To summarise the present results, by sonicating the tooth, significantly more bacteria could be detected compared to the swab, as shown by the higher number of OTUs in sonication samples. Furthermore, even the microbial composition of the analysed samples differed between the tested procedures. Moreover, it was possible to find some bacteria which could not be found in the standard procedure for microbiological testing. Focusing on antibiotic resistance rate, no significant difference between sonication fluid, vortex fluid, or swab could be found. A significant difference could be shown in the comparison of the resistance rate of the evaluated antibiotics.
The knowledge of the bacterial composition is crucial for a targeted and effective antibiotic regime. Next-generation sequencing using 16S rRNA gene has shown good results for identification of the oral microbiome [30,31,32,33,34,35]. The microbial composition reported is similar to earlier studies focusing on the oral bacterial composition [30, 32, 33, 36]. The most commonly used sample types to study the healthy oral microbiome and its changes in various diseases were saliva, oral rinse, or niche-specific samples, e.g. supra- or subgingival plaque or tongue swab [31]. Whole teeth have never been analysed before.
For evaluation of the purity of the collected specimen, one control sample for every procedure was taken and analysed separately, as recommended by Zaura et al. [31]. The results showed a low number of reads and diversity, confirming negligible contamination due to the kit-specific microbiome.
The number of reads is a semi-quantitative tool. It is highly affected by the number of PCR cycles performed, the taxa identified, and the sequencing runs itself [37, 38]. Therefore, no precise quantitative measurement is possible regarding how many times more DNA can be found in the sonication fluid compared to the swab or vortex.
In the present study, all patient samples were located on the same sequencing run, with the same number of cycles and same primer respectively, which provides comparability between the swab, vortex, and sonication fluid (Additional file 3).
Taking all of this into consideration, it is highly likely that there is a higher amount of DNA in the sonication fluid compared to the swab and vortex in each patient.
The procedural difference between sonication in contrast to the swab is that they are performed on extracted teeth, resulting in biofilm loosening on the whole surface, whereas the swab was only taken from the alveolus. Therefore, the quantity of the material gathered is probably higher and might have a distorting effect.
The composition of each oral microbiome is different, not only in the number of reads but also in the taxa found. A possible reason for this could be the diversity of the different microbiome surfaces and inter- and intra-individual variations [39]. The oral biofilm development over time is a complex interaction of different species which colonise oral surfaces to form an organised multispecies community with a specific composition. This is caused by the different prevailing physical and biological conditions in the oral habitat, such as surface texture, cell desquamation, or aerobic capacity in the specific niche [6]. Faust et al. demonstrated that the microbiome in different types of samples is similar but nevertheless different [40], highlighting that the source of sampling is crucial for proper microbial testing and especially so for antimicrobial susceptibility testing. This difference could be a reason for the differing microbiological results of the swab and vortex or sonication fluid. Especially regarding the aerophilic capacity of each bacterium, a more anaerobic bacterial composition should be expected in the alveolus or the periodontal pocket than on the tooth surface. By comparing the whole tooth surface, symbolised in the sonication fluid, the bacterial distribution is expected to be different from the bacterial distribution of the alveolus itself.
Apart from this, the contamination of sample extraction kits, during production, can have a potentially misleading impact on the microbiome analysis and consequent conclusion [41].
Even though the biofilm disruption on the whole surface of the tooth could be a potential bias, every bacterium located on the tooth could be able to cause further or could be the reason for the specific infection. Therefore, this setup resembles the reality of biofilm behaviour in the extraction setting in which the potential dissemination of parts of the biofilm can occur, resulting in severe consequences such as infectious endocarditis [42].
A potential criticism could be the amplification of the 16S rRNA gene. Using this procedure, there is no differentiation between bacteria, dead or alive.
Nonetheless, detection of dead bacteria is a potential benefit, because the procedure of swab taking negatively affects the viability of anaerobic bacteria [9]. In normal culture-based analysis, only living bacteria can be examined.
So, the procedure of sonication could be a further influencing factor and cause potential bias. During sonication, little air bubbles explode on the surface of the tooth, leading to the loosening of the biofilm. Bacteria, which are anaerobes or facultative anaerobes, will be highly affected by this excess oxygen [43], even though it could be shown that these bacteria are still alive after sonication of endoprosthesis [18]. Using the 16S rRNA gene, this correlation can be ignored, due to the stability of the genome, even when the bacteria are compromised.
Focusing on the study design, neither orthopaedic nor else studies, which are comparing direct 16S-rRNA-gene analysis and the difference in the microbiome distribution in different sampling modalities exist. Only a few studies have been investigating, whether using 16S-rRNA-gene analysis resulted in similar or even improved results, than normal microbiological testing [26, 44,45,46]. Due to this fact and the appropriate results of 16S-rRNA-gene analysis in microbiome analysis of the oral cavity, we assume, that this is a reliable tool for such investigation [30, 32, 34, 47]. Also in settings, where a dental infection is a potential causing of more severe disease such as medication-related osteonecrosis of the jaw, this sampling method is a potential tool to evaluate and identify the disease-causing bacteria, also in areas hard to reach or where contamination of the normal probing method, swab or tissue sample, is to be expected.
The present results revealed that the amount of Actinomycetales is underrepresented in the normal probing procedure. One potential life-threatening disease, which is hard to diagnose, is craniofacial actinomycosis [48]. In most cases, it is associated with odontogenic infections [49]. Therefore, it can be assumed, that by only taking a swab in combination with normal microbial culture, there is a general underestimation of this disease. Especially in patients undergoing or following radiotherapy due to head and neck cancer, this disease is a feared complication [48]. Also, in medication-related osteonecrosis of the jaw, Actinomyces spp. seem to play a major role in disease progression [50,51,52].
Focusing on Bacteroidales and Mycoplasmatales, which had a higher abundance in swabs, the read numbers of these orders were higher in sonication than in the swab. Showing that there were no bacteria missing but because of the higher amount of Actinomycetales, the relative abundance was lower than in the swab.
The antimicrobial susceptibility to specific antibiotics was not tested in bacterial culture or by molecular genetic analysis, which is a downside of this study. Nevertheless, it was not the primary objective of this investigation to precisely analyse antibiotic resistances, but rather to evaluate the procedure of sonication or vortex as a new tool in microbiological testing in oral surgery. Due to this limitation, no exact statement regarding proper antibiotic treatment can be made.
Heim et al. have already shown the increasing level of resistance found in cultures [53]. Literature-based resistances show the same percentage of resistance as that assigned by enzyme-based evaluation. In particular, the anaerobic species, which are hard to cultivate, show a higher resistance to clindamycin [54], which is still the antibiotic of choice in penicillin-allergic patients [2, 55] according to the German Guidelines [8]. By also evaluating doxycycline and levofloxacin, it could be shown that the resistance rate of amoxicillin with clavulanate was similar to levofloxacin and clindamycin to doxycycline.
Focusing on the analysed resistance rates, a change to the use of levofloxacin as a first-line alternative in severe cases of odontogenic infections should be considered in patients with an allergy to penicillin. Levofloxacin is similar to moxifloxacin, which is already in use in odontogenic infections [2]. In 2011, a comparison of clindamycin and moxifloxacin showed similar outcomes, but with a lower rate of adverse effects in moxifloxacin [56]. Taking this into consideration, levofloxacin or moxifloxacin should replace clindamycin as the first line alternative to amoxicillin with clavulanate in severe odontogenic infections. Due to the high rates of overprescribing of dental antibiotics [1], this change should be exclusively provided for hospitalised patients. The inpatient setting is also preferable for fluoroquinolones, due to their possible side effects like increasing the QT-interval [55]. Other possible side effects are tendinopathy, especially with long-term use [57], and drug-drug interactions. Because of the chondrotoxicity, it is not recommended to use this during pregnancy and in children.
In conclusion with the help of sonication, it was possible to find additional species which were not found in the standard procedure, swabs. The whole microbiome constitution differs, showing a potential incongruence between the standard method and our shown procedure. Still, the microbiomes found were similar in the swab, vortex, and sonication. Consequently, in high-risk patients requiring a more targeted antibiotic treatment (e.g., complex infections, former or ongoing radiotherapy, former or ongoing bisphosphonate medication, congenital heart disease, or immunosuppression) sonication of the tooth should be considered to gain a more complete image of the potential disease-causing microbiome. This sample provides the option to obtain as much information on the bacterial colonisation of the tooth as currently possible, meaning that it can therefore improve treatment as well as clinical outcomes. Early targeted treatment or the prevention of severe complications in high-risk patients can be necessary for their survival.
Limitations
This evaluation was performed on healthy patients, where the hosts’ anti-infective capability is high and severe complications are rare. Due to the small cohort and the characteristic as a pilot study, further investigation should be performed not only focusing on the differences between sonication and the standard procedure for microbial testing in the treatment of infections of the maxillo-facial region. Additionally, the focus should be laid on changes in the oral microbiome in immune-compromised patients. No differentiation has been performed regarding the location of the teeth or the grade of carious destruction.
Yet, sonication could be a tool, especially for immune-incompetent patients, to improve the overview of the bacteria in the infected area, allowing for a more targeted antimicrobial therapy. Also, differences in conventional microbiological testing: bacterial culture, identification, and bacterial susceptibility are planned, to validate the found data, in a bigger cohort. If sonication is not accessible, we could show that vortex could also be considered for loosening of the biofilm on extracted teeth.