Sage Journals: Discover world-class research

Abstract

Objectives

Deep neck infections (DNIs) are a common and intractable disease encountered in ENT clinics that impose a significant medical and financial burden on affected individuals and their families. However, insufficient data are currently available for predicting outcomes in cases of DNI. The present study thus sought to develop a novel model capable of predicting treatment outcomes of DNI patients just using indicators at the visit.

Methods

Patients with DNIs treated from 2010 to 2022 were included in the present study. Patient data were retrospectively collected from medical records. Risk factors associated with mortality were identified using logistic regression models. A predictive model was constructed based on odds ratios for factors calculated using a multivariate regression model.

Results

In total, 153 patients were enrolled in the present study. Risk factors associated with mortality included age >50 years, residence in a rural area, dyspnea at visit, the involvement of multiple infected sites, serum albumin<34 g/L, renal insufficiency, mediastinitis, pulmonary infection, and septic shock. A multivariate regression model revealed that mediastinitis (OR: 7.308, P < 0.001), serum creatinine>95 μmol/L (OR: 23.363, P < 0.05), and serum albumin<34 g/L (OR: 13.837, P < 0.05) were independent predictors of mortality in deep neck infection patients, with serum creatinine>95 μmol/L being particularly critical to the outcomes. Diabetes was not the predictor of mortality but was associated with long-term hospitalization (P < 0.001).

Conclusions

In summary, the model constructed in the present study was capable of estimating the potential for poor outcomes in DNI patients before the initiation of treatment. These findings may help improve doctor–patient communication, especially for those struggling financially.

Keywords

deep neck infection predictive model mortality malnutrition renal insufficiency mediastinitis

Introduction

Deep neck infections (DNIs) refer to cellulitis or abscesses that develop in the deep cervical fascial space. These infections are commonly encountered in ENT clinical practice and can arise from progressive infections of the tonsils, odontogenic structures, pharynx, esophagus, trachea, or salivary glands. The neck contains several cavities and fascial recesses, including the retropharyngeal, peritonsillar, masseteric, pterygopalatine maxillary, parapharyngeal, and submandibular spaces, all of which are interconnected and filled with soft connective tissue, adipose tissue, muscles, blood vessels, and nerves. DNIs can spread to several of these tissue sites along interconnected soft tissues, potentially resulting in severe symptoms and poor health outcomes. Common presenting symptoms in DNI patients include odontalgia/pharyngalgia/cervicodynia, fever, dyspnea, restricted mouth opening, and the presence of a cervical mass.

Deep neck infections have the potential to induce serious complications. For example, infections involving the pre-tracheal, retropharyngeal, or parapharyngeal spaces can compress the airway and cause respiratory tract obstruction. Moreover, when infections spread downward into the mediastinum, which is also filled with soft connective tissue, mediastinitis can result, causing chest pain and fever. Descending necrotizing mediastinitis, which is often a consequence of the downward spread of deep neck infections, can result in high mortality rates of 10–30%.^1-3 When these infections continue progressing, pericarditis and empyema may result, and bacterial toxins can enter the bloodstream, potentially causing life-threatening septicemia and multiple organ failure.

The rapid onset and progression of DNIs and their potential for high morbidity and mortality rates can impose significant physical, psychological, and economic burdens on affected patients and their families, in addition to placing a substantial financial burden on the government as a whole⁴. Prior evidence suggests that the duration of hospitalization for DNIs is generally approximately 2 weeks,^5,6 with the exact duration being dependent upon whether complications develop and whether the infection involves multiple spaces within the neck,^5,7 although these findings are not universal.⁸ Overall DNI mortality rates range from 0 to 2.7%,^4,9,10 with death often coinciding with the presence of significant co-morbidities. However, few studies have explored the prediction model for mortality before the treatment.

The potential for mortality may be an important consideration for patient decision-making. The present study was thus designed to construct a model capable of predicting the poor outcomes before treatment in patients with DNIs to promote better communication between doctors and patients in daily clinical practice.

Patients and methods

Study participants and ethical approval

The medical records of 198 DNI patients were examined between 2010 and 2022 in the Otorhinolaryngology Department of the first affiliated hospital of Anhui Medical University for the current retrospective analysis. Those patients diagnosed with cervical abscess/cellulitis were included in this study. Patients were excluded from these analyses when sufficient data were not available. The First Affiliated Hospital Ethics Committee of Anhui Medical University approved this study. Figure 1 outlines the general study design in detail.

Figure 1.

Enrollment, data collection, risk factor screening, and model construction approaches.

Of the 198 evaluated patients, 153 were enrolled in the final study, 48 were ruled out because insufficient data. Treatment approaches (pharmaceutical intervention, drainage, tracheotomy, etc.) were based on clinical findings. Variables extracted from patient medical records included demographic information, hospitalization duration, symptoms, laboratory examinations, radiological examinations, and co-morbidities corresponding to patient condition at presentation.

Statistical analysis

Continuous variables are given as means with standard deviations when normally distributed or medians with interquartile ranges when not normally distributed and were compared via Mann–Whitney U-tests. Categorical variables are given as frequencies with percentages and were compared via X² tests.

Logistic regression models were utilized to explore the relationship between risk factors and mortality, with odds ratios (ORs) and 95% confidence intervals (CIs) calculated for each variable. Those significant risk factors in univariate logistic regression analyses were then incorporated into multivariate logistic regression models, with stepwise regression being used to filter out significant factors and adjusted ORs being calculated accordingly. Receiver operating characteristic (ROC) curves were then constructed and used to identify optimal cutoff points associated with predictors of high treatment costs. Area under the curve (AUC) values were additionally calculated for each predictor. Cutoff values for continuous variables were selected based on the maximal AUC. Model calibration was performed with the Nagelkerke R². A P-value of < 0.05 was the significance threshold for this study, and all statistical analyses were conducted using SPSS 20.0 (IBM, NY, USA).

Results

Patient characteristics

In total, 153 patients (87 male, 66 female) were enrolled in this retrospective study, with a median age of 48.0 years (range: 1–89 years). The mean duration of hospitalization for these patients was 17 days, the mean white blood cell count in the overall patient population was 14.5 ± 1.1×10³ cells/mm³, and the mean neutrophil ratio was 79.1 ± 2.4%. The mean serum creatinine was 72.0 ± 14.6 μmol/L, and the mean serum albumin (ALB) was 37.0 ± 2.5 g/L.

Of these 153 patients, 37 (24.2%) reported smoking and 29 (19.0%) reported drinking daily. For co-morbidities, 57 (37.3%) had diabetes, and 30 (19.7%) were previously diagnosed with high blood pressure. At the time of hospital presentation, 62 (40.8%) patients had a fever, and 32 (21.1%) were experiencing dyspnea. Imaging analyses revealed infections of the parapharyngeal space (45.8%), retropharyngeal space (15.7%), submandibular space (45.1%), sternocleidomastoid region (69.3%), and masseteric space (6.5%) in these patients, of whom 49.7% exhibited infections of multiple spaces. A subset of these patients additionally presented with serious complications at the time of hospitalization, including mediastinitis (13.1%), septic shock (5.9%), and pulmonary infection (6.5%), with 37.5% of patients requiring immediate tracheotomy (Table 1).

Table 1.

Demographics of deep neck infection patients at the time of clinical presentation.

Characteristic	All, n = 153	Cured, n = 138	Not cured, n = 15	P-value
Gender, male	87 (56.9%)	76 (55.1%)	11 (73.3%)	0.139
Age (years)	48.0 (18.0-78.0)	46.5 (14.5-78.5)	63.0 (38.0-88.0)	0.006*
Place of residence, rural	92 (60.1%)	78 (56.5%)	14 (93.3%)	0.005*
Length of stay (days)	17.3 (15.1-19.4)	17.8 (15.2-19.2)	17.9 (4.7-31.2)	0.836
Smoking	37 (24.2%)	33 (23.9%)	4 (26.7%)	0.759
Alcohol drink	29 (19.0%)	25 (18.1%)	4 (26.7%)	0.486
BMI	21.5 (20.5-22.5)	21.6 (20.5-22.6)	20.4 (16.7-24.1)	0.661
Diabetes	57 (37.3%)	48 (34.8%)	9 (60%)	0.089
High blood pressure	30 (19.7%)	24 (17.5%)	6 (40%)	0.079
Symptoms
Fever (>37.3°C)	62 (40.8%)	57 (41.6%)	5 (33.3%)	0.592
Dyspnea	32 (21.1%)	25 (18.2%)	7 (46.7%)	0.018*
Blood test
WBC (×10⁹/L)	14.5 (13.4-15.6)	14.7 (13.5-15.8)	12.9 (9.2-16.5)	0.346
Neutrophil ratio (%)	79.1 (76.7-81.6)	79.3 (77.0-81.5)	78.0 (63.2-92.8)	0.762
GFR, mL/(min×1.73 m²)	114.9 (108.3-121.5)	120.3 (114.5-126.1)	64.4 (36.0-92.9)	<0.001*
Creatinine (μmol/L)	72.0 (57.3-86.6)	65.7 (50.5-80.9)	126.2 (76.5-175.9)	0.013*
BUN (mmol/L)	6.8 (6.0-7.6)	6.2 (5.5-6.9)	12.1 (7.8-16.5)	<0.001*
Albumin (g/L)	37.0 (34.4-39.5)	38.3 (35.5-41.1)	25.6 (22.3-28.9)	0.003*
ALT (U/L)	38.7 (32.2-45.1)	38.9 (32.0-45.8)	36.9 (18.5-55.2)	0.852
Bicarbonate (mmol/L)	23.3 (22.6-24.0)	23.6 (22.9-24.4)	20.6 (17.8-23.5)	0.009*
Blood glucose (mmol/L)	8.7 (7.8-9.6)	8.7 (7.7-9.7)	8.7 (6.6-10.7)	0.977
Infection location in CT
Parapharyngeal space	70 (45.8%)	60 (43.5%)	10 (66.7%)	0.106
Retropharyngeal space	24 (15.7%)	19 (13.8%)	5 (33.3%)	0.062
Submandibular space	69 (45.1%)	58 (42.0%)	11 (73.3%)	0.028*
Sternocleidomastoid region	106 (69.3%)	93 (67.4%)	13 (86.7%)	0.151
Masseteric space	10 (6.5%)	8 (5.8%)	2 (13.3%)	0.255
Multiple space infection	76 (49.7)	63 (45.7%)	13 (86.7%)	0.003*
Co-morbiditis
Mediastinitis	20 (13.1%)	13 (9.4%)	7 (46.7%)	0.001*
Septic shock	9 (5.9%)	3 (2.2%)	6 (40%)	<0.001*
Pulmonary infection	10 (6.5%)	5 (3.6%)	5 (33.3%)	0.001*
Requiring tracheotomy	57 (37.5%)	47 (34.1%)	10 (71.4%)	0.009*

* denotes a significant difference between groups of patients that were not successfully cured. BMI: Body mass index; WBC: White blood cell; GFR: Glomerular filtration rate; BUN: Blood urea nitrogen.

Factors associated with mortality

Of the 153 patients involved in this trial, 143 underwent successful treatment and recovered or stabilized, while the remaining 15 underwent unsuccessful treatment and eventually worsened or passed away.

The average age of cured patients was 44.9 ± 3.5 years, significantly younger than those of uncured ones (60.4 ± 9.0 years). Of the cured patients, 78 lived in rural areas (56.5%), whereas 14 of the uncured patients (93.3%) were from rural areas, with a significant difference in these rates between groups (P < 0.05). Moreover, 46.7% of uncured patients presented with dyspnea, which is significantly lower than cured patients (18.2%). The serum creatinine (126.2 ± 49.7 g/L compared to 65.7 ± 15.2 g/L) and blood urea nitrogen (BUN) (12.1 ± 4.4 g/L compared to 6.2 ± 0.7 g/L) were significantly higher among uncured ones compared to cured ones.

However, the serum albumin (25.6 ± 3.3 g/L compared to 38.3 ± 2.8 g/L) and bicarbonate (20.6 ± 2.9 g/L compared to 23.6 ± 0.7 g/L) were significantly lower in the uncured group. In addition, rates of mediastinitis, submandibular space infection, the infection of multiple spaces, septic shock, and pulmonary infection were all higher among uncured patients relative to cured patients (Table 1).

Factors associated with long-term hospitalization

Mortality was not the only parameter used to assess the DNI's prognosis. The length of hospital stay for those who survived reveals the severity of the condition. For the present study, DNI patients were stratified based upon the length of hospital stay into a short-term group (< 2 weeks) and a long-term expense group (> 2 weeks). Patients in the long-term group were elder (52.1 ± 3.9 years compared to 37.3 ± 5.4 years, P < 0.001) and heavier in weight (60.4 ± 3.6 years compare to 53.6 ± 3.5 kg, P < 0.001) than those in the short-term group. Approximately 17.9% of patients in the short-term group presented with diabetes, whereas these rates were significantly higher in the long-term group (50.7%, P < 0.001). Just 9% of patients in the short-term group exhibited dyspnea at the time of presentation, compared to 27.1% of patients in the long-term group, with these latter rates being significantly higher (P = 0.007).

At the same time, the white blood cell count (WBC) (16.6 ± 1.8 × 10⁹/L compared to 12.7 ± 1.4 × 10⁹/L, P = 0.001), neutrophil ratio (84.0 ± 2.3% compared to 74.3 ± 3.7% P < 0.001), glomerular filtration rate (GFR) [111.6 ± 6.8 mL/(min × 1.73 m²) compared to 131.4 ± 9.0 mL/(min×1.73 m²), P = 0.001], ALT (46.4 ± 10.9 U/L compare to 30.3 ± 7.9 U/L, P = 0.02), and blood glucose at visit (10.6 ± 1.6 mmol/L compare to 6.4 ± 0.9 mmol/L, P < 0.001) were significantly higher in the long-term group. Rates of mediastinitis, Parapharyngeal space infection, the infection of multiple spaces, Gas formation, and requiring tracheotomy were also significantly higher in the long-term group relative to the short-term group (Table 2).

Table 2.

Factors associated with long-term hospitalization (Stay > 2 weeks).

Characteristic	Stay < 2 weeks, n = 67	Stay > 2 weeks, n = 71	P-value
Age	37.3 (31.9-42.7)	52.1 (48.2-56.0)	<0.001*
Weight	53.6 (49.1-58.1)	60.4 (56.8-64.0)	0.02
Diabetes	12 (17.9%)	36 (50.7%)	<0.001
Symptoms
Dyspnea	6 (9.0%)	19 (27.1%)	0.007
Blood test
WBC (×10⁹/L)	12.7 (11.3-14.1)	16.6 (14.8-18.4)	0.001
Neutrophil ratio	74.3 (70.6-78.0)	84.0 (81.7-86.3)	<0.001
GFR, mL/(min×1.73 m²)	131.4 (122.4-140.4)	111.6 (104.7-118.4)	0.001
BUN (mmol/L)	5.3 (4.5-6.0)	7.1 (6.0-8.2)	0.01
ALT (U/L)	30.3 (22.4-38.2)	46.4 (35.5-57.3)	0.02
Blood glucose (mmol/L)	6.4 (5.5-7.3)	10.6 (9.1-12.2)	<0.001
Infection location in CT
Parapharyngeal space	23 (34.3%)	37 (52.1%)	0.04
Multiple space infection	18 (26.9%)	45 (63.4%)	<0.001
Gas formation	3/11 (37.3%)	11/14 (78.6%)	0.017
Co-morbiditis
Mediastinitis	0 (0)	13 (18.3%)	<0.001
Requiring tracheotomy	5 (7.5%)	42 (59.2%)	<0.001

*denotes a significant difference WBC: White blood cell; GFR: Glomerular filtration rate; BUN: Blood urea nitrogen.

The development of a mortality prediction model

To establish a model capable of predicting mortality for patients at the time of clinical presentation, we incorporated the factors identified above into a logistic model. Factors in the univariate model included age >50 years (OR 4.903), rural residence (OR 10.768), dyspnea (OR 3.920), GFR <94 mL/(min×1.73 m²) (OR 33.250), creatinine >95 μmol/L (OR 28.571), BUN >7.86 mmol/L (OR 14.274), parapharyngeal space infections (OR 6.067), ALB <34 g/L (OR 30.049), Bicarbonate <18.8 mmol/L (OR 11.844), submandibular space infections (OR 3.793), the infection of multiple spaces (OR 7.738), mediastinal infections (OR 8.413), septic shock (OR 30.00), pulmonary infections (OR 13.300), and requiring tracheotomy (OR 4.840).

When these factors were incorporated into a multivariate model, only creatinine >95 μmol/L, ALB <34 g/L and mediastinal infections were retained for the final predictive model (Table 3). This model was constructed based on the factors we screened. Under this model, condition 1 applied to patients with creatinine >95 μmol/L, while condition 2 applied to patients with mediastinitis and ALB <34 g/L (both match) (Figure 2).

Figure 2.

The prediction model for mortality in deep neck infection patients. Condition 1 applies to patients with creatinine >95 μmol/L, while condition 2 applied to patients with mediastinitis and ALB <34 g/L (both match).

Based on condition 1 and condition 2, patients were stratified into three groups. In the Recoverable group (neither condition 1 nor condition 2 fits), 98.4% of patients survived. While in the High-risk group (either condition 1 or condition 2 fits), roughly half of the patients eventually deteriorated or died, and in the Fatal group (both of the 2 conditions fit), two-thirds of these patients died (Figure 3). The area under the ROC curve for this model is 0.896, with an 86.67% sensitivity and a 92.03% specificity. The Nagelkerke R² value is 0.552 for this model.

Figure 3.

The distribution of mortality rate in different groups of the model, the number in the column means how many cases in each group.

Discussion

Reports of DNI rates have been rising in clinical contexts in both Western¹¹ and Eastern countries,¹² often resulting in severe, life-threatening complications, prolonged hospitalization, and high mortality rates. In the current study, we investigated the risk factors for fatalities, and extended hospital stays in DNI patients, establishing a prediction model that can estimate the likelihood of fatalities depending on particular baseline parameters. This predictive model is capable of evaluating treatment outcomes of DNI, just relying on some clinical factors which are available before treatment is initiated, and its use has the potential to contribute to more reliable communication between doctors in their patients, aiding informed medical decision-making.

Table 3.

Odds ratios for mortality-related risk factors.

	One variable model		Multiple independent variable models
	Odds ratio	P-value	Odds ratio	P-value
Age>50 years	4.903 (1.324-18.151)	0.017
Place of residence, rural	10.768 (1.377-84.198)	0.024
Dyspnea	3.920 (1.301-11.813)	0.015
GFR<94 mL/(min×1.73 m²)	33.250 (5.882-187.95)	<0.001
Serum Creatinine>95 μmol/L	28.571 (7.392-110.436)	<0.001	23.363 (3.924-139.682)	<0.001
BUN > 7.86 mmol/L	14.274 (4.147-49.129)	<0.001
Serum albumin <34 g/L	30.049 (3.821-236.311)	0.001	13.837 (1.569-122.177)	0.015
Serum Bicarbonate <18.8 mmol/L	11.844 (3.61-38.864)	<0.001
Submandibular space infection	3.793 (1.150-12.509)	0.029
Multiple space infection	7.738 (1.682-35.590)	0.009
Mediastinal infection	8.413 (2.627-26.949)	<0.001	7.308 (1.473-36.252)	0.015
Septic shock	30.00 (6.422-140.138)	<0.001
Pulmonary infection	13.300 (3.292-53.734)	<0.001
Requiring tracheotomy	4.840 (1.441-16.262)	0.011

The optimal cutoff line of age, GFR, Creatinine, BUN, albumin and Bicarbonate was based on a receiver operating characteristic (ROC) curve. GFR: Glomerular filtration rate; BUN: Blood urea nitrogen.

The model developed in this study was established through multivariate logistic regression analysis and incorporated three factors: mediastinal infection status, ALB <34 g/L and creatinine >95 μmol/L. This model was constructed using the ORs associated with these variables, among which creatinine >95 μmol/L was found to be a key determinant of the final overall treatment cost for DNI patients.

First, Patients with mediastinitis have been previously reported to exhibit high mortality rates of 10-30%.^1-3 Severe mediastinitis can induce tissue necrosis, contributing to subsequent septicemia and multiple organ failure that can rapidly lead to death without sufficient treatment. In our cohort, 35% of the DNI patients with mediastinitis turned out dead, confirming the close relationship between mediastinitis and mortality.

Secondly, our present research highlights the role of renal insufficiency in affecting DNI patients’ treatment outcomes. Previous research has reported a higher incidence rate of DNIs in early-stage renal disease patients.^13,14 In a study on the correlation between the mortality rate of DNI and renal insufficiency, the result showed that the mortality rate of patients with DNI in the early-stage renal disease group was more than 2-fold higher than that in the control group. A recent study in 2021 supports this standpoint,¹⁵ but no specific cutoff point values of kidney function indicators were set up for high-risk DNI patients. However, Park et al¹⁶ declared that renal failure is a risk factor for necrotizing fasciitis, mediastinitis, and CRP >20 mg/dL but not for clinical outcomes.

Our current investigation proved that renal insufficiency is fatal for DNI patients. We set up a cutoff value of serum creatinine for high-risk DNI, with more than half of the patients finally dying if the serum creatinine is more than 95 μmol/L at the visit. This might be because patients with renal insufficiency may have poor immunologic function, causing infections to spread. Furthermore, when antibiotics are poured in, the kidney must bear a significant weight of drug metabolism, causing severe impairment to kidney function.

Thirdly, our research indicated that low serum albumin is also an important factor for poor treatment outcomes, which was seldom mentioned in previous research. Park et al¹⁶ reported that serum albumin level <3.0 g/dL contributed to a higher rate of simultaneous mediastinitis, necrotizing fasciitis, greater extent of infection, higher initial CRP, and delayed treatment response, and longer duration of hospitalization. Our report herein strengthens the role of hypoalbuminemia in decreasing the survival rate of DNI. Albumin is thought to play an important role in mediating inflammatory response¹⁷ and; hypoalbuminemia could be both the reason and result of systemic inflammation.

The present study explored other risk factors associated with poor patient prognosis. The analyses revealed that age >50 years, residence in a rural area, dyspnea at time of presentation, submandibular space infection, the involvement of multiple sites, serum bicarbonate <18.8 mmol/L, septic shock, and pulmonary infections at the visit were all associated with patient outcomes. In our cohort, age >50 was a predictor of poor outcomes. The negative impact of advanced age on prognosis may be related to more complications in elderly people, but not to age itself.¹⁸ Residence in a rural area is an interesting predictor we found.

According to the previous reports,⁶ almost all DNI patients were of low socioeconomic status, with roughly 70% being illiterate. Many of these patients also exhibited poor oral hygiene and presented with dental infections. Primary healthcare facilities are also lacking in rural areas, so treatment may be delayed when infections arise in these regions, leading to further deterioration. The finding that infections with multiple sites of involvement,^10,19 and dyspnea^13,20 are associated with greater risk is consistent with other reports indicating the presence of a wide range of lesions at the time of treatment initiation, potentially complicating treatment efforts. Patients with septic shock²¹ have also previously exhibited high mortality rates, which is now confirmed in our study. However, septic shock is relatively rare compared to mediastinitis.

It is worth mentioning that diabetes is not the risk factor for mortality, according to our research, but would prolong the length of hospital stay, which is in line with previous reports.^22,23 One recent study reported the average postoperative hospitalization duration for DNI patients with and without diabetes was 7 and 3 days, respectively.²² Additional treatments to normalize blood sugar levels can contribute to prolonged hospitalization duration. Mediastinitis is a life-threatening complication of such infections and may necessitate prolonged observation in the ICU. In addition, the mediastinum is located in a deep, low location that can be difficult to drain thoroughly, further prolonging the length of hospitalization. Tracheotomy is often critical to managing DNIs, particularly in patients with airway obstruction. Older age, multiple infected spaces, mediastinitis, and gas formation have all been identified as tracheotomy-related risk factors in DNI patients in the past.^19,24 These characteristics are frequently linked to a more severe clinical presentation, which could lengthen the hospital stay and call for more difficult medical procedures.

The present study has several advantages. For instance, all patients were treated in the same tertiary hospital using the same medications and surgical equipment, reducing the possibility of treatment approach-related bias. Moreover, the sample size was sufficient to permit robust statistical analyses. In our study, we set up a series of the cutoff value for mortality-related continuous variables, which could be used immediately for daily practice. Even so, there are some limitations to this study. For one, all data were collected retrospectively from medical records, which has the potential to introduce certain inaccuracies. Secondly, patients were from a single center, resulting in the potential for selection bias towards more severe or complicated cases.

Conclusions

In summary, the present study successfully identified risk factors associated with poor prognosis for patients suffering from deep neck infections. The predictive model developed herein may aid in efforts to forecast the potential for mortality before the therapeutic intervention, thereby enabling better communication between doctors and patients, particularly for patients faced with economic hardship.

Footnotes

Acknowledgments

The authors would like to thank the Medical Records and Statistics Room of the First Affiliated Hospital of Anhui Medical University for its technical support and statistical advice

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Ethical approval

This study was approved by the Ethics Committee for Research and Publication of The First Affiliated Hospital of Anhui Medical University. Written informed consent was waived owing to the use of deidentified retrospective data.

ORCID iD

Yehai Liu

Abbreviations

References

Ridder

Maier

Kinzer

Teszler

Boedeker

Pfeiffer

. Descending necrotizing mediastinitis: contemporary trends in etiology, diagnosis, management, and outcome. Ann Surg. 2010;251:528-534.

Prado-Calleros

Jimenez-Fuentes

Jimenez-Escobar

. Descending necrotizing mediastinitis: Systematic review on its treatment in the last 6 years, 75 years after its description. Head Neck. 2016;38(Suppl 1):E2275-E2283.

Palma

Giuliano

Cracchiolo

, et al. Clinical features and outcome of patients with descending necrotizing mediastinitis: prospective analysis of 34 cases. Infection. 2016;44:77-84.

Hurley

Douglas

Montgomery

Clark

. The hidden cost of deep neck space infections. Ann Royal College Surg England. 2018;100:129-134.

Boscolo-Rizzo

Marchiori

Zanetti

Vaglia

Da Mosto

. Conservative management of deep neck abscesses in adults: the importance of CECT findings. Otolaryngol Head Neck Surg. 2006;135:894-899.

Agarwal

Sethi

Mrig

Chopra

. Role of socioeconomic factors in deep neck abscess: A prospective study of 120 patients. Br J Oral Maxillofacial Surg. 2007;45:553-555.

Brito

Hazboun

Fernandes

, et al. Deep neck abscesses: study of 101 cases. Brazil J Otorhinolaryngol. 2017;83:341-348.

Bal

Unal

Delialioglu

Oztornaci

Ismi

Vayisoglu

. Diagnostic and therapeutic approaches in deep neck infections: an analysis of 74 consecutive patients. Brazil J Otorhinolaryngol. 2020;88(4):511-522.

Adovica

Veidere

Ronis

Sumeraga

. Deep neck infections: review of 263 cases. Otolaryngol Polska. 2017;71:37-42.

10.

Staffieri

Fasanaro

Favaretto

, et al. Multivariate approach to investigating prognostic factors in deep neck infections. Euro Arch Oto-rhino-laryngol. 2014;271:2061-2067.

11.

Martinez Pascual

Pinacho Martinez

Friedlander

Martin Oviedo

Scola Yurrita

. Peritonsillar and deep neck infections: A review of 330 cases. Brazil J Otorhinolaryngol. 2018;84:305-310.

12.

Yang

Xirasagar

Cheng

Kao

Lin

. A nationwide population-based study on the incidence of parapharyngeal and retropharyngeal abscess-a 10-year study. Int J Environ Res Public Health. 2021;18(3):1049.

13.

Wang

Kuo

Tsai

Huang

. Characterizations of life-threatening deep cervical space infections: a review of one hundred ninety-six cases. Am J Otolaryngol. 2003;24:111-117.

14.

Huang

Liu

Chen

Tseng

Yeh

Chen

. Deep neck infection: Analysis of 185 cases. Head Neck. 2004;26:854-860.

15.

Tsai

Yang

Huang

, et al. Pathogens and prognosis of deep neck infection in end-stage renal disease patients. Laryngoscope. 2022;132:1403-1409.

16.

Park

Kim

, et al. Initial nutritional status and clinical outcomes in patients with deep neck infection. Clin Experi Otorhinolaryngol. 2018;11:293-300.

17.

Casaer

Van den Berghe

. Nutrition in the acute phase of critical illness. New Engl J Med. 2014;370:2450-2451.

18.

Marioni

Castegnaro

Staffieri

, et al. Deep neck infection in elderly patients. A single institution experience (2000-2004). Aging Clin Experi Res. 2006;18:127-132.

19.

Lin

Gao

Yue

, et al. A novel risk score for the prediction of airway management in patients with deep neck space abscess: A multicenter retrospective cohort study. J Intensive Care. 2021;9(1):41.

20.

Mejzlik

Celakovsky

Kotulek

, et al. Univariate and multivariate models for the prediction of life-threatening complications in 586 cases of deep neck space infections: Retrospective multi-institutional study. J Laryngol Otol. 2017;131:779-784.

21.

Suehara

Rodrigues

AAN

Menezes

, et al. Predictive factors of lethality and complications of deep fascial space infections of the neck. Rev Colegio Brasil Cirurgioes. 2020;47:e20202524.

22.

O’Brien

Snapp

Dugan

Westgate

Gupta

. Risk factors affecting length of stay in patients with deep neck space infection. Laryngoscope. 2020;130:2133-2137.

23.

Wang

Tai

Kuo

Chien

. Predisposing factors of complicated deep neck infections: 12-year experience at a single institution. J Otolaryngol. 2010;39:335-341.

24.

Chen

Young

Tsai

Chien

Kang

Liao

, et al. Factors affecting the necessity of tracheostomy in patients with deep neck infection. Diagnostics. 2021;11(9):1536.

A novel model for predicting mortality in the management of deep neck infections

Abstract

Objectives

Methods

Results

Conclusions

Keywords

Introduction

Patients and methods

Study participants and ethical approval

Statistical analysis

Results

Patient characteristics

Factors associated with mortality

Factors associated with long-term hospitalization

The development of a mortality prediction model

Discussion

Conclusions

Footnotes

Acknowledgments

Declaration of conflicting interests

Funding

Ethical approval

ORCID iD

Abbreviations

References