Does the Prosthetic Phase of Dental Implants Justify the Prescription of Preventive Antibiotics in Healthy Patients? A Systematic Review Open Access

Antibiotic therapy in Oral Implantology can be classified as “preventive” when it is used to avoid getting an infection, or “therapeutic” when it is used as a treatment against an infection already established.¹  Preventive antibiotic therapy (PAT) has been accepted as a strategy to avoid systemic bacteremia, but also to achieve an adequate antibiotic concentration in the blood to prevent bacterial contamination during surgery, as the oral cavity per se is a septic cavity.²  In addition, many professionals are using the prescription of PA as a defensive clinical practice that increases their feeling of safety in the face of a possible demand.³  However, it is still unclear whether the administration of PA in implant surgery procedures is necessary, as there are not enough studies with sufficient methodological quality to support it.¹  In this regard, in the last 2 years, clear recommendations have been established regarding the prescription of PA in healthy patients for the insertion of dental implants without anatomical conditions⁴  or in bone augmentations with single or 2-stage implant insertion to avoid bacterial contamination of the implant and/or the grafted bone particles.⁵  In these cases, it is recommended to prescribe 2 or 3 g of amoxicillin 1 hour before surgery.^4,5  However, there are still several procedures in which guidelines are lacking, such as sinus lifts, the insertion of immediate implants or the implant prosthetic phase. The latter includes second phases, that is, uncovering of implants for the placement of a prosthetic attachment around which the peri-implant mucosa will heal, taking impressions, and placing prostheses on implants.

Several surveys have been published in the last years aimed at studying the prescription patterns of PA in implant treatment. However, few authors have evaluated the administration of these drugs during the implant prosthetic phase.^6–8  In this respect, 6.6%⁸  to 11.1%⁶  of Jordanian and Spanish dentists prescribe PA in second phases, and up to 7% and 7.6% when taking impressions and placing the prostheses on implants,⁶  respectively. These differences highlight the lack of consensus that exists among professionals. Moreover, the urge to further evaluate the use of PAT stands as evident when these findings are placed in the current context of bacterial resistance to antimicrobials. Antibiotic resistance has become a serious threat to public health, as an increasing number of infections are becoming more and more difficult to treat due to a loss of efficacy of these drugs. On the other hand, these drugs cause undesirable effects, such as allergic reactions, which can even compromise the patient's life.⁹  In this sense, it is estimated that for every million patients treated with a single dose of antibiotic, mild, moderate, and severe allergic reactions occur in 2400, 400, and 0.9 patients, respectively.¹⁰  Other problems related to the use of antibiotics include direct toxicity, such as gastrointestinal disorders (nausea, vomiting, diarrhea, and abdominal pain), hematological problems (neutropenia, thrombocytopenia, and hemolysis), nephrotoxicity (proteinuria or renal failure), neuropathies (nerve dysfunction or peripheral neuropathy), hepatobiliary dysfunctions (jaundice or hepatitis), and interactions with other drugs with which the patient is being treated. In addition, antibiotics do not distinguish between beneficial bacteria and the bacteria they are intended to fight, which can lead to opportunistic infections or pseudomembranous colitis.¹¹ 

Therefore, the present study aims to conduct a systematic review to determine whether the prescription of PA is needed in the implant prosthetic phase and if so, to establish recommendations for its use.

This systematic review was structured according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocols (PRISMA) statement.¹²  The protocol was registered at the International Prospective Register of Systematic Reviews under the registration number CRD42021277959.

The study aimed to answer the following PICO (population, intervention, comparison, and outcome) question based on the PRISMA guidelines (Table 1):

In healthy patients starting the implant prosthetic phase (P), does the prescription of PA (I) compared with not prescribing PA (C) decrease the incidence of infectious complications (O)?

The term “implant prosthetic phase” also includes: (1) the performance of implant exposure surgery for the connection of a prosthetic attachment around which the peri-implant mucosa will heal, (2) impression-taking, and (3) placement of the prosthesis on implants.

From the invention of dental implants to the present day, millions of dental implants have been placed that have required and still require the completion of corresponding prosthetic stages. Recent surveys have shown that a significant percentage of professionals prescribe PA for these procedures to reduce the possible appearance of infections or biological complications.^6–8 

Before starting, exclusion and inclusion criteria were defined for the resulting articles.

Included studies were the following: (1) human studies, (2) meta-analyses, (3) systematic reviews, (4) randomized clinical trials (RCTs), (5) cohort studies, (6) observational studies, (7) comparative studies, (8) multicentric studies, (9) consensus statements, (10) reviews, and (11) guidelines. The procedures considered within the papers were the following: the performance of implant exposure surgery for the connection of a prosthetic attachment around which the peri-implant mucosa will heal, impression-taking, and placement of the prosthesis on implants. Any type of antibiotic was considered for analysis in these procedures. No restrictions were applied by the language of publication.

The exclusion criteria determined the exclusion of the following: (1) experimental laboratory studies, (2) animal studies, (3) studies whose main topic was not the analysis of infection complications in the implant prosthetic phase, (4) duplicated articles, (5) books or book chapters, (6) letters to the Editor, (7) comments, (8) case reports, and (9) unpublished articles.

A comprehensive search of the literature was conducted in the following databases: PubMed/MEDLINE, Web of Science, Google Scholar, and LILACS. A search for unpublished studies (grey literature) was conducted on the OpenGrey database. Besides, we examined the bibliographic references of the selected articles for publications that did not appear in the initial search and might be of interest.

The search was performed by 2 independent researchers (A.-O.S.-P. and J.-F.P.-C.). The search was not time-restricted and was updated to September 2021. MeSH (Medical Subjects Headings) terms, keywords, and other free terms were used with Boolean operators (“OR”, “NOT”, “AND”) to combine searches (Table 2). The same keywords were used for all search platforms and followed the syntax rules of each database.

Two researchers (A.-O.S.-P. and J.-F.P.-C.) independently compared results to ensure completeness and removed duplicates. Then, the full title and abstracts of the remaining papers were screened individually. Finally, full-text articles to be included in this systematic review were selected according to the criteria described above. Disagreements over which eligible studies were to be included were discussed with a third reviewer (J.-L.G.R.), and a consensus was reached. The reference list of the included studies was also reviewed for possible inclusion. Agreement between reviewers was measured with the κ coefficient. The results were also expressed as the concordance between reviewers (%).

Two independent reviewers (A.-O.S.-P. and J.-F.P.-C.) evaluated the methodological quality of eligible studies following Joanna Briggs Institute Critical Appraisal Tool,¹³  which incorporates 10 domains. The studies were classified as low-quality assessment studies (0–5 domains), or as high-quality assessment studies (5–10 domains).

The search strategy resulted in 331 results, of which 309 remained after removing the duplicates. Then, 2 independent researchers (A.-O.S.-P. and J.-F.P.-C) reviewed all the titles and abstracts and excluded 298 papers that were outside the scope of this review. Thus, we obtained 11 potential references. After reading the full text of those 11 papers, 9 were discarded for not being related to the prescription of PA in the placement of dental implants (n = 8) and 1 for being a survey. When analyzing the references of the reviewed articles, 1 article was included as an ancillary paper. Therefore, 3 studies were included in the systematic review^14–16  (Figure 1).

There was a 99.18% concordance between the 2 authors (A.-O.S.-P. and J.-F.P.-C.) with a κ coefficient of 0.85 (SE, 0.04 [95% CI, 0.77–0.93]) for titles and abstracts, and a 97.43% concordance with κ coefficient of 0.93 (SE, 0.08 [95% CI, 0.74–1]) for full-text articles, respectively.

None of the studies specifically compared postoperative infection rates related to whether PA was prescribed in these procedures. However, the risk of infection associated with the performance of second stages, and impression-taking was obtained in relation to the performance of these procedures.

The main findings are described in the following sections.

The search provided 3 articles related to the risk of infection and/or recommendations on the prescription of PA in second-stage implant surgeries.^14–16  The resulting articles were a retrospective clinical study,¹⁶  a consensus document,¹⁵  and a clinical protocol.¹⁴ 

The first of these, published by Powell et al¹⁶  in 2005, referred globally to the risk of infection in first- and second-stage implant surgeries as being 1.14% (2/175) and those related to the use of subepithelial connective tissue autografts (3.66%; 3/82) and epithelialized free soft tissue autografts (5.88%; 1/17). For this study, infection was defined as increasing and progressive soft tissue swelling with the presence of suppuration. Patients were not required to have had a documented fever or submandibular and/or cervical lymphadenopathy. Any unscheduled return in which the provider placed a patient on antibiotic therapy after surgery was also considered an infection. Records were reviewed for infections by 3 calibrated clinicians. Moreover, patients who received PAT as part of the surgical protocol had an infection prevalence of 2.85%, which was not significantly different from the 1.81% prevalence rate seen when PAT was not used. A limitation of the study was that it involved professionals of varying skill levels. Specifically, 82% of infections (18/22) occurred in surgeries performed by residents and general dentists, compared with only 18% (4/22) when performed by board-certified periodontal faculty members.

Subsequently, Gutiérrez et al¹⁵  (2006) published a consensus document that brought together several Spanish scientific societies (Spanish Society of Oral Surgery, Spanish Society for Oral Medicine, Spanish Society for Conservative Dentistry, Spanish Society for Dental Prosthesis, Spanish Society of Implants, Spanish Society for Periodontics, Spanish Society for Paediatric Dentistry, Spanish Society for Chemotherapy and Spanish Society for Oral and Maxillofacial Surgery). This consensus document did not refer directly to second-stage implant surgeries but to “mucosa-gingival surgery,” classifying them as a procedure with a high risk of infection, indicating the prescription of PA in healthy patients, specifically amoxicillin/clavulanic acid 2 g/125 mg orally and, in patients allergic to penicillin, clindamycin 600 mg orally, both 1 hour before the intervention.

Finally, Resnik and Misch¹⁴  (2008) described the protocols developed by the Misch International Implant Institute. These authors were the only ones to explicitly refer to second-stage implants, classifying these procedures in category 1, that is, with a low risk of bacterial contamination and surgical site infection, and therefore not recommending the prescription of PA. Chlorhexidine 0.12%, twice a day for 2 weeks, was suggested as a preoperative and postoperative agent to decrease postoperative infection risk and promote soft tissue healing.

Only 1 consensus document published by Gutiérrez et al¹⁵  (2006) was found that referred to impression-taking in Oral Surgery as a procedure with a low risk of infection; therefore, the prescription of PA would not be indicated.

No specific studies were found that assessed the risk of infection or bacteremia secondary to implant-borne prosthesis placement.

Risk of bias and study quality analyses were performed independently by 2 review authors (A.-O.S.-P. and J.-F.P.-C.). Using the predetermined 10 domains for the methodological quality assessment according to the Joanna Briggs Institute Prevalence Critical Appraisal Tool,¹³  it was determined that 2 included papers^14,15  have a low-quality assessment (0–5 domains) and 1 of them¹⁶  has a high-quality assessment (5–10 domains). Table 3 shows a more detailed description of the articles included.

The evidence related to the prescription of PA in these procedures is very limited. Specifically, only 2 authors^14,16  explicitly referred to second-stage implants. However, they did not define which procedures this designation implies, since second-stage surgeries may present a greater or lesser risk of infection depending on the complexity of the technique used. These procedures can range from making a minimal invasive linear incision in the mucosa to techniques used to increase the thickness and/or width of the keratinized mucosa by harvesting autologous soft tissue grafts or using other types of biomaterials. Gutiérrez et al¹⁵  referred to “mucosa-gingival surgeries” and could refer to second-stage surgeries involving this type of more invasive surgical management of the peri-implant mucosa and, therefore, with a higher associated risk of infection. Likewise, these surgeries could have a similar risk to periodontal and/or mucogingival surgeries performed on teeth. However, in many cases, the risk could be considerably lower, as many of these surgeries are performed on submucosal implants, where there are no prior bacterial components beyond those already present on the mucosal surface or salivary level. According to Altemeier et al,¹⁷  the wounds created in these procedures could be considered class I or “clean” wounds, as these are incisions made on healed, undrained, nontraumatic tissue, without clinical signs of inflammation and in the absence of unusual contamination, with an associated infection rate of 1%–5% in the absence of PAT, and lower than 1% after administration of these drugs. A group of experts of the 10th European Workshop on Periodontology¹⁸  (2014) concluded that systemic prescription of peri- or postoperative PAT is not indicated in periodontal plastic surgeries. This is because several studies analyzed the effect of PAT on various types of periodontal surgeries and observed no significant differences between the test and control groups. More specifically, Checchi et al¹⁹  described postoperative infection rates in the test group (tetracycline hydrochloride 250 mg, 4 times daily, or minocycline, 2 times daily, for 7 days) of 3.80%, while in the control group (without PAT) it was 4.4% (P > .05) (231 patients/498 periodontal surgeries). On the other hand, Pack and Haber²⁰  prescribed PA only to patients with diabetes or rheumatic fever (penicillin or erythromycin 250 mg orally, 4 times daily for 7 days, starting the day of surgery), observing a postoperative infection rate of 2.33%, while that of the control group (healthy patients not prescribed PA) was 0.99% (P > .05) (218 patients/927 periodontal surgeries). Other authors described postsurgical infection rates of 3.70% and 3.60% in the test and control group, respectively (P > .05) (n = 207 patients and 220 periodontal surgeries). In conclusion, the benefit of using PAT in these cases is of no clinical significance.

It is considered prudent to apply these same recommendations in peri-implant plastic surgery given the lack of specific scientific evidence and for responsible prescription of PA. In this regard, Oswal et al²¹  (2014) carried out an RCT in which they analyzed the effect of prescribing (1) 1 g of amoxicillin 1 hour preoperatively, (2) amoxicillin 500 mg/8 hours/5 days postoperatively, and (3) not prescribing PA in mucogingival and other periodontal surgeries. As no postoperative infections were recorded, they recommend not to use antibiotics in healthy patients, except for surgeries of long duration (>2-hour duration) or when biomaterials are grafted extensively.

Although Gutiérrez et al¹⁵  recommended prescribing amoxicillin/clavulanic acid 2 g/125 mg, in case of considering PA administration, and until further studies are carried out, from the present study it is recommended to prescribe 2 g of amoxicillin 1 hour before the second phase. In particular, this regimen produces a concentration of amoxicillin in venous plasma of 4.21 ± 2.12 μg/mL (range, 1.6–9.98 μg/mL) and 5 ± 2.63 μg/mL (range, 2.04–11.18 μg/mL) at the implant site, with no statistically significant differences between them.²²  Both measurements are sufficient to reach the breakpoint minimum inhibitory concentration recommended by the National Committee for Clinical Laboratory Standards for anaerobic bacteria (0.5 μg/mL).²³  The effectiveness of PAT is achieved when the minimum inhibitory concentration for a given bacterium is exceeded 2 to 4 times.²⁴  Exceeding this therapeutic range creates a window of therapeutic overdose known as the “mutant selection window,”²⁵  which modifies the susceptibility of bacteria to antibiotics and selects antibiotic-resistant mutant populations.²²  Amoxicillin has demonstrated an activity greater than 80% against sensitive strains of periodontopathogens, such as Aggregatibacter actinomycetemcomitans, Peptostreptococcus spp, Prevotella spp, Porphyromonas spp, Fusobacterium spp, and Streptococcus oralis, which are also found in implants with peri-implantitis.²⁶ 

On the other hand, it has been suggested that patients allergic to penicillin should be prescribed 600 mg of clindamycin 1 hour before the second-stage implant surgery. However, the current trend suggests using this drug with caution. In this regard, a recent study in the United Kingdom suggested that a single dose of clindamycin may cause complications, including death, from Clostridium difficile infection.²⁷  For this reason, clindamycin is no longer recommended for antibiotic prophylaxis for a dental procedure according to the American Heart Association.²⁸  Also, Salgado-Peralvo et al²⁹  (2022) related the use of clindamycin with an increased risk of implant failure, possibly because this drug may favor an increase in the proportions of resistant Prevotella species in saliva³⁰  and some, such as P intermedia and P aeruginosa, are often found in implants with peri-implantitis.³¹  In this context, an in vitro study found that 1 or more pathogenic species found in implants with peri-implantitis, especially P intermedia, Tannerella forsythia, and A actinomycetemcomitans, are resistant at therapeutic concentrations in 46.70% of cases to clindamycin.³²  Besides, several studies have linked the prescription of antibiotics other than β-lactams to an increase in methicillin-resistant Staphylococcus aureus.³³  In this regard, S aureus has been found at high concentrations in implants with peri-implantitis, as have other clindamycin-resistant bacteria mentioned above (A actinomycetemcomitans, P intermedia, and T forsythia) (P < .001).³⁴  Also, if S aureus is part of the early colonizing bacteria of implants, this bacterium will be present 1 year later,³⁵  thus increasing the risk of future peri-implantitis.^35–37  In addition, it generates stepwise microbial resistance, that is, at the slightest exposure of bacteria to clindamycin, they become resistant. Zirk et al³⁸  studied the type of antibiotic appropriate for the treatment of odontogenic maxillary sinusitis, concluding that clindamycin is the antimicrobial with the most unfavorable results with 50% of tested pathogens resistant.³⁹  Pigrau et al⁴⁰  (2009) studied the effect of various antibiotics in the treatment of osteomyelitis in a sample in which 92.48% of patients had previously been exposed to clindamycin for various reasons, including 15.22% for prophylaxis before implant placement. These authors observed that Streptococci viridans was susceptible in 81% to penicillin and 96% to fluoroquinolones, but only 11.5% to clindamycin. At least 1 clindamycin-resistant species was present in 92.10% of the samples, indicating the rapid emergence of resistance in patients previously exposed to clindamycin.

Furthermore, approximately 10%–20% of patients report an allergy or reaction to penicillin; however, these are rarely hypersensitivity or immunoglobulin E mediated reactions, so these drugs could be administered safely.^41–43  Moreover, studies have shown that 80%–99% of patients may no longer be considered allergic after allergy testing.^44–46  Therefore, this study recommends that patients self-declared allergic to penicillin should be tested for sensitivity to β-lactams and,²⁹  if a true allergy is confirmed, the prescription of azithromycin 500 mg 1 hour before surgery is recommended,⁴⁷  as it has been found in concentrations of 224 and 203 mg/L in gingival and peri-implant crevicular fluid 13 days postsurgery, respectively. Azithromycin has also important effects on early inflammation and healing by decreasing levels of granulocyte colony-stimulating factor, interleukins 6 and 8, macrophage inflammatory protein 1β, and interferon-induced protein 10 kDa, reducing mobilization of granulocyte precursors and recruitment of immune and inflammatory cells during the healing phase, with higher bioavailability than amoxicillin.⁴⁸  This antimicrobial has demonstrated 30%–80% action against sensitive strains of periodontopathogens.¹⁵ 

Due to the low infection rates associated with these procedures, there is no justification for prescribing PA. Comparing these procedures with those performed on teeth is not entirely possible because they differ in their characteristics. There are implants where an intermediate abutment is connected, the mucosa heals completely, and an impression transfer or a definitive abutment can be connected to the implant without breaking the peri-implant mucosal seal. In contrast, when healing abutments are connected directly to the implant connection, a small wound is produced by temporarily exposing the connective tissue that was attached to the surface of the prosthetic attachment.⁴⁹  Connections/disconnections of prosthetic attachments at this stage will be identified by the body as a “wound” in the connective tissue, which may favor the entry of micro-organisms.⁵⁰  Moreover, during continuous disconnections and reconnections, microbiologically contaminated fluids might infect the connection, creating the correct environment for anaerobic bacteria selection.⁵¹ 

No studies were found that quantified the magnitude of bacteremia associated with the connection/disconnection of prosthetic attachments. Previous studies that evaluated taking alginate impressions on teeth have shown no statistically significant differences between bacteremia following alginate impression-taking and bacteremia at baseline. In this respect, Umeh et al⁵²  (2016) observed that the prevalence of bacteremia after impression-taking was 8% (n = 2) compared with 0% at baseline (P = .500). The intensity of bacteremia was very low, with values of 0.52 colony-forming units per mL (CFU/mL) (± 1.33) vs 0.28 CFU/mL (± 0.74), respectively (P = .060). Lucas et al⁵³  (2002) described similar data, with the prevalence of bacteremia before and after taking alginate impressions being 23% and 31%, respectively, as well as the intensity of bacteremia (0.2 CFU/mL [± 0.7] vs 0.3 CFU/mL [± 0.6]). For these reasons, the magnitude of this possible bacteremia is currently unknown, but its significance is not considered clinically relevant from the point of view of the occurrence of infectious complications. In this regard, a consensus document published in 2006¹⁵ classified impression-taking as a procedure with a low associated risk of infection.

To avoid bacteremia associated with these procedures, it is recommended to use topical antimicrobials such as chlorhexidine digluconate, as it has shown to significantly reduce biofilm growth on healing abutments compared with a placebo group.⁵⁴  It is also recommended to avoid reusing prosthetic abutments between patients, as the routine methods (mechanical, chemical, and steam) used for cleaning and sterilizing the healing abutments did not result in complete removal of decontaminants. Despite this, no studies were reporting adverse consequences for dental implants, such as infections, bone loss, mechanical complications, implant failure, or harm to patients.⁵⁵ 

This systematic review presents several strengths, such as an unrestricted search in the literature (including grey literature), the searching process of studies, data extraction, and risk analysis bias performed in duplicate.

Nonetheless, some limitations may be related to this systematic review. First, the low number of studies available in the literature. Second, 2 included articles^14,15  that presented a high risk of bias and their data should be interpreted with caution.

Future studies should include the design of RCTs to specifically analyze the influence of prescribing PA in second-stage implant surgeries, with and without the use of soft tissue grafts, vs not prescribing them. It would also be interesting to know the magnitude of possible bacteremia associated with the connection/disconnection of the prosthetic components of the implants produced during impression-taking and placement of the prosthesis on implants.

Considering the limitations of the present study, prescribing PA in second-stage implant surgeries, impression-taking and/or implant prosthetic placement are not justified. In second-stage implant surgeries, or in peri-implant plastic surgery procedures lasting more than 2 hours and/or in which soft tissue grafts and/or biomaterials are used extensively, the administration of PA may be indicated. In these cases, and given the current lack of evidence, it is recommended to prescribe 2 g of amoxicillin 1 hour before surgery. In patients allergic to penicillins, the prescription of 500 mg of azithromycin 1 hour before surgery may be indicated.