Abstract
Background:
Necrotizing soft tissue infections (NSTIs) are rapidly progressing, life-threatening diseases associated with substantial morbidity and mortality, especially in patients 65 years or older. We aimed to evaluate clinical factors associated with mortality and discharge disposition after NSTIs in elderly patients.
Patients and Methods:
Retrospective data were obtained from the 2007–2017 American College of Surgeons–National Surgical Quality (ACS-NSQIP) database. Patients aged 65 years or older with a post-operative diagnosis of an NSTI (defined as gas gangrene, necrotizing fasciitis, or Fournier gangrene) were included. Univariable and multivariable analyses were performed to identify independent clinical and demographic factors associated with mortality and with discharge disposition.
Results:
A total of 1,460 patients were included. Median age was 71 years, 43% were females. Overall, 30-day mortality was 18.5% and 30-day morbidity was 63.6%. The most important predictors of mortality included pre-operative septic shock (odds ratio [OR], 6.36; 95% confidence interval [CI], 3.61–11.18), pre-operative dialysis dependence (OR, 2.99; 95% CI, 1.77–5.05), coagulopathy (international normalized ratio [INR], >1.5, OR, 2.25; 95% CI, 1.51–3.37), hepatobiliary disease (bilirubin >1.0 mg/dL; OR, 2.05; 95% CI, 1.38–3.04) and aged 80 years or older (OR, 3.36; 95% CI, 2.08–5.44). Patients without any of these risk factors had a mortality of 7.3%. Predictors of discharge to inpatient rehabilitation or skilled care included age 80 years or older (OR, 2.49; 95% CI, 1.44–4.30), American Society of Anesthesiologists (ASA) ≥3 (OR, 2.05; 95% CI, 1.03–4.05)] and amputation as opposed to debridement (OR, 2.53; 95% CI,1.48-4.32).
Conclusions:
We identified several pre-operative clinical factors that were associated with increased post-operative mortality and discharge to post-acute care. The next steps should focus on determining if optimization of modifiable predictors would improve mortality.
Necrotizing soft tissue infections (NSTIs) are rapidly progressing, life-threatening diseases characterized by inflammation and necrosis of subcutaneous fat, fascia, or muscle [1]. Despite the increased attention to early diagnosis of NSTIs and the improvements in their surgical management, they continue to be associated with substantial morbidity and mortality, and reduced quality of life [2,3]. Large sample studies estimate the mortality after NSTIs to be between 10% and 12% in the general population [4–6], and it is speculated to be even higher in elderly patients because of their reduced physiologic reserve and frailty [7].
The elderly (aged 65 and above) comprise one of the fastest growing segments of the population in the Unites States [8]. This population increased by more than one-third over the last decade, and by 2030, one in every five American is expected to be above the age of 65 years [8,9]. Age-related structural and functional changes to the skin and soft tissues put elderly patients at increased risk of infections [10]. In particular, one in every five patients diagnosed with NSTIs is estimated to be 65 years or older [4]. Other risk factors that are less specific to the elderly include penetrating skin trauma, cirrhosis, diabetes mellitus, and neutropenia [11]. However, signs and symptoms of skin and soft tissue infections tend to be subtle and atypical in older adults and often result in greater delays in diagnosis [10,12]. Given the rapidly expanding nature of the disease, failure to identify it and manage it early can lead to worse outcomes [13,14]. Additionally, elderly patients tend to have several comorbidities at baseline, and constitute a significant portion of hospital admissions, often requiring continued care even after discharge [15,16]. Although risk factors for NSTIs and outcomes after delayed diagnosis have been well studied in older adults, little is known about the impact of pre-operative factors on mortality and discharge disposition after NSTIs in this population.
Given the expanding geriatric population in the United States and the limited data on mortality from NSTIs in this specific age group, we decided to use the American College of Surgeons–National Surgical Quality (ACS-NSQIP) database to identify demographic and clinical predictors of mortality and discharge disposition in elderly patients with NSTIs. We hypothesized that in elderly patients with NSTIs, pre-admission comorbidities are the major determinant of 30-day mortality and discharge disposition.
Patients and Methods
Patient population
All patients 65 years and older, with a post-operative diagnosis of NSTI in the ACS-NSQIP from 2007 to 2017 were included. Patients who died in the emergency department were excluded from the analysis.
Defining necrotizing soft-tissue infections
Given that International Classification of Diseases, 10th Revision (ICD-10) coding was implemented in October 2015, cohort selection was based on both ICD-9 and ICD-10 codes for NSTIs. Specifically, an NSTI was defined as gas gangrene, necrotizing fasciitis, or Fournier gangrene (ICD-9-CM codes 040.0, 728.86, and 608.83 and ICD-10-CM codes A480, M726, and N493, respectively).
Defining surgical procedures
Primary surgery was based on the Current Procedural Terminology (CPT) code of the Principal Operative Procedure, defined by ACS-NSQIP as “the most complex of all the procedures performed by the primary operating team during the trip to the operating room” [17]. Additional procedures at time of primary surgery were derived from the CPT code of the Other Procedure variable, defined as “an additional surgical procedure performed by the same surgical team, under the same anesthetic which has a CPT code different from that of the Principal Operative Procedure” [17], and from the CPT code of the Concurrent Procedure variable, which refers to “an additional operative procedure performed by a different surgical team (i.e., a different specialty/service) under the same anesthetic that has a CPT code different from that of the Principal Operative Procedure” [17].
Data analysis
Descriptive statistics were used to report demographics, pre-operative comorbidities, laboratory values, operative procedure, post-operative outcomes, and discharge disposition of the study cohort. Univariable analyses were used to compare survivors and non-survivors. Clinically relevant variables were retained for the multivariable analysis. Multivariable logistic regression models were constructed to control for confounders and identify predictors of 30-day mortality. The five predictors with the highest odds ratio for mortality were selected out of those that achieved statistical significance. Mortality rates were calculated in the presence of one, two, or three predictors to identify combinations of risk factors that put the patient at highest risk of death. Admission from long-term care facility was excluded from consideration because it does not represent a pre-operative disease or comorbidity and was addressed separately. Post-acute care was defined as inpatient rehabilitation or skilled care facility. Univariable and multivariable analyses were used to identify predictors of discharge to post-acute care. Rates of post-acute care utilization were calculated for individual predictors as previously described for mortality.
STATA, version 15.0 (StataCorp, College Station, TX) was used for statistical analyses in this study. Categorical variables were presented as absolute values and percentages. Continuous variables were presented as either mean and standard deviation (SD) for normal distribution or median and interquartile range (IQR) for non-normal distribution. For univariable analyses, χ2 test or Fisher's exact test were performed for categorical variables and either Wilcoxon rank-sum tests or t-tests were performed for continuous variables, based on the distribution. A p value <0.05 was considered statistically significant for all tests. Missing data was addressed using multiple imputation with chained equations (MICE), creating 10 iterations for variables with missing data points.
Ethical oversight
This study was reviewed and approved by the Partners Human Research Committee.
Results
Study population
From a total of 6,485,915 patients, 1,460 patients were included. Table 1 summarizes the demographics, clinical characteristics, and post-operative outcomes of subjects. Median age was 71 years, 43% were females, 76.1% were white, and 52.4% had a body mass index (BMI) of 30 or more. The three most common comorbidities were hypertension (73.4%), presence of an open wound (62.4%), and diabetes mellitus on medications (53.2%). Hypoalbuminemia (albumin <3 g/dL), was the most common laboratory derangement, found in 74.4%. The mean duration from hospital admission to operation (standard deviation [SD]) was 2.9 (SD, 6.8) days. Overall, 30-day mortality was 18.5% and 30-day morbidity was 63.6%. Fifty-seven percent of patients were discharged to post-acute care, whereas 24.3% went home. The two most common post-operative complications were failure to wean off ventilator for more than 48 hours (26.7%) and post-operative septic shock (23.6%).
Comparison of Characteristics of Survivors versus Non-Survivors
IQR = interquartile range; ASA = American Society of Anesthesiologists; BMI = body mass index; COPD = chronic obstructive pulmonary disease; RBC = red blood cell; BUN = blood urea nitrogen; INR = international normalized ratio; SGOT = serum glutamic oxaloacetic transaminase; WBC = white blood cell; CVA = cerebrovascular accident; CPR = cardiopulmonary resuscitation; MI = myocardial infarction.
Of the patients in the “Other” category, the following had “Other procedure” or “Concurrent procedure”: 72 had a debridement, three had an amputation, three had tissue/organ excision, 10 had abscess drainage or fasciotomy, 53 had miscellaneous diverse procedures and 55 had no recorded “Other procedure” or “Concurrent procedure”. These numbers add up to 196 rather than 193 because three patients had more than one “Other procedure” or “Concurrent procedure”.
When it comes to the admission source, 64.2% of patients were admitted from home and 7.9% from a chronic care facility. Compared with patients admitted from home, patients who were living in a long-term care facility were more likely to have three or more comorbidities (74% vs. 59%; p = 0.002), have an ASA of four or more (66% vs. 51%; p = 0.002), had a higher 30-day mortality (28% vs. 17%; p = 0.006), but were less likely to be in septic shock pre-operatively (20% vs. 26%; p = 0.04).
Predictors of 30-day mortality
From the multivariable analysis (Table 2), the following variables were identified as predictors of mortality: pre-operative sepsis, pre-operative septic shock, pre-operative dialysis dependence, history of severe chronic obstructive pulmonary disease (COPD), disseminated cancer, hypoalbuminemia, hyperbilirubinemia, and international normalized ratio (INR) >1.5. Patients admitted from a long-term care facility had more than twice higher odds of 30-day mortality compared with those admitted directly from home. Being overweight was the only protective factor against mortality.
Multivariable Analysis: Mortality
COPD = chronic obstructive pulmonary disease; INR = international normalized ratio; BMI = body mass index.
Table 3 summarizes the selected predictors with the respective 30-day mortality. As expected, the presence of more predictors was associated with higher mortality. In dialysis-dependent patients, mortality was 39.5% (51/129), increasing to 61% (36/59) in those who were dialysis-dependent and were in septic shock pre-operatively, and increasing further still to 80% (4/5) in those 80 years and older with dialysis dependence and pre-operative septic shock. In patients younger than 80 years old who are not in pre-operative septic shock, not on dialysis, and have a normal bilirubin, and INR, mortality decreases to 7.3% (23/315).
Selected Predictors of Mortality
INR = international normalized ratio.
Predictors of discharge disposition
From the multivariate analysis (Table 4), the following variables were identified as predictors of discharge to post-acute care: Age more than 80 years, ASA of 3 or more, transfer from acute care hospital, pre-operative septic shock, hypoalbuminemia, and amputation surgery. Table 5 summarizes the selected predictors with the respective rates of discharge to post-acute care. As expected, the presence of more predictors was associated with higher rates of post-acute care utilization. Among subjects who received an amputation, 82.6% (123/149) were discharged to post-acute care, increasing to 94.4% (17/18) in patients who were transferred from an acute care hospital and received an amputation.
Multivariable Analysis: Discharge Disposition
ASA = American Society of Anesthesiologists.
Selected Predictors of Discharge to Post-Acute Care
PAC = post-acute care; ASA = American Society of Anesthesiologists.
Discussion
Pre-operative risk factors and mortality
Using the well-validated ACS-NSQIP database, we present an analysis of the pre-operative factors associated with 30-day mortality in geriatric patients diagnosed with NSTIs. Early identification of patients at high risk of mortality can help provide targeted treatment to reduce the burden of comorbidities, while also allowing an informed discussion between the surgeon and the patient and their family. One strength of our study is that we identified mortality in the presence of a single or a combination of clinical or demographic characteristics in patients older than 65 years presenting with NSTIs.
Our results show that 30-day mortality among patients 65 years and older admitted with a diagnosis of NSTI is as high as 18.5%. This is higher than the reported 10%–12% mortality in the general population [4–6]. Elderly people tend to have more comorbidities and develop more severe complications because of underlying frailty compared with their younger counterparts [18–21]. Notably in our analysis, patients 80 years or older were three times more likely to die of an NSTI relative to patients between 65 and 69 years. This is comparable to what Hajibandeh et al. [22] demonstrated in a meta-analysis looking at mortality in octogenarians after emergency general surgery [22].
Among pre-hospital characteristics, predictors of mortality included pre-operative dialysis dependence, hyperbilirubinemia (bilirubin >1.0 mg/dL) and coagulopathy (INR >1.5). McCarty et al. [23] , similarly found that pre-existing liver and kidney diseases were associated with increased odds of mortality in geriatric patients diagnosed with NSTIs. One possible explanation is that failure of these organ systems puts patients at greater risk of infections and can precipitate failure of other organs as well.
Hypoalbuminemia was also a predictor of mortality present in a significant portion of our population with almost three in every four patients having a low albumin pre-operatively. However, it is difficult to establish a link between albumin levels and malnutrition in patients with an active infection because albumin is a negative acute phase reactant [24]. Despite the ongoing infection, malnutrition is likely to play a role in elderly patients but establishing a direct link is not possible with the available data. Regardless of nutritional status, low albumin levels in hospitalized elderly patients have been reported to be associated with higher mortality [25–27].
Although pre-operative factors such as patient age and hyperbilirubinemia cannot be modified, early operative management to prevent patients from going into pre-operative septic shock and early initiation of dialysis may play a role in lowering the risk of mortality and avoiding multi-organ failure.
Patient selection and survival rate
Mortality is low in a considerable number of geriatric patients in the absence of the five most lethal pre-operative risk factors that we identified. In fact, patients between 65 and 79 years, without a history of kidney disease or coagulopathy, who are not in pre-operative septic shock, and have a normal bilirubin are at substantial chance of survival after a diagnosis of NSTI. Their 30-day survival was almost 93%. With a high overall mortality in the geriatric population, clinicians might be inclined to expect unfavorable outcomes after a diagnosis of NSTIs. However, we demonstrate that even among geriatric patients, outcomes can vary significantly, and knowledge of pre-operative risk factors is key to accurate prognostication. This speaks to the importance of individualized decision-making and pre-operative risk counseling in elderly patients diagnosed with NSTIs.
Admission source and mortality
Our data show that older patients admitted from a long-term care facility have more than twice the odds of dying of an NSTI than their counterparts who are admitted from home. They also have more comorbidities and a higher ASA at baseline but have lower rates of pre-operative septic shock, which can be used as an objective measure of infection severity. In other words, the higher mortality in this patient population can be traced back to their worse health at baseline. Previous studies show that elderly patients living in chronic care facilities often have significant frailty, multiple comorbidities, and reduced reserve capacity [28]. They are also more prone to be re-admitted to a hospital and have a more complicated disease course [28]. Special attention should be given to those patients when diagnosed with an NSTI as they tend to have worse outcomes.
Pre-operative factors and discharge to post-acute care
Although most patients were admitted from home, only 24.3% were discharged back home. As age increased, notably with patients older than 80 years, the odds of being discharged to a facility increased, with almost 80% of those patients being discharged to post-acute care. This is in line with previous reported data after emergency general surgeries that suggests that as age increased, the likelihood of comorbidities increased, requiring additional care after discharge [29]. Additionally, receiving an amputation resulted in higher odds of post-acute care utilization compared with getting a debridement. The latter would typically involve thorough removal of all devitalized tissue, resulting in large defect that require substantial wound care. However, limb amputations can have a more substantial impact on quality of life and patients often require rehabilitation in order to preserve a certain functional level.
Hypoalbuminemia is another predictor of post-acute care utilization. Albumin is well known for its importance in wound healing and recovery after surgery. Despite the absence of data on the effect of hypoalbuminemia on recovery after surgery for NSTIs, several studies have correlated low pre-operative albumin levels with worse functional outcomes and slower recovery after surgery for hip fracture [30,31]. As a result, those patients require continued care beyond the initial hospital admission justifying their discharge to skilled care facilities or inpatient rehabilitation. Studying discharge to post-acute care is of particular importance because it can serve as an indicator of a patient's functional outcome at the time of discharge [32].
Limitations
Our study has some limitations. First, we are limited to the variables that are reported in the ACS-NSQIP database and as such the risk of residual confounders that were not accounted for cannot be excluded. One such variable is insurance status, which could play a role in influencing discharge disposition. Second, it would have been interesting to report on the bacteriology responsible for these infections and describe evolving trends, but this data was not available in the ACS-NSQIP database. In addition, we could not provide information on the site of infection because these data are not available in the dataset. Third, because of the absence of variables such as C-reactive protein, hemoglobin, and glucose we could not calculate the Laboratory Risk Indicator for Necrotizing Fasciitis (LIRNEC) score and study its association with post-operative outcomes. Although the accuracy of LIRNEC in identifying NSTIs is debatable, it can be used in the appropriate clinical settings to aid in diagnosis.
Conclusions
In elderly patients diagnosed with NSTIs, the major predictors of mortality are pre-operative septic shock, pre-operative dialysis dependence, elevated bilirubin and INR, and age 80 years or older. In the absence of those predictors of mortality, survival is high, thus patient selection based on pre-operative factors is key during pre-surgical counseling. Residents of long-term care facilities require increased attention as their baseline comorbidities put them at a higher risk of mortality after NSTIs. The next steps should focus on determining if optimization of modifiable predictors would improve mortality.
Footnotes
Authors' Contributions
All the co-authors were integrally involved in the writing of this manuscript via study conception/design and/or data acquisition and analysis/interpretation. Furthermore, all authors made significant contributions to the drafting or critical revisions of the manuscript, and all authors gave final approval prior to submission for publication and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Funding Information
No grants or other forms of funding or assistance were utilized in the creation of this manuscript.
Author Disclosure Statement
We certify that all listed co-authors have no commercial associations that might create a conflict of interest in connection with submitted manuscripts.