Clinical Microbial Identification of Severe Oral Infections by MALDI-TOF Mass Spectrometry in Stockholm County: an 11-Year (2010 to 2020) Epidemiological Investigation

ABSTRACT

Growing evidence suggests that oral infections can modify the course of systemic diseases. To date, epidemiological data on microbial oral infections are scarce. Here, we performed a comprehensive analysis of the trend and microbial diversity in oral infection specimens referred for clinical microbiology analysis from 2010 to 2020. The microbes were isolated by culture and were identified via matrix-assisted laser desorption ionization-time of flight mass spectrometry technology (MALDI-TOF MS) throughout the study period. A total of 1,014 referred samples from dental clinics in Stockholm County with dentoalveolar abscesses and jaw osteomyelitis being the main reason were identified. Overall, the microbial composition was dominated by Firmicutes (51%), followed by Bacteroidetes (19%), Proteobacteria (12%), and Actinobacteria (5%). At the genus level, Streptococcus spp. (36%), Prevotella spp. (18%), and Staphylococcus spp. (11%) were among the most frequently reported. Interestingly, a strong increase in trend was noted for Streptococcus anginosus, Streptococcus mitis, Streptococcus sanguinis, Eikenella corrodens, Actinomyces spp., Aggregatibacter aphrophilus, Staphylococcus epidermidis, and Granulicatella adiacens during the study time (R = 0.66 to 0.89, P < 0.05), and a minor increase was noted for Enterococcus faecalis and Klebsiella spp., whereas steady levels were noted for most of the others. The present study shows the diversity of bacteria that have been involved in dental infections during the last decade in the capital of Sweden, as well as the emerging oral microbiota trend, with clear clinical implications on the oral-systemic link.

IMPORTANCE Oral diseases and associated microbes are a risk factor for systemic diseases and can change the courses of these diseases. To date, epidemiological data on microbial oral infections are scarce, and longitudinal reports are lacking. We present for the first time the microbial composition of severe oral bacterial infections determined via the MALDI-TOF mass spectrometry technique in a comprehensive study between 2010 and 2020 (11 years) in Stockholm County. The trend and microbial diversity of oral infections were analyzed on referred clinical microbiological samples and were processed by standardized protocols. Trend increase was noted for Streptococcus anginosus, Streptococcus mitis, Streptococcus sanguinis, Eikenella corrodens, Actinomyces spp., Aggregatibacter aphrophilus, Staphylococcus epidermidis, Granulicatella adiacens, Enterococcus faecalis, and Klebsiella spp. Our results provide new insights into the diversity and trend of oral microbiota that were involved in serious oral infections over the past decade in the capital of Sweden and may influence the oral-systemic link.

INTRODUCTION

The oral cavity is a crucial component of the human body and is essential to overall health and well-being. Despite advancements in oral health care systems, oral diseases are still listed among the top leading noncommunicable diseases (1). Oral disease manifestations not only contribute to the increasing disease burden in our societies, as noted in the recent Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) report, but also raise medical concerns (2). Especially in aging, socially deprived, or medically deprived communities, common oral/dental diseases, such as dental caries, periapical periodontitis, soft tissue abscesses, and periodontal diseases, may progress and exacerbate the burden of other diseases (2). Consequently, serious infections might be developed, and these might require emergency and urgent care.

Oral infections commonly originate from an odontogenic source and are typically caused by opportunistic microbes (3–5) that disseminate via deep dental caries, endodontic, periodontal diseases, or tissue infections (6). Deep dental caries and endodontic infections are the main causes of dental abscesses (6), whereas periodontal diseases, including periodontitis, or pericoronitis cause periodontal abscesses (7). Clinically, the effects range from localized abscesses to deep head and neck space infections, which may require hospitalization (8). Jaw osteomyelitis and odontogenic maxillary sinusitis constitute other serious complications following odontogenic infections (9, 10). Osteonecrosis is also frequently associated with immunosuppression (11) as well as head/neck radiation, and it can be medically induced (12–15).

Accumulating research suggests that oral diseases are an increasingly common risk factor for systemic diseases (16) and are an emerging challenge for health care (17, 18). Oral pathobionts or harmless commensals may “turn against” the host and cause tissue damage or reactivity if an opportunity is present (19). In this context, severe infections in the oral cavity and associated microbes are important to mine and monitor. Earlier studies using culture-based methods reported mixed oral microbiota, which mainly consisted of facultative anaerobes with annotations limited to the Streptococcus spp. and strict anaerobes, particularly Prevotella and Fusobacterium spp. (4).

Matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) became a standard clinical microbiology method nearly a decade ago and is the method of choice for the identification of microorganisms in clinical laboratories. MALDI-TOF MS is an accurate, fast, and cost-effective technology developed for routine clinical diagnostics (20). It is highly sensitive to a variety of clinical samples and allows microorganisms to be identified using their proteomic fingerprints, which are primarily made up of ribosomal proteins (20). To date, the usage of this technique to describe the microbiological profile of oral infections at a population level has not been reported.

Here, we conducted an 11-year retrospective epidemiological study in Stockholm County to determine the microbial profile longitudinally of severe oral infection samples that were referred to the Karolinska University Clinical Microbiology Laboratory.

RESULTS

Demographic distribution.

Samples were analyzed from a total of 1,014 patients. The average age was 56.34 (± a standard deviation of 21.70) years. 469 were female, and 545 were male. Dentoalveolar abscess was the most prevalent clinical diagnosis, followed by jaw osteomyelitis (Table 1). Adults in age groups ≥60, 40 to 59, or 18 to 39 each accounted for around 30% of the total patients, and 5.5% were from individuals who were <18 years old. A total of 102 bacteria species were reported during the entire study period, with a range of 1 to 7 species per sample. Proportions of mono-respective polymicrobial culture findings were similar across indicated subgroups, except for the jaw osteonecrosis subgroup, in which polymicrobial cultures were more likely to occur, as noted by a logistic regression analysis, (P < 0.001).

TABLE 1.

Multivariable logistic regression analysis assessing clinical diagnosis, gender, age, and sample material to predict the occurrence of polymicrobial culture^a

Variable	Sample (%)	Microbial culture (%)b		Coe.	SE	OR (%)	CI (95%)	P value
		Mono-	Poly-
Clinical diagnosis
Dentoalveolar abscesses	55.6	50.5	49.5	-	-	1	-	Reference
Jaw osteomyelitis	15.3	51.9	48.1	0.136	0.223	1.14	0.74 to 1.77	0.540
Jaw osteonecrosis	8	30.9	69.1	0.952	0.294	2.59	1.46 to 4.61	0.001*
Odontogenic maxillary sinusitis	4.1	56.1	43.9	0.245	0.431	1.28	0.55 to 2.98	0.570
Periradicular pathosis	3	36.7	63.3	0.873	0.480	2.39	0.94 to 6.13	0.068
Periodontal diseases	2	60	40	−0.322	0.512	0.73	0.27 to 1.98	0.530
Age in years
≥60	33.5	50.3	49.7	-	-	1	-	Reference
40 to 59	32	52.7	47.3	0.100	0.170	1.11	0.79 to 1.54	0.557
18 to 39	29	45.8	54.2	0.131	0.167	1.14	0.82 to 1.58	0.435
<18	5.5	53.9	46.2	−0.120	0.323	0.89	0.47 to 0.67	0.712
Gender
Male	53.8	49.6	50.4	-	-	1	-	Reference
Female	46.3	51.4	48.7	0.012	0.130	1.01	0.78 to 1.31	0.930
Sample material
Abscess	72	48.5	51.5	-	-	1	-	Reference
Mucosa	10	62.1	37.9	−0.483	0.307	0.62	0.33 to 1.12	0.116
Tissue	7	51.4	48.7	−0.424	0.347	0.65	0.33 to 1.29	0.222

^aCoe., coefficient; SE, standard error; OR, odds ratio; CI, confidence interval.

^bThe dashed line (−) indicates that no value has been delegated to Coe., SE, or CI in the reference variables. *, a P-value with statistical significance (P < 0.5).

Microbial composition and annual trend from 2010 to 2020.

To investigate the microbial composition and trends of the identified bacteria, the data were arranged at the phylum, genus, and species levels. As shown in Fig. 1A, 4 phyla constituted the vast majority of the detected microbial species: in ranked order, Firmicutes (51%; genera Streptococcus, Staphylococcus, Enterococcus, Lactobacillus, and Granulicatella), Bacteroidetes (19%; genera Prevotella, and Capnocytophaga), Proteobacteria (12%; genera Haemophilus, Eikenella, Klebsiella, Escherichia, Enterobacter, Aggregatibacter, and Neisseria), and Actinobacteria (5%; genera Actinomyces, Cutibacterium, and Rothia); with the remaining 13% being unclassified. At the genus level, Streptococcus spp. and Prevotella spp. account for almost 50% of the reported bacteria, followed by Staphylococcus spp. (11%). The microbial distribution at the phylum, genus, and species levels also exhibited a predominance of Streptococcus spp., Prevotella spp., and Staphylococcus spp. across the 11 years included (Fig. 1B–D).

FIG 1.

Overall microbes in oral infections. A global view of the oral bacteria reported during the entire 11-year study period, classified at the phylum, genus, and species levels, presented as a donut chart (A). Stacked bars illustrate the yearly frequency of the microbial phyla (B), genera (C), and species (D) reported for the oral infection samples. The data are expressed in percentages.

A subsequent trend analysis based on the annual data (Fig. 2) indicated that among the Streptococcus spp., the frequency of Streptococcus anginosus, Streptococcus mitis, and Streptococcus sanguinis had increased significantly (R = 0.79, R = 0.86, and R = 0.66, respectively; P < 0.05) (Fig. 2). Moreover, significant increases were also noted for Staphylococcus epidermidis (R = 0.69), Eikenella corrodens (R = 0.77), Actinomyces spp. (R = 0.73), Aggregatibacter aphrophilus (R = 0.89), and Granulicatella adiancens (R = 0.70; P < 0.05) (Fig. 2). A mild to moderate increase was also noted for Enterococcus faecalis, Klebsiella spp., Capnocytophaga spp., Lactobacillus spp., and Neisseria spp., (R = 0.30, R = 0.38, R = 0.16, R = 0.35, and R = 0.39, respectively; P > 0.05) (Fig. S1). On the other hand, a moderate decline was found for Prevotella spp., Haemophilus spp., Escherichia coli, and Streptococcus pyogenes (R = −0.42, R = −0.70, R = −0.50, and R = −0.40, respectively; P > 0.05) (Fig. S1).

FIG 2.

Trends of the reported microbes in oral infections over the 11-year study period. A linear regression analysis indicates a significant increase of Streptococcus anginosus, Streptococcus mitis, Streptococcus sanguinis, Eikenella corrodens, Actinomyces spp., Aggregatibacter aphrophilus, Staphylococcus epidermidis, and Granulicatella adiacens in oral infection samples. P values of <0.05 are considered to be indicative of a statistically significant result.

Microbial distribution and annual trend by Gram staining, morphology, and oxygen metabolism.

Overall, the proportions of Gram-negative and Gram-positive bacteria were 55.4% and 44.6% over the past 11 years (Fig. 3A). The frequency of Gram-positive species increased significantly at the expense of Gram-negative species over the study period (R = 0.68; P < 0.05) (Fig. 3B and C). As for microbial morphology, bacilli accounted for 47% of the total reported microbes, followed by cocci (33%), and coccobacilli (20%) (Fig. 3D). While the frequencies of bacillus species remained stable over time (R = 0.24; P > 0.05), the frequency of the coccobacilli and cocci changed significantly over time (R = −0.68 and R = 0.60, respectively; P < 0.05) (Fig. 3E–G). As for oxygen metabolism, facultative anaerobes were the most frequently detected, followed by anaerobic and aerobic species (Fig. 3H). A trend analysis indicated no significant changes in either of these species (Fig. 3I–K).

FIG 3.

Distribution and trends of reported microbes of oral infections, as classified by Gram staining, morphology, and oxygen metabolism. Stacked bars illustrate the yearly frequencies of Gram-positive versus Gram-negative bacteria (A), coccobacilli, coccus, and bacillus (D), facultative anaerobe, aerobic, and anaerobic (H) bacteria that were detected in oral infection samples. A linear regression analysis indicates the trends of the Gram-positive (B), Gram-negative (C), coccobacilli (E), coccus (F), bacillus (G), facultative anaerobe (I), aerobic (J), and anaerobic (K) bacteria noted in oral infection samples over the 11-year study period. The data are expressed in percentages. P values of <0.05 are considered to be indicative of a statistically significant result.

Microbial profiles by clinical diagnosis and sample material.

Among the 1,014 identified samples, 890 were assigned to six different clinical diagnoses: dentoalveolar abscesses, jaw osteomyelitis, jaw osteonecrosis, odontogenic maxillary sinusitis, periradicular pathosis, and periodontal diseases. 124 samples were listed as undiagnosed. The most frequently involved facial spaces were the submandibular site (43%), followed by the buccal site (38%), submental region (11%), palate (3%), canine (3%), and pterygomandibular region (2%). As shown in Fig. 4A–F and in Table 1, dentoalveolar abscesses accounted for 55.6% of the total samples, in which S. anginosus, Prevotella spp., and Haemophilus parainfluenzae were frequently noted. Interestingly, S. anginosus was also commonly associated with jaw osteomyelitis, jaw osteonecrosis, odontogenic maxillary sinusitis, and periradicular pathosis. Other frequent microorganisms in jaw osteomyelitis were S. epidermidis, H. parainfluenzae, Staphylococcus aureus, and E. faecalis. In periradicular pathosis samples, E. faecalis, E. coli, and Prevotella spp. were frequently detected, whereas in periodontal diseases, Prevotella spp. and Streptococcus spp. were detected in over 50% of the cases. Across the clinical diagnoses, five bacterial species were found to be in common: S. anginosus, S. aureus, E. faecalis, E. corrodens, and S. mitis.

FIG 4.

Microbial profile of oral infections based on their clinical diagnosis group. Bars illustrate the frequencies of the top-10 microbes reported in dentoalveolar abscesses (A), jaw osteomyelitis (B), jaw osteonecrosis (C), odontogenic maxillary sinusitis (D), periradicular pathosis (E), and periodontal diseases (F), respectively. Stacked bars in inserted panels indicate the proportions of Gram-positive versus negative, facultative anaerobe, aerobic, and anaerobic bacteria phenotypes, respectively. The data are expressed in percentages. (G) Hierarchical clustering heat map representation of the occurrence of bacterial species between different oral infections. Colors represent the level of species occurrence, which is indicated as low (light green) to high (light red), based on the Euclidean distance from the maximum level of each species among the diagnosed groups.

Although a general predominance of Gram-positivity and facultative anaerobes was noted, the periodontal disease group frequently showed Gram-negativity as well as anaerobic species (Fig. 4F). A deeper examination via a heat map analysis on the microbial species between the disease conditions also showed a substantial interecology occurrence (Fig. 4G). Interestingly, while the strictly anaerobic Fusobacterium nucleatum and Aggregatibacter actinomycetemcomitans (a fastidious HACEK organism) were noted, they occurred mainly in cases of jaw osteomyelitis or dentoalveolar abscess. Other fastidious HACEK organisms, such as E. corrodens, H. haemolyticus, H. parainfluenza, and A. aphrophilus, also occurred mainly in jaw osteonecrosis and periradicular pathosis cases, and occasionally in cases of periodontal disease, as revealed by the clinical MALDI-TOF approach.

As for the sample material, abscesses were the most prevalent (72%), followed by mucosa swabs (10%), and tissue samples (7%). 11% of all samples were comprised of material of an unspecified origin. Interestingly, the Venn analysis presented in Fig. 5 revealed that 21 species are shared by all 3 sample types and could represent the core microbes of the oral infection samples reported in this study. Among the core members were E. corrodens, E. faecalis, S. anginosus, S. mitis, and S. epidermidis as well as members of gamma-proteobacteria, such as Enterobacter cloacae, E. coli, Klebsiella oxytoca, and Klebsiella pneumoniae. Intriguingly, all of these microbes were also noted earlier among the emerging bacteria during this 11-year study period (Fig. 2; Fig. S1), and none of them were uniquely restricted to any of the reported oral diseases (Fig. S2).

FIG 5.

Unique and shared core microbiota in oral infections. The Venn diagram depicts the relation between the identified bacteria species, organized according to the indicated sampling materials. The overlapping subset indicates the shared core species found in all three sample types.

DISCUSSION

A central clinical question in oral infections is the identification of microbes associated with these infections. While previous studies have often reported mixed oral microbiota and presented a limited overview at the population level, this study focuses on oral bacteria at the species level and the epidemiology over an 11-year period of analysis. Here, we report the microbial composition of oral bacterial infections, as determined by MALDI-TOF MS technology, for the first time and confirm that dentoalveolar abscesses remained the most prevalent clinical diagnosis.

According to a recent nationwide survey in the United States, the number of periapical abscess-related emergency visits increased from 460,260 in 2008 to 545,693 in 2014, whereas the average hospital stay increased in duration from 2.96 days in 2008 to 4.5 days in 2014 (21). A similar observation has also been seen in our study, with dentoalveolar abscesses remained the main reason for microbiological testing over the study period. Our findings also revealed that microbes associated with dentoalveolar abscesses are diverse in nature, with a clear predominance of two genera, namely, Streptococcus spp. (in particular, S. anginosus) and Prevotella spp., which is consistent with the results of previous culture-dependent studies (22, 23). Moreover, S. anginosus alone, as a single species, accounted for over 20% of the bacterial infections that were reported in this study. The unique characteristic of the S. anginosus group (group D streptococci) that sets them apart from other pathogenic streptococci, such as S. pyogenes (group A streptococci) and Streptococcus agalactiae (group B streptococci), is their ability to cause abscesses in different body organs (24). However, some of the Gram-negative bacteria, such as Veillonella spp. that had been previously isolated from dentoalveolar abscesses via the traditional culturing method (22, 23) were not reported by MALDI-TOF MS in the current study. The absence of such bacteria could be attributed to culture-related factors, as a minimum concentration of microbes or a certain culture medium might be required for their clinical isolation and recognition.

Clinical microbiological examination investigates disease-causative and/or disease-associated agents and is mainly requested for difficult-to-treat cases. We present an epidemiological and longitudinal overview at the population level that has been lacking in the literature, in part because cultivated oral microbiology samples could not be classified to the species level before the MALDI-TOF MS analysis era. Monitoring the changes over time via trend analysis has given some insights into the context of the oral-systemic disease link. Given that severe tissue infections are seldom confined only to the locally affected area, it is sensible to monitor disease-associated microbes, as they may cross tissue barriers, evade immune surveillance, and thus disseminate to distant organs through the circulatory and lymphatic systems. If managed incorrectly, systemic infections and fatal conditions affecting the neural, cardiovascular, or pulmonary regions could occur (25, 26). Hence, microorganisms identified in oral infections have gained attention as potential biomarkers for some systemic disorders (18, 25).

Our findings demonstrate that a number of microorganisms have appeared much more frequently in oral infections over the past 11 years, specifically S. anginosus, S. mitis, S. sanguinis, S. epidermidis, E. corrodens, A. aphrophilus, E. faecalis, Klebsiella spp., Granulicatella spp., Lactobacillus spp., Neisseria spp., Capnocytophaga spp., and Actinomyces spp. Intriguingly, these microorganisms are frequently mentioned in the context of cancer development and progression, as well as in various systemic diseases, such as pericarditis and endocarditis (17, 18, 27, 28). Many of the gamma-proteobacteria reported here are also detected inside tumor lesions including in early precursors to invasive pancreatic cancer via microbiome and proteome analysis, which may interfere with cancer chemotherapy (29–31). S. anginosus and S. mitis remain common causes of abscesses in the liver and spleen, where immunosuppression, diabetes, and heavy alcohol use are among the known risk factors (24). The immunomodulatory effects of, for example, bacterial endotoxins, metabolites, and inflammatory mediators may modify the course of systemic health in (including but not limited to) cardiovascular diseases and malignancies along the gastrointestinal tract.

A recent report found that approximately 13% of the global cancer burden is attributed to microbial infections (32, 33), based on all known “oncomicrobes” to date. New evidence, however, suggests that the oncomicrobe category may be broader than previously thought and that “complicit” microbes specialised with microbial features such as genotoxin-mediated mutagenesis, may promote carcinogenesis (34). Deeper investigations to uncover such risk factors for earlier detection and intervention could be important in management of patients at risk.

Our trend analysis revealed an increase of the Gram-positive cocci pathogens in oral infections over the past 11 years. This supports recent findings from urinary tract and blood-based clinical microbiology and from epidemiological studies that report significant increases in Gram-positive infections over the past decade (35, 36), including multiple drug-resistant strains of Gram-positive cocci (37), which remain as a major global concern. This new information has substantial therapeutic implications, especially for odontogenic infections, including the possible use of NSAID/cyclooxygenase inhibitors to block the Gram-positive bacteria-associated NF-B overactivation in order to dampen tissue destruction (38). While Sweden remains in a favorable position compared to many other countries in terms of antibiotic resistance, the number of cases with a multiresistant Gram-positive infection is on the rise (39). The oral cavity can be a reservoir of such microbes, and oral health awareness could thus enable early intervention.

As estimated by Siqueira Jr. et al. via 16S rRNA sequencing (40), the microbial community of acute dental abscesses is a mix of 20 to 30 taxa per sample. However, data from the Human Oral Microbiome Database demonstrates that 24 to 46% of the oral taxa were uncultivated phylotypes. Thus, the clinical MALDI-TOF approach is a complementary alternative to 16S rRNA sequencing, as up to seven unique viable species could be classified simultaneously in a clinical routine laboratory. Possibilities for experimental functional studies of live oral microbiota and the whole-genomic sequencing of various specialized ecologies could improve the current understanding of human oral microbiota. The downside of this clinical approach is that some species are difficult to cultivate and require special culture conditions. Hence, our frequently reported live core species did not exactly meet the taxa reported earlier, (i.e., F. nucleatum, Porphyromonas spp., Treponema spp., Parvimonas micra, Tannerela forsythia, and Dialister spp.) (40).

Jaw osteonecrosis is another oral condition, and it is caused by radiation-induced injury or medication-induced osteonecrosis (41). In our study, samples from patients diagnosed with jaw osteonecrosis were found to contain a high predominance of facultative anaerobes with a high frequency of Streptococcus spp. and Prevotella spp. These findings support the findings of reports from recent research (42, 43). Moreover, microbes that are thought to be the leading causes of delayed wound healing in osteonecrosis (44), such as Streptococcus spp., S. aureus, and Pseudomonas aeruginosa, have also been found in our study. Contrary to other studies (42, 43), our findings showed lower levels of Actinomyces spp. and P. aeruginosa in osteonecrosis samples, as well as a lack of Veillonella species. The absence of Veillonella spp., as well as the reporting of Actinomyces spp. and P. aeruginosa at a lower frequency, could be attributed to differences in the site of the harvested samples, history of antibiotic therapy, overgrowth of concomitant microorganisms, or culturing methods.

Another type of oral infection is jaw osteomyelitis, which represents an inflammation of the jawbone, triggered by microbial infection (45). In this study, we found Streptococcus, Staphylococcus, Prevotella, Enterococcus, and Cutibacterium species, with a greater proportion of facultative anaerobes being found in the jaw osteomyelitis samples. These findings are in line with those of earlier reports (45, 46). However, while Parvimonas micra in jaw osteomyelitis has previously been described (45), our findings did not confirm the presence of this isolate. This could be explained by the challenges associated with the cultivation of this microorganism.

There is currently limited data on the microbiology of odontogenic maxillary sinusitis. Earlier studies showed that most microorganism populations consist of mixed aerobic and anaerobic species (10). In line with this, our findings revealed that samples from odontogenic maxillary sinusitis may consist of a range of different bacterial genera of aerobic (Streptococcus spp., Staphylococcus spp., and Neisseria spp.) and anaerobic (Prevotella spp. and Haemophilus spp.) bacteria. A relatively similar microbial composition to that found in dentoalveolar abscesses is also noted here. For the periapical pathosis samples, Streptococcus spp. and E. faecalis appear to be the most frequent. The latter is commonly associated with failed endodontic therapy and postendodontic treatment complications (47). Meanwhile, E. faecalis has a high antimicrobial resistance, can survive in harsh environments with a limited nutrient supply, and induces hydroxyapatite deposition and biofilm calcification as resistance mechanisms (47). Other frequently reported microbes found here include E. coli, Prevotella spp., Actinomyces spp., and Cutibacterium spp., all of which are in line with earlier research (48).

The current study’s microbiological findings from the periodontal disease samples are consistent with those of previous reports (49), in which Prevotella spp., Capnocytophaga spp., and Eikenella spp. are frequently reported. According to Haffajee and Socransky’s classification (50), these microbes are regarded as strong pathogens in periodontal diseases, indicating plaque-induced gingivitis or peri-implantitis rather than chronic periodontitis conditions. However, the absence of Fusobacteria spp. in the periodontal samples could be explained by the relatively small sample size of this group and the category of the periodontal infections submitted to this analysis (i.e., acute rather than chronic cases). Other factors which may have affected our results include the competitive metabolic microenvironments, cross-feeding, fastidious microorganisms, or overgrowth of single microorganisms during the cultivation process (51). We opt to present clinical oral microbiology from a large, real-world, clinical cohort, as we believe it could bring insights into the context of the clinical management of oral infections (that is, the selection of antimicrobial therapy, the determination of whether hospitalization is necessary, and the change of mortality and progression of infections).

One of the main strengths of our study is that it is the first study to describe the microbial profile of different severe oral infections at the population level, using MALDI-TOF MS technology, with a relatively large sample size. This technology allows for the precise identification of microorganisms that remained unidentifiable by conventional phenotypic methods, is rapidly becoming a cost-effective alternative to 16S rRNA gene sequencing, and presents with unquestioned reliability (20). We also analyzed for the first-time longitudinal trends in oral infections over an 11-year study period, core microbes across different oral infections, and various sampling materials. Further, the adopted culturing technique and microbial identification procedures were conducted using a standardized laboratory protocol throughout the whole 11-year investigation period.

We acknowledge that this study has a number of limitations. First, it is retrospective in design, which limits the evaluation of factors that are not regularly documented, including histopathological evaluations and the use of antibiotics. Another limitation is that clinical details, including treatment outcomes, were not collected. Combining this information with microbiology findings will shed more light on the current study and help us to understand the potential emerging pathogenicity of oral microbes. Moreover, the microbial load/density and antimicrobial susceptibility results were not integrated into the current microbiological database, and the microbial findings in the current study are based exclusively on reports of bacterial outputs. Since oral infections are typically polymicrobial in nature, combining the MALDI-TOF results with a 16S rRNA gene sequencing analysis in future studies will provide complementary insights into the microbial ecology of these infections. Lastly, as this study was conducted in the capital of Sweden, our results may not be generalizable to other parts of the world. Comparisons of microbial trends with other geographical areas thus remain to be performed.

MATERIALS AND METHODS

Ethical approval.

The study was approved by the regional ethical review board in Stockholm, Sweden (Dnr 2021-03974) and was deemed to be in compliance with good clinical practice, the principles of the Declaration of Helsinki, and the standards of Karolinska University Hospital.

Sample identification.

The study was carried out retrospectively by using the sample data/records from within the Qlik-View software system (Qlik, Lund, Sweden) at Karolinska University Hospital’s Clinical Microbiology lab, spanning the years from 2010 to 2020. All clinical samples with microbial growth were subjected to a MALDI-TOF MS (Bruker, Daltonik, GmbH, Bremen, Germany) analysis. The exported data were sorted in Excel and filtered with the following key terms: dental, tooth, oral care, gum, mouth, jaw, craniofacial, and maxillofacial. All oral infection samples submitted by dental clinics and hospitals in Stockholm County were collected. In addition to the microbiological data, clinical variables, including the clinical diagnosis (dentoalveolar abscesses, jaw osteomyelitis, jaw osteonecrosis, odontogenic maxillary sinusitis, periradicular pathosis, and periodontal diseases), involved facial space/s, and sampling materials (abscess, mucosa swabs, or tissues) were collected, as were general variables, such as the age and gender of each patient.

Culturing technique.

All samples were cultured on blood, chocolate, cystine lactose electrolyte deficient (CLED), and Brucella laked blood agar plates and were incubated under aerobic and anaerobic conditions. For the aerobic conditions, the blood and CLED plates were incubated at 37°C in a relevant atmosphere, while the chocolate plate was incubated at 37°C in a 5 to 10% CO₂ environment. For the anaerobic conditions, an anaerobic chamber (Anaerocult A; Merck, Darmstadt, Germany) was used to incubate the blood agar and Brucella lacked blood agar plate with kanamycin and vancomycin at 37°C. All of the agar plates were read twice, after 24 h and 48 h for the aerobic plates, and after 48 h and 6 days for the anaerobic plates. After microbial isolation, the colonies were conveyed to MALDI-TOF MS identification.

MALDI-TOF MS identification.

A well-formed colony from the culture plates was directly spread on the MALDI targeted plate. After adding formic acid (1 μL of a 70% solution) and air-drying, 1 μL of matrix solution was applied at room temperature. Using the MALDI Biotyper (version 3.0) and FlexControl (version 3.3) software packages, the ionization and acceleration of the agent macromolecules embedded in small matrix molecules in an electromagnetic field of 10 to 30 k permitted the identification of different flight times. The microbial spectra were then matched against the references in the database. The obtained score value (log) data ranged from ≥2 for species-level identification and ≥1.7 for genus-level identification. Microbes with a score of <1.7 were described by their morphology and were classified as others in this study.

Statistical analysis.

The data were analyzed as primary data. To minimize potential technical errors and to increase the ecological relevance of oral infections, data filtration processes were applied by removing bacteria with a count of less than 2. The descriptive analysis, average number of identified isolates, male or female origin, and age groups were reported as means (± standard deviations) and proportions (%). Microbial frequencies were reported as proportions of phyla, genera, species, gram staining, morphology, oxygen metabolisms, different clinical diagnoses, and sampling materials. For the inferential analyses, a multivariable logistic regression analysis, modeling the odds of monomicrobial or polymicrobial culture, was conducted to identify independent predictors of the clinical diagnosis, age group, gender, and sample material. First, the multivariable analysis was used to calculate partially adjusted odds ratios (OR) and 95% confidence intervals (CI), which were adjusted for the age group, gender, and sample material. The final adjusted analysis took into consideration the simultaneous effects of relevant predictors to investigate their independent impacts on monomicrobial and polymicrobial cultures. Additionally, a linear regression model was used to examine trends of microbial frequency from 2010 to 2020. Based on the R values, the trend was described as increased, reduced, or steady. A hierarchical clustering heat map was created, using the R package pheatmap (R-4.0.4; Core Team, 2021, Vienna, Austria), using the Euclidean distance measure and the Ward clustering algorithm. The diversity of the detected microorganisms was plotted using Venn diagrams. The data were analyzed using SPSS version 24.0 (IBM Corp., Armonk, NY, USA). A P value of <0.05 was considered to be indicative of a statistically significant result.

Data availability.

Data will be made available upon request to the corresponding author, following the publication of the present article.