events was given. Only one participant in the placebo group reported adverse events (diarrhoea) two days after implant placement. In the study by Esposito et al. (2008), one week after implant placement, one adverse event occurred in the placebo group (itching for one day) and one in the AB group (diarrhoea and somnolence); however, no significant difference was reported (p = 0.1), and no major complications linked to the use of ABs occurred. Seven studies did not report on AB-related adverse events [25,27,28,[33][34][35]37].
In the study by Payer et al. [28], the patient self-reported outcomes used a visual analogue scale (VAS) (0-10) every day for the first two postoperative weeks. Results from mean and median VAS scores of bleedings, swelling and pain revealed no significant difference at day 14 (p > 0.05), but the mean VAS scores decreased over time (p < 0.001). The patients of Tan et al. [27] self-reported outcomes at days 1, 7 and 14. There was no statistically significant difference among study AB groups for bleeding, swelling, pain and bruising, indicating no superiority for the prophylactic regimen (p > 0.05). All other studies did not report on this outcome [25,26,29,[31][32][33][34][35][36][37].
Nolan's [33] patients kept records of the number of paracetamol 500 mg tablets taken for one week, and a VAS question was used for reporting. Higher postoperative VAS scores were recorded by the placebo group after two days postoperatively, and the difference between groups was significant (p = 0.003 at first day and p < 0.001 after 7 days). Durand's [29] patients were given daily diaries to evaluate their postoperative pain severity for 1 week postoperatively, using self-administered questionnaires (VAS score). The median VAS score was 2 in the control group and 0 in the AB group. The overall median pain severity observed in both groups during the first seven days after surgery was considered mild. However, the perceived pain intensity difference in participants taking the postoperative placebo was statistically significant on the fourth day at noon (p = 0.047) and at night (p = 0.036) and on the 5th day at night (p = 0.036), indicating that patients in the postoperative placebo group experienced significantly more pain. However, it is worth noting that their control group experienced a longer implant surgery duration and higher number of implants inserted compared to those in their AB group. Durand et al. [29] reported that the mean number of supplemental analgesics taken by the participants in each group was 1.5 ± 4.5 tablets, and for those in the placebo group, it was 1.0 ± 2.8 tablets at the 1-week follow-up, which was not significant.
Tan et al. [27] reported no significant difference in analgesic consumption groups. Paracetamol three times a day for two days after surgery was recommended by Payer et al. [28]. The percentage of patients who took analgesics did not differ significantly (p > 0.05) between the two treatment groups. The other studies did not report on this outcome [25,26,31,32,[34][35][36][37].
The postoperative use of rescue ABs was reported by Kashani et al. [34]. Seven patients with 10 implants required postoperative AB treatment. Of 10 infected implants, seven healed after 10 days of ABs, whereas three implants were lost and subsequently reoperated.

Number Needed to Treat Calculation
To Prevent Implant Failure Due to Infection, by Implant NNT: 32. This is >5, which indicates that the intervention (use of prophylactic ABs) is not effective in preventing infection [22,23].
To Prevent Implant Failure Due to Infection by Patient NNT: 14. The number of patients needed to treat (NNT) with antimicrobial prophylaxis to prevent one implant failure due to infection in previous studies was indicated as from 25 to 48 [40,41]. This review found that 14 patients undergoing DIP surgery need to receive ABs in order to prevent one implant failure due to infection occurring. The NNT is more than 5, which indicates that the intervention (use of prophylactic ABs) is not effective in preventing infection [22,23].
To Prevent Complications Due to Infection NNT: 2523. The NNT is more than 5, which indicates that the intervention (use of prophylactic ABs) is not effective in preventing complications caused by infection [22,23].
To Cause Adverse Events NNH: 528. The NNH is not a negative figure but more than 5, which means that the possible harm caused by the use of ABs exists, but it is very small and does not negate AB use when required [22,23].

Discussion
The major limitation of this systematic review is the small number of well-designed randomised controlled trials on the efficacy of ABs for DIP surgery identified. This may in part be due to this unfunded study not having the resources for wider searching or obtaining translations of non-English language publications. However, the RCTs that were included covered a range of countries and different healthcare systems and thus reflect a wide range of practices. The trials were well documented RCTs rated with a low RoB. This review provided evidence broadly supporting the administration of prophylactic ABs in those undergoing implant surgeries. When implant failure due to infection was analysed by implants, it was noted that the implants placed in the AB group were 67% more likely to survive than those when receiving placebo, indicating that implants placed in the placebo group were three times more likely to fail. In the studies, 3496 implants were placed in both groups; 23 failed in the AB groups and 68 in the placebo group. The effectiveness analysis of implant failure by implant suggests an antibiotic NNT of 32 to prevent one implant loss from occurring. This broadly agrees with the recommendations of the fourth European Association for Osseointegration Consensus 2015 [42].
When implant failure due to infection was analysed by patients, statistically beneficial results were found for the administration of ABs compared to those with controls. However, for the postoperative complications, no evidence of AB use benefit was found. Differences in clinical efficacy of Abs based on postoperative complications did not demonstrate significance due to the variability and small sample sizes of included studies. The most suggested AB was amoxicillin 2 g, but there is insufficient evidence to recommend this specific dosage. The number of adverse events occurring in those receiving ABs was not significantly different from that in those who received no ABs, confirming that ABs are safe to use when indicated. From 12 included trials, 6 reported a significant difference between ABs and placebo [29,31,[33][34][35]37]. In contrast, the remaining studies showed no significant benefit [25][26][27][28]32,36]. Nolan et al. [33] reported that the number of implants placed appeared to significantly affect the osseointegration of DIPs and that when more implants were placed, the surgery duration was longer. A different implant system was used in all trials, and no statistically significant differences were reported for the success or failure of osseointegration. Most studies implemented a variety of single and combined additional therapies with anti-inflammatory and pain-relief medications postoperatively to manage postoperative pain following DIP surgery. The Esposito et al. [31] study found a correlation between immediate postextractive implants and the number of failed implants and concluded that patients who received immediate post-extractive implants were more likely to fail compared with patients receiving delayed implants. The study of Nolan et al. [33] reported that longer procedures showed a modest correlation with post-operative pain (on VAS scores), interference of daily activities and the number of analgesics used. Only one study found smoking to have affected the outcome results, and heavy smoking was associated with an increased risk of implant failure. This does not rule out moderate smoking as a risk factor for failed implants. Abu Ta'a et al. [32] reported on parafunctions, finding they would increase the risk of implant failure. In this review, only one study addressed immediate post-extractive implants as a potential risk of failed implants. In most of the studies, the implant surgery was performed by oral and maxillofacial surgeons for placement of the implant. In one study by Nolan et al. [33], the implant surgical procedure was performed by two postgraduate students supervised by an experienced periodontist; however, the results of the study appear to be unaffected. Regarding the surgical protocol of one-stage and two-stage surgeries, none of the studies found significant factors associated with oneor two-stage surgeries with the failure or success of an implant. It is worth noting that emerging technologies, such as laser, photo-disinfection and airflow, etc., will contribute to infection risk reduction [new]. All studies implemented additional therapies with chlorhexidine oral rinse pre-and postoperatively or administered analgesic before and after implant surgery, which could have affected the results of the review. The studies also reported maintaining a high standard of infection control in surgical procedures. The results from this review were compared to previous studies (Table 4). Table 4. Comparison of findings with previous systematic reviews.

Citation Comparison
Canullo et al., 2020 [43] This study found a statistically Ata-Ali, Ata-Ali and Ata-Ali, 2014 [40] The study by Ata-Ali et al., (2014) included only four, large-sample size double-blinded placebo-controlled trails with a total of 2063 implants (including 1077 in the antibiotic treatment group and 986 in the control group). The results affirmed that antibiotic treatment provides a statistically significant beneficial reduction in implant failure (p = 0.003) with an NNT of 48, and it was concluded that antibiotic administration lowers the odds of implant failure by 66.9%. In contrast, failing to reach statistical significance was a reduction in the incidence of postoperative infection (p = 0.754). The current meta-analysis is consistent with previously published meta-analyses but included a larger number of studies regardless the sample size of studies (eight double blinded placebo-controlled trails, including a combined total of 2086 implants, including 1930 implants in the antibiotic versus 1566 in the placebo group) and revealed a statistically significant reduction in implant failure (67%). This meta-analysis used risk ratio as the measure of effect, while other meta-analyses in the literature have used the odds ratio (OR) as the primary measure. To add to the limitations, the study included RCTs assessing the effect of amoxicillin as the antibiotic treatment, whereas this meta-analysis included any type of antibiotic.

Citation Comparison
Jain, Rai, Singh and Taneja, 2020 [44] The metanalysis implemented by Jain et al. (2020) included five RCTs rated as having a low risk of bias assessing the use of preoperative amoxicillin 1 h prior to surgery with a total of 1032 patients (514 in antibiotic and 518 in placebo groups) and 1919 implants (952 in the antibiotic group and 967 in the placebo group) for the prevention of both the number of implant failures and the number of patients with implant failures after a minimum of 3 months, from implant placement in healthy individuals, and concluded that the number of implants failed was statistically fewer in the antibiotic group, p = 0.006, with 0% heterogeneity. The number of patients that suffered from implant failure was significantly lower in the antibiotic group, p = 0.006 with 0% heterogeneity. This study is consistent with the current finding; however, our study included a larger number of randomised controlled trials including any antibiotic regardless of the timing of administration and revealed statistically significant results.
Braun, Chambrone and Khouly, 2019 [45] The systematic review and metanalysis performed by Braun et al., (2019) included six studies with a combined total of 1162 participants and a total combined number of implants placed of 2250, assessing the efficacy of antibiotic use in the prevention of both the number of implant failures and the number of patients with implant failures after a minimum of 3 months after dental implant placement in otherwise healthy individuals receiving dental implants and suggested that the number of implants performed with the use of antibiotic had a significant effect on the survival rate (p = 0.005; RR, 0.35; 95% CI, 0.16 to 0.72), with an NNT of 43 for the number of implants, and the number of patients with implant failure was statistically lower in the antibiotic groups (p = 0.002; RR, 0.33; 95% CI, 0.16 to 0.67), with an NNT of 24 for the number of patients.
The current meta-analysis is consistent but included a higher number of total participants and found a smaller NNT for both implant failure analysis by the number of implants and by patient numbers: NNT, 32 and NNT, 14. The research conducted by Braun et al. (2019) assessed antibiotic-associated adverse events and prosthetic failure as a secondary outcome. The results showed no statistically significant difference for any of the other outcome adverse events (RR, 1.00, 95% CI, 0.06 to 15.85) and prosthetic failure (p = 0.08; RR, 0.43; 95% CI, 0.17 to 1.11). The conclusion for the antibiotic-associated adverse events was based on the inclusion of three RCTs, while the current study included five RCTs with three studies reporting no adverse events in either group. Braun included two studies on prosthetic failure in the meta-analysis, while the current meta-analysis combined the total number of prostheses failed due to implant loss and total number of implant failures together.

Limitations
The limitation of this systematic review is the small number of well-designed randomised controlled trials on the efficacy of ABs for DIP surgery available for review to date.
In addition, due to the small number of studies included, the funnel plot was not used. Thus, it was not possible to objectively evaluate the publication bias in these meta-analyses. All RCTs that were included in these meta-analyses were conducted in various countries with different healthcare systems. Therefore, drawing definitive conclusions about the effects of ABs is a problem.
The study search was completed in January 2022; it is planned that a new systematic review will be conducted to cover studies published in 2022 and 2023.

Conclusions
Antibiotic use was reported to be statistically significant in preventing infection (p < 001), but not significant in the prevention of complications (p = 0.96), and the NNT was larger than 5 (14 and 2523 respectively), which indicates that the intervention is not sufficiently effective to justify its routine use. The occurrence of side effects was not significant (p = 0.63), and the NNH was 528, indicating that the possible harm is very small and does not negate the AB use when required. Given that the studies only evaluated healthy men and women, the study can be applied to systemically and periodontally healthy individual undergoing straightforward oral implant surgery placement under sterile conditions. It is important that future studies are rigorously designed and adequately powered to better inform future practice around the secondary outcomes and for complex patients with co-morbidities.