Cellulitis in the maxillofacial region is acute, diffuse inflammations of the subcutaneous tissue that spread through the intercellular spaces to the surrounding anatomical areas [1]. If not treated in time, common odontogenic infections spread to various contact areas of the head and neck and damage dangerous anatomical zones [2,3].

The spread of infection outside the bone depends on several factors, such as: the state of the patient's immune system and the immunological response, the virulence of the pathogens and their resistance to antimicrobial agents. Diseases that interfere with the normal immune response, such as diabetes mellitus, liver cirrhosis, prolonged alcohol consumption or immunosuppressive drugs, increase the likelihood of developing cellulitis [4].

Cellulitis in the maxillofacial region can become a life-threatening condition. According to a study by Blankson et al. [5], the mortality rate for complications is up to 5.8%. Recent studies have reported increasing microbial resistance to antibiotics, and a higher incidence of ambulance calls due to impairment from facial and maxillofacial cellulitis of odontogenic origin [2]. There are a number of studies in the scientific literature that examine the characteristics of cellulitis of odontogenic origin in hospitals. Therefore, it is important to evaluate the aetiopathogenesis and treatment of maxillofacial odontogenic cellulitis by analysing the most common pathogens of odontogenic cellulitis and their sensitivity/resistance to antibiotics, as well as their interaction with prescribed antibiotic therapy. It is also important to identify factors associated with longer overall duration of treatment of maxillofacial odontogenic cellulitis and duration of treatment in the intensive care unit (ICU). The aim of this retrospective cohort study is to evaluate the aetiopathogenesis and treatment of odontogenic maxillofacial cellulitis by analysing the medical records of patients treated at the Hospital of Lithuanian University of Health Sciences (Kauno klinikos). This is based on the null hypothesis that the duration of the patient's treatment depends on the pathogen causing the maxillofacial odontogenic cellulitis, its resistance to antibiotics and the type of antibiotic therapy prescribed for treatment.

This retrospective study was conducted at the Hospital of Lithuanian University of Health Sciences (Kauno klinikos) by examining the medical records of patients with facial and maxillofacial cellulitis of odontogenic origin treated at the clinic in the period from January 1, 2018 to January 1, 2023. Before starting the research, approval was obtained from the Centre for Bioethics (No. 2023-BEC2-113) and approval from the Clinics Higher Education and Research Coordination Unit (Kauno klinikos) (No. SPBT-130). The selection of research material is purposive. All participants have read and signed informed consent form. A total of 710 medical records were collected from patients corresponding to the diagnosis of the International Classification of Diseases (ICD) code K12.2 (cellulitis and abscess of mouth) in the presented period.

Selection criteria

The main criteria for inclusion in the study have been established:

The exclusion criteria were as follows:

Of the 710 patients who met the criteria for ICD code K12.2, 32 medical records of children under the age of 18 were rejected. In addition, the study did not include the medical records of 285 individuals who had other inflammatory conditions of the facial and maxillofacial region that did not correspond to the clinical signs typical of facial and maxillofacial cellulitis. Thirty seven cases of cellulitis of non-odontogenic origin were rejected, as were 138 cases in which the patients were not treated in the ICU. A further 119 cases were also excluded from the study, 96 of which had no growth of pathogens in pus cultures and 23 of whom had no data uploaded to the hospital information system on the pathogen causing the cellulitis and its antibiotic resistance and sensitivity. After applying all selection criteria, 99 patients met them. Thereafter, the secondary collection of data from medical documents and their analysis was performed (Figure 1).

Second stage of data collection

Of the 99 patients treated in the Hospital of Lithuanian University of Health Sciences (Kauno klinikos)for cellulitis of odontogenic origin in the facial and maxillofacial region, anonymized data on the patient's age, gender, general health status, duration of treatment in the ICU, total duration of inpatient treatment and complications were selected from the medical documents entered into the Hospital of Lithuanian University of Health Sciences (Kauno klinikos) information system. In addition, from the results of the microbiological culture of the pus entered into the information system, the primary pathogen of the disease, its resistance and sensitivity to antibacterial agents were selected and the patient was prescribed antibacterial therapy.

Statistical analysis

The data collected as part of the study was entered into the Microsoft Office Excel^® 2016 software (Microsoft Co.; Redmond, Washington, USA) and the statistical analysis of the data was carried out using the IBM SPSS Statistics 29.0 (IBM Corp.; Armonk, New York, USA). Descriptive statistics for quantitative variables include mean, standard deviation (SD), median, minimum, and maximum values. Frequencies and percentages are provided to describe qualitative variables. For the comparison of quantitative variables that did not meet the normality assumption (the Shapiro-Wilk and Kolmogorov-Smirnov criteria were used to check this assumption), a non-parametric Mann-Whitney U criterion was used to compare two independent samples (the results are described in the median [minimum; maximum]).

To assess the relationship between quantitative variables, Spearman's coefficient of moral correlation was used (when the variables did not meet the normality assumption). The difference was considered statistically significant if P < 0.05.

General characteristics of study sample

Of the 99 subjects, 61 men (61.6%) and 38 women (38.4%) were identified. The oldest person with the disease was 90 years old and the youngest was 20 years old. The average age of the subjects was 49.7 (SD 8.3) years. On average, the subjects spent 17.8 (SD 10) days in the inpatient ward and 6 (SD 6) days in the ICU before being discharged from the medical facility. An analysis of the course of the disease revealed that 95 subjects (96%) recovered, while 4 subjects (4%) died.

Primary pathogens

When analysing the primary pathogens identified in the pus, these were divided into groups. The group of streptococci includes: Streptococcus group F, group CFG, pyogenes, beta-hemolytic, viridans. For the coagulase-negative staphylococci: Staphylococcus haemolyticus, hominis. For the Proteus group - Proteus mirabilis and vulgaris, as well as Prevotella - P. buccae, P. species, P. oris and P. nigrescens.

In the pus cultures, streptococci (group F, group CFG, pyogenes, beta-hemolytic, viridans) were found as the most frequent primary pathogens of the purulent infection in 40 subjects (41.4%), Proteus mirabilis or Proteus vulgaris in 12 subjects (12.1%) and Peptostreptococcus in 10 subjects (10.1%). The bacterium Enterococcus faecium was found least frequently, namely in only 1 of the subjects.

Sensitivity of pathogens and antibiotic resistance

Based on the data from antibiograms and the analysis of the sensitivity and resistance of the pathogens to the most frequently administered empirically prescribed antibiotics (penicillin G, cefuroxime, metronidazole, cefazolin, ampicillin), it was found that the pathogens in the streptococcus group were 95% sensitive and only 5% resistant to penicillin. In the Prevotella group, just under half (n = 4 [44.4%]) of all microorganisms found were resistant to penicillin, and (n = 5 [55.6%]) were sensitive. Half of all pathogens in the Proteus group were resistant to cefuroxime (n = 6 [50%]), while the other half (n = 6 [50%]) were sensitive to it. Twenty percent (n = 2) of the Peptostreptococcus bacteria were resistant to metronidazole. Fusobacterium nucleatum was sensitive to metronidazole in all cases, but resistant to penicillin in 50% (n = 2) of cases. And S. aureus showed sensitivity to clindamycin in all cases (Table 1).

Empirically prescribed antibiotics

A combination of antibiotics is most frequently prescribed empirically, a combination of penicillin G and metronidazole (IV) in 58% of cases, a combination of cefuroxime and metronidazole (IV) in 13% of cases examined and a combination of cefazolin and metronidazole (IV) in 11 of cases examined. In the rarest cases, in only 1% of cases, empirical antibiotic therapy with vancomycin (n = 1) was prescribed (IV).

Constancy of antibiotic therapy

An examination of the subjects' antibiograms showed that 66 of the subjects (66.7%) were not prescribed empirical antibiotics, while the remaining 33 subjects were replaced with antibiotics: 10 of them (10.1%) according to the clinical course and 23 patients (23.2%) according to the results of microbiological cultures. Patients whose antibiotic therapy was not modified had significantly shorter treatment duration both overall (13 days) and in the ICU (2 days) compared to those whose antibiotics were changed (23 and 11 days, respectively). The total duration of treatment and the duration of treatment in the ICU were statistically significantly (P < 0.001) longer when the antibiotics were changed (Table 2).

Prescribed antibiotic and its effectiveness

After analysing both the empirical therapy and the change of prescribed antibiotics and their efficacy, the results showed that the empirically designed combination of penicillin G and metronidazole was effective in 61% of cases (n = 36) without changing to another antibiotic and ineffective in 37.3% of cases (n = 37.3). The empirically prescribed combination of cefuroxime and metronidazole was effective in 76.9% of cases (n = 10) and ineffective in 23.1% of cases (n = 3) when prescribed empirically. Ciprofloxacin, clindamycin, cefuroxime, sultamycillin and a combination of cefoperazone and sulbactam were effective as second-line antibiotics in all cases of their administration, although no empirical administration was prescribed (Table 3).

Pathogens and total duration of treatment in the ICU

The interaction between pathogens causing cellulitis and antibiotics used for their empirical treatment was analysed, taking into account the total duration and time of treatment of the patients in the ICU. After identifying the most common pathogen, which belonged to the Streptococcus group, a combination of penicillin and metronidazole was the most frequently prescribed empirically (n = 24 [58.5%]). The shortest total duration of treatment (10 days on average) and duration of treatment in the ICU (2 days on average) was determined by empirical administration of a combination of amoxicillin and clavulanic acid (n = 1 [50%]) (10 days on average). However, these data are not particularly reliable as this combination is only intended for one of the subjects studied. For all pathogen groups, the combinations of antibiotics with metronidazole were associated with a shorter overall duration, and a shorter duration of treatment in the ICU than the same antibiotic substances without metronidazole. And empirically, for both penicillin/metronidazole and cefazolin/metronidazole combinations, the infections caused by Streptococcus (n = 24 [58.5%] and n = 5 [12.2%]), coagulase-negative staphylococci, Prevotella (n = 4 [44.4%] and n = 3 [33.3%]), Proteus groups (n = 6 [50%] and n = 1 [8.3%]) and S. aureus, Peptostreptococci, Cutibacterium and S. marcescens, the shorter overall duration of treatment and longer time to treatment were the result of the shorter duration and time to ICU treatment observed with the cefazolin/metronidazole combination. On the contrary, when a combination of cefuroxime and metronidazole was used, the total duration of treatment and time to ICU treatment was longer than with empirical antibiotic therapy with penicillin and metronidazole (Table 4).

Age of patients and total duration of treatment in groups of different cellulitis

In total, 34.3% of the subjects were diagnosed with widespread cellulitis, 25.3% of the patients had submandibular cellulitis, 13.1% of the subjects had cellulitis of the floor of the mouth and submental area, 10.1% of the subjects had cellulitis of the pterygomandibular space, and 4% had cellulitis of the temporal area.

The highest age of patients was found in patients with floor of mouth cellulitis (median 55 years [range 31 to 86]) and the lowest in patients with submental cellulitis (median 35 years [range 20 to 79]). However, the duration of treatment did not differ statistically significantly (P ˃ 0.05) between the anatomical areas affected by cellulitis (Table 5).

Total duration of treatment across groups with different comorbidities

Forty six (46.5%) of the examined persons had one or more general diseases, while 53 (53.5%) of the examined persons had no comorbidities. Among the subjects, the majority had arterial hypertension (n = 28) and diabetes mellitus (n = 8). When analysing the correlation of comorbidities and total duration of treatment in the different groups of cellulitis, it was found that the correlation was not significant. The duration of treatment in hospital was statistically significantly longer for pulmonary diseases (P = 0.036) and diabetes mellitus (P = 0.049) than for these general pathologies (Table 6).

Relationship between complications and duration of treatment

Based on the receiver operating characteristic (ROC) analysis, we identified threshold values for total hospitalization duration (17.5 days) and ICU stay duration (3.5 days) in relation to the development of complications. If the patient's total hospitalization duration exceeds 17.5 days, the risk of developing complications significantly increases (95% CI [confidence interval]) (Figure 2).

The total duration of treatment (P < 0.001) and the duration of treatment in the ICU (P < 0.001) were significantly higher in patients with complications. If no complications occurred, the total duration of treatment was 13 (interquartile range [IQR] = 10 to 17) days, if they occurred - 24 (IQR = 17.3 to 29.5) days. If there were no complications, the length of stay in the ICU was 2 (IQR = 1 to 5) days and if there were complications, 10 (IQR = 6 to 18.8) days (Figure 2).

Cellulitis of odontogenic origin in the maxillofacial area is still a common inflammatory disease, both in Lithuania and worldwide. Rastenienė et al. [6] study, analysed the patients treated for facial and maxillofacial cellulitis at Žalgiris Clinic of Vilnius University Hospital in Lithuania, stated that this disease accounted for an average of 206 cases per year and affected about 0.05% of the total population of Lithuania. As the present study only included adult patients who were in intensive care and whose pus samples showed pathogen growth, the annual number of affected patients was significantly lower, averaging 19.8 cases per year.

In the presence of concomitant diseases that cause immunosuppression of the body, there is an increased risk of contracting infectious diseases [7]. According to a study by Rahimi-Nedjat et al. [8], the likelihood of developing cellulitis in diabetes mellitus is 1.28 times higher than in patients without immunosuppressive disorders. Stathopoulos et al. [9] argued that poorly controlled diabetes mellitus was complicated by necrotizing mediastinitis, and the results obtained showed that such patients are significantly more prone to develop cellulitis, with the infection being more difficult to control and spreading more rapidly (P < 0.01). In the present study, we found that patients with diabetes mellitus had statistically significantly longer hospitalization.

Most cellulitis in the oral and maxillofacial region is caused by polymicrobial microflora [10,11]. A study by Plum et al. [12], conducted in the New York City, USA, showed that the flora of odontogenic maxillofacial cellulitis is polymicrobial in adults and accounts for up to 70% of cases in children in 59.5% of cases. In the present study, individual bacterial strains were identified as the primary pathogens. van der Merwe et al. [7] argues that this could be due to the improper storage environment of pus during transportation of the microbiology specimen to the laboratory.

In the cases examined in the Hospital of Lithuanian University of Health Sciences (Kauno klinikos), the predominant pathogens, accounting for 41.4% of all cases, are group Streptococcus (viridans, group F, group CFG, pyogenes, beta-hemolytic). Similar studies from abroad, in which the pathogens causing cellulitis of odontogenic origin in the facial and maxillofacial region were examined, are also dominated by streptococcal strains, the frequency of which varies between 22.2% [13] and 71.26% [7,14-16]. There is controversy over Staphylococcus aureus as the causative agent of cellulitis, as Staphylococcus is capable of producing enzymes that stimulate the production of fibrin, forming well-encapsulated abscesses [12,17]. Nevertheless, the pathogen S. aureus was found in the cultures of the test subjects in 4% of cases.

The present study revealed that empirical antibiotic therapy is most often prescribed, a combination of penicillin and metronidazole (58%). Bhagania et al. [18] in the study say that the combination of penicillin and metronidazole, even with the growth of resistance, remains effective and clinically acceptable, with a set percentage of failure, with an empirical treatment of just 3.5%. Therefore, the guidelines for empirical antibacterial therapy in Lithuania remain similar. According to foreign researchers, clindamycin is often prescribed empirically, especially in patients allergic to penicillin. According to a study by Tancawan et al. [19], the success rate of penicillin antibiotics and clindamycin in the treatment of odontogenic purulent infections remains similar, 88.2% and 89.7%, respectively. However, many researchers report a growing resistance of microorganisms to clindamycin [4,20]. Kim et al. [14] compared the data for 2009 to 2014 with the data for 1997 to 2003 and noted that the resistance of clindamycin in the Streptococcus group increased, while in the Staphylococcus group decreased. In Lithuania, clindamycin remains a reserve antibiotic to prevent the formation of resistant microorganisms [6].

Several authors such as Bhagania et al. [18] and Sebastian et al. [21] point out that the use of metronidazole in combination with other antibiotics enhances the action of the antibacterial drug against anaerobes. Bali et al. [22] found that the sensitivity of anaerobic microorganisms to amoxicillin with clavulanic acid was lower (45%) than to metronidazole (75%) and indicated that in the presence of cellulitis of odontogenic origin, it is often appropriate to add metronidazole. Our study revealed similar results: combinations of antibiotics with metronidazole were associated with the average rates of shorter treatment durations and shorter ICU stays than with the same antibiotics without metronidazole.

When it is necessary to change antibiotic treatment, without obtaining an adequate response with empirical treatment, and when microorganisms resistant to empirically prescribed antibiotics are found, leads to an increase in the total duration of treatment in hospital and ICUs (P < 0.001). This is also reported by other researchers [10,14].

Limitations

The main limitation of the study was that the selection of test subjects was considerably restricted by the criteria for inclusion and rejection. Of the 710 cases examined, only 99 met the criteria for our study, which corresponds to only 13.9% of all cases. In 96 cases, the pus culture samples showed no growth of microorganisms or normal oral microflora with untested antibiotic sensitivity and resistance. In 23 cases, the medical records did not contain antibiograms. Similarly, most microorganisms were not tested for resistance/susceptibility to metronidazole as the most commonly used antibacterial agent. Due to the small selection, the groups of most of the microorganisms detected were extremely less, which may have influenced the results of the statistical analysis. Future research should be broader in scope and include more anaerobic and aerobic pathogens, as well as their antibiotic resistance and susceptibility characteristics in order to reveal a more complete model of empirical antibiotic therapy.

The most common pathogens of the Streptococcus group, which cause cellulitis of odontogenic origin in the maxillofacial area, are highly sensitive to penicillin and clindamycin, as well as to third-generation cephalosporins. Proteus mirabilis and Proteus vulgaris have high resistance to cefuroxime and ampicillin. Pathogens of the S. aureus, F. nucleatum and Prevotella group are highly resistant to penicillin.

The most effective empirical antibiotic therapy is the combination of cefazolin and metronidazole, as well as cefuroxime and metronidazole. The least effective empirical therapy – penicillin G. Combinations of antibiotics with metronidazole are associated with a shorter duration of treatment than without it.

The duration of treatment is prolonged by diabetes mellitus and pulmonary pathologies, the replacement of ineffective empirical antibiotic therapy with another, and the complications that have occurred. The duration of treatment, both in the hospital and in the intensive care unit, does not differ depending on the pathogen.