This study demonstrated that the most frequently observed abscess sites in DNI were the lateral cervical and submandibular regions. However, it should be noted that 35 patients with peritonsillar abscess were excluded from the study because of unavailable blood tests. This was the reason for the low rate of peritonsillar abscess in our series. An etiology could be identified in 71.4% of patients, with tonsillopharyngitis (19,5%) and lymphadenitis (18,2%) being the most common causes. In single-region abscess cases, the lateral cervical region was the most affected site. Abscesses caused by cysts or lymphadenitis were exclusively limited to single regions without any complication, whereas those of cancer and odontogenic etiologies had higher complication rates.

The frequency of odontogenic infections (9,1%) was found to be lower in our study compared to the literature. This finding may be attributed to the relatively low prevalence of chewing tobacco habits in Turkey, which was the leading cause of odontogenic infections in other studies [11, 12]. While the prevalence of chewing tobacco is significantly high in South Asian countries (20–30%), it is considered negligible in Turkey [13, 14]. In our study, patients with odontogenic infections showed a complication rate of 28.6%. This result underscores the importance of close monitoring in odontogenic infections to reduce complication risks and hospitalization duration. Initiating secondary group broad-spectrum antibiotic combinations early, especially in these cases, could improve clinical outcomes.

The SII index was recently developed as an indicator to reflect the balance of the inflammatory response and immune status of the host. Malignancies, underlying immunosuppression, age, usage of antibiotics are all important variables that may directly influence SII index levels. However, it is important to note that only three patients in our study had malignancy. With multivariate regression analysis, although an interaction was observed between only age and SII index, no relationship was observed between age and other variables such as the presence of complications, duration of treatment, extent of abscess, and etiology. Apart from this, there were no patients with immune deficiency among our patients and since all patients received antibiotic treatment, they were on equal terms in this care. For these reasons, we think that the main factor determining the SII index cut-off value according to ROC analysis is the DNI. Our study aimed to evaluate the potential utility of the SII index in infectious diseases—particularly in clinical settings where diagnostic limitations are present—as a prognostic or triage tool, and to assess its predictive value in identifying complication risk in DNI. In this regard, we would like to emphasize that the SII index should not be interpreted in isolation, but rather within a comprehensive and integrative clinical framework that includes primary diagnostic findings, imaging results, and clinical follow-up data.

In our study, 29.9% of our patients were found to have abscesses in two or more regions (Group 2). Patients in Group 2 had significantly higher SII index values, prolonged treatment durations, and extended hospital stays. Notably, complications were observed in Group 2, with a rate of 30.4%, while no complications were observed in Group 1 (single region abscess). Lin et al., reported that multi-region abscess and gas detection on CT were independent predictors of complications [11]. These findings highlight the critical importance of early identification and management of multi-region abscesses to predict complication risks and treatment trajectories. In cases where adequate imaging methods are not available, it may be difficult to understand the extent of the abscess. In our study, an SII threshold of 1616.5 was determined to predict multi-region abscesses with 65.2% sensitivity and 75.9% specificity. The SII index is an easy-to-use test developed as a marker reflecting the balance between the host’s inflammatory response and immune status and elevated SII should serve as red flags for serious multi-region infections [8*]. Therefore, it is a useful tool in the initial evaluation of DNI patients in terms of predicting the severity of the disease.

In our cohort, patients with complications had significantly higher mean SII index values (5302.4) compared to those without complications (1514.1). These patients also experienced longer hospital stays (19.7 vs. 5.4 days). Similar to our findings, González et al. reported an SII threshold of 2975 for predicting mediastinitis, tracheotomy need, and mortality in DNI patients [8*]. Early prediction of complications using biomarkers like SII index can facilitate timely intervention, potentially reducing morbidity, mortality, and healthcare costs.

In our study, only 23.4% of patients had positive culture results, with Staphylococcus aureus being the most frequently isolated pathogen. In our hospital, anaerobic culture antibiogram cannot be performed, only aerobic and facultative aerobic pathogens are identified with an automated system and an antibiogram is performed. However, we acknowledge that the inability to perform anaerobic culture and susceptibility testing in our center represents a diagnostic limitation. This deficiency compels infectious disease specialists to rely more heavily on clinical judgment in the diagnostic process and contributes to the widespread use of treatment strategies with empirical broad-spectrum antibiotic combinations. We would like to emphasize that the majority of patients were managed with combination antibiotic regimens, with antibiotic selection individualized based on clinical presentation, radiological findings, and patient-specific risk factors. The low rate of culture positivity may be attributed to high incidence of prior oral antibiotic use or the involvement of polymicrobial or anaerobic pathogens. This finding highlights the need for advanced diagnostic techniques, such as molecular and serological methods, to identify elusive pathogens [15]. Epidemiological determination of resistance patterns is also very important in terms of selecting effective empirical treatment in infections with high mortality risk. Beka et al. emphasized the importance of using molecular analyses in addition to classical methods in selecting empirical treatment; in their article, the detection of bacterial DNA in 33 samples that did not show growth in culture, in particular, demonstrated the advantage of faster diagnosis with molecular methods [16]. In our study, we observed that based on the patients’ epidemiological histories and sociocultural backgrounds, tuberculosis cultures were performed, and three patients were identified as having tuberculous lymphadenitis as an etiological factor. Considering the significant migration movements that our country has experienced in recent years, tuberculosis has been increasingly scrutinized as a potential underlying cause of DNIs. However, no drug resistance and complications were detected in the tuberculosis cases included in our study.

In our study, the majority of the treatment regimens were pPGA combination therapies, commonly used in community-acquired infections and generally considered as first-line empirical treatment options. However, in cases involving resistant pathogens (e.g., Methicilline resistant staphylococcus aureus, anaerobic microorganisms) or infections with extensive and complicated multi-region involvement, the effectiveness of these agents may be limited. Therefore, it is plausible that the therapeutic efficacy of pPGA-based regimens was suboptimal in certain patients. In our study, treatment regimens included in the pSGA therapy never involved the use of narrow-spectrum antibiotics as monotherapy. Acknowledging the limitations in diagnostic capabilities, our treatment strategy was predominantly based on combination regimens. These regimens were particularly preferred in cases of multi-region abscessess or when there was a high risk of complications, with the aim of enhancing clinical efficacy. Furthermore, treatment decisions were not static but dynamically adjusted based on clinical and laboratory findings. In cases of poor response to initial empirical therapy, or upon receipt of culture and susceptibility results, treatment regimens were re-evaluated and modified on an individual basis. This approach was intended to ensure that each patient received the most appropriate, effective, and targeted therapy based on their specific clinical condition.

It has been reported that synergism of microorganisms in polymicrobial abscess formation can cause gaseous infections and lead to more serious complications in the clinical course of patients [17*]. In a study on bacteriology, Huang et al. detected 35.9% polymicrobial agents in 128 DNI patients and reported the anaerobic culture rate as 59.3% in odontogenic infection [18]. Vavro et al. reported that the highest resistance in their antibiograms was determined to be against Clindamycin, and that resistance to Metronidazole, Penicillin and Amoxicillin-Clavulanate was also quite high [19*]. In our study, patients initially treated with pPGA combinations were transitioned to pSGA combinations in 38% of cases. Because of our low rate of positive culture test results, most of the changes were due to inadequate response to empirical antibiotic therapy administered within 48 h. The high rate of treatment escalation emphasizes the importance of initiating secondary group broad-spectrum combination therapy in DNI cases, especially those with odontogenic etiologies involving multiple abscess sites.

Limitations of the study should be considered. This study had a retrospective design and was conducted in a single center. The low rate of positive culture, the inability to perform anaerobic culture and susceptibility testing in our center are also limitations in evaluating the effect of microbiological factors. However, to our knowledge, this is the first study to assess the prognostic value of clinical and serological factors in conjunction with antimicrobial therapy in surgically drained deep neck infections that failed to respond adequately to broad-spectrum empirical combination therapies.