Severe skin and soft tissue infections (SSTIs) and necrotizing soft tissue infections (NSTIs) represent a significant burden of disease, with over 1 million annual admissions for SSTI and over 20 thousand annual admissions for NSTI in the United States alone. While overall intensive care unit (ICU) admission rates for SSTIs are low, they are still increasing relative to other diseases. Severe SSTI and NSTIs represent a time-sensitive disease process with prompt surgical evaluation and management as a cornerstone of treatment. A previous study suggested that delayed surgical interventions of >12 h among patients with necrotizing fasciitis were associated with increased number of surgical debridement and higher incidence of septic shock and AKI.
Recognition of severe SSTI and NSTI can be challenging, and external physical examination frequently cannot distinguish NSTI. Other testing such as blood markers, composite scores such as laboratory risk indicator for necrotizing fasciitis (LRINEC), and imaging can be helpful but are not perfectly sensitive.
While many patients with severe infection may have overt shock and hypotension at the index presentation, a significant number of patients develop septic shock in a delayed fashion after hospital and ICU admission, which is associated with worse outcomes. Associated organ dysfunction that can occur with such infections is also associated with increased mortality. Therefore, the 2016 Consensus Definitions of Sepsis and Septic Shock highlights further understanding of a dysregulated host response to infection and focuses on organ dysfunction, which is associated with increased mortality generally.
While septic shock is often typified by the need for vasopressors to maintain blood pressure despite adequate fluid administration, the published data showed that elevated lactate >2 mmol/L or elevated sequential organ failure assessment (SOFA) score is associated with significantly increased risk-adjusted mortality. One study of NSTI patients requiring debridement noted an increase in SOFA score at 1 day following admission with gradual decline in SOFA scores afterward; it also reported that elevated admission SOFA score over 3 was associated with higher risk of mortality, length of stay, and acute kidney injury (AKI). Patients with soft tissue infection and delayed manifestations of organ dysfunction may be considered to have an occult form of septic shock (occult septic shock [OSS]), with a different phenotype as those with persistent hypotension at presentation, which in this study we termed early septic shock (ESS).
Early interventions for septic shock are associated with improved outcomes. Therefore, early recognition and aggressive treatment of patients who are at risk for OSS is paramount for improving outcomes among patients with SSTI, while they are waiting for definitive surgical intervention, if indicated. Therefore, we sought to investigate factors associated with shock and organ dysfunction in a diverse cohort of patients with SSTI and NSTI.
Study design and setting
This was a retrospective, chart review study. The setting was a single academic, urban hospital, with a specialized Soft Tissue Surgery Service which serves as a regional referral consultant for patients with SSTI. On arrival at our medical center, the patient is evaluated by the Soft Tissue Surgery Team, which is composed of an attending physician, an advanced practice provider, and surgical residents. During patient's ICU stay, the Soft Tissue Surgery Team coordinates with the ICU team to jointly manage the patient.
The study period was between January 1, 2014, and December 31, 2016. Eligible patients were adults with SSTI who were admitted directly from an emergency department (ED) to the Soft Tissue Surgery Team at our academic center. We excluded patients who were admitted from other hospitals' inpatient units since our interests focused on early clinical factors and patient management in the ED setting. We also excluded patients who did not have complete ED records when they were transferred to our medical center. The study was approved by our institutional review board.
Data for a patient's ED clinical course were extracted from referring ED records. Patients' clinical courses and outcomes at our academic center were extracted from our electronic medical records.
Data extracted from referring ED records included date, time, and vital signs at an ED triage. For a patient's care intensity and disease severity, we extracted the emergency severity index, components of the LRINEC score for NSTI, and components of the SOFA score. Additional serum laboratory values during the ED stay such as serum lactate and white blood cell count were collected. We also recorded time from ED arrival to ED intervention.
ED clinicians' interventions were recorded. Interventions that were ordered but not administered to patients were not recorded. We cross-checked these interventions from both referring ED records and accepting nursing reports at our academic center's accepting units.
From our academic medical center's electronic records, we extracted initial vital signs, initial laboratory values, time from arrival to the interventions, surgical interventions, and estimated blood loss (EBL) of the index surgical operation. Besides SOFA score and the LRINEC score, the EBL served as a surrogate marker for patients' severity index. We assigned a value of 0 EBL for patients not undergoing surgical intervention. We also collected blood culture results, tissue culture results, and final diagnoses from our institution's electronic medical records.
Data were extracted by investigators who were individually trained by the principal investigator (PI). Investigators were supervised until the accuracy of their data collection reached 90% in agreement with data extracted by the PI or another senior investigator. Up to 20% of data were subsequently checked by another investigator to maintain at least 90% interrater agreement. To further reduce bias, investigators who were not blinded to the study hypothesis collected data in separate sections. For example, investigators extracting laboratory data did not have access to data regarding treatment nor outcomes. Differences were adjudicated by discussion and agreement between investigators and the PI during our meetings occurring every other month. Data were entered into a standardized database (Microsoft Access, Microsoft Corp., Redmond, Washington, USA).
Occult septic shock and early septic shock
Our primary outcome was the prevalence of OSS, which was operationally defined as the initiation of vasopressors within the first 24 h of arrival at our quaternary care center. ESS was defined as the initiation of vasopressors when patients first presented to the ED.
Acute kidney injury
Our secondary outcome of AKI was defined in accordance with the Kidney Disease Improving Global Outcome (KDIGO) scale [Appendix 1]. The stages of the KDIGO scale indicate AKI from 0 to 3, with 3 being the most severe stage of kidney injury. Patients with history of end-stage renal disease were included in our analysis; however, they were not considered to have AKI.
First, we categorized our data into groups: all patients, patients with NSS, patients with ESS, and patients with OSS. We presented our data with mean ± standard deviation (SD) or median with interquartile range (IQR) and number with percentage (N, %) as appropriate. For comparison of continuous data, we used the Student's t-test or Mann–Whitney U-test as indicated. Categorical data were compared via Pearson's Chi-square or Fisher's exact tests as appropriate.
We used logistic regressions to measure the association between patients' demographic and clinical data with the outcomes of interest. All variables were identified a priori. We measured the goodness-of-fit of our multivariable logistic regression with the Hosmer-Lemeshow test, for which a P > 0.05 demonstrates a model with good fit. We also evaluated the collinearity of our independent variables by their variance inflation factor (VIF). Any factor with VIF >5.0 would be considered having collinearity with other factors and would be removed from our multivariable logistic regression. The discriminatory capability of each multivariable logistic regression model was evaluated further by using the area under the receiving operating characteristic (AUROC) curve. An AUROC curve of 0.6–0.7 is considered to have good discriminatory capability, while an AUROC curve of 0.8–0.9 is considered to have very good discriminatory between dichotomous outcomes (OSS vs. not, AKI vs. not). All variables with two-tailed P < 0.05 were considered statistically significant.
All data analysis was performed with Minitab version 19 (www.minitab.com; State College, Pennsylvania, USA).
There were 902 patients who were admitted to the Soft Tissue Surgery Service at our medical center during the study period. After excluding 633 patients who were transferred from other hospitals' inpatient units, we identified and included 269 patients who were admitted directly to our medical center from other hospitals' EDs. Among 269 patients, 218 (81%) had NSS, 16 (6%) patients had ESS, while 35 (13%) patients had OSS [Table 1].
Patients who had OSS had higher prevalence necrotizing fasciitis (30/35 patients, 83%), compared with patients with NSS (92/218, 42%; P < 0.001); however, there was no significant difference between the groups with ESS and OSS. Patients who had OSS (5 [3–11]) had significantly higher median (IQR) SOFA score, in comparison of those with NSS (2 [1–3], P < 0.001). Similarly, patients with OSS (7 [4–9]) also had significantly higher median (IQR) LRINEC score in comparison of those with NSS (5 [3–8], P = 0.017). In contrast, both SOFA and LRINEC scores were not statistically significant between patients with ESS and OSS.
Occult septic shock versus early septic shock
During the ED course, patients with OSS had lower mean (±SD) of shock indexes (OSS 0.90 ± 0.22; ESS 1.21 ± 0.35; P = 0.004) and were given lower volumes of intravenous fluids (IVFs) as milliliter per kilogram of body weight (OSS 23 ± 17; ESS 39 ± 16; P = 0.003) compared to those with ESS [Table 2].
Within the first 24 h of arrival at our quaternary medical center, patients with OSS had higher median [IQR] number of debridement (OSS 3 [1–5]; ESS 1 [1–2.75]; P = 0.015) and an increase in shock index (OSS 0.04 ± 0.29; ESS −0.32 ± 0.25; P < 0.001) [Table 2].
Occult septic shock versus no septic shock
During the ED course, patients with OSS had higher serum lactate values (OSS 3.5 ± 2.6; NSS 2.1 ± 1.6; P = 0.004). However, there was no significant difference in administration of IVF between groups (OSS 3.5 ± 2 mmol/L; NSS 2.1 ± 1 mmol/L; P < 0.01). Within 24 h of arrival at our quaternary medical center, patients with OSS continued to show higher markers of shock such as shock index and serum lactate compared to patients with NSS [Table 2].
Primary outcome: Occult septic shock
As reported above, the prevalence of OSS within our patient population who were admitted from the ED was 13% (35/269 patients). An increase of one point on the SOFA score increased the likelihood of developing OSS by 41% (odds ratio [OR] 1.41, 95% confidence interval [CI] 1.23–1.62, P < 0.001). Increase of shock index differences (OR 21.27, 95% CI 3.35–50+, P = 0.001) was also associated with higher likelihood of developing OSS during hospital stay. Diagnosis of necrotizing fasciitis (OR 12.69, 95% CI 2.10–50+, P = 0.006) or abscess (OR 11.94, 95% CI 1.42–50+, P = 0.023) was associated with higher likelihood of developing OSS during hospital stay. All factors showed no collinearity as their variance inflation factor was <5. The logistic regression model showed good fit as the P value for the Hosmer-Lemeshow test was 0.51. Furthermore, the model demonstrated very good discriminatory capability, with its AUROC value of 0.86 [Table 3].
Secondary outcome: Acute kidney injury
As reported above, 18% of the patients admitted from the ED had AKI. Patients with NSS had an 84% lower risk of developing AKI (OR 0.16, 95% CI 0.08–0.33, P < 0.001). This model showed good fit of data, with the Hosmer-Lemeshow test's P = 0.938. The discriminatory capability was good, with the model's AUROC value of 0.67 [Table 3].
Figure 1 illustrates the statistically significant association (coefficient 0.2, P < 0.001) between SOFA score and patients' probability of developing OSS, using probit logit regression. A SOFA score of 5 denotes an 18% probability of developing OSS. When a patient has an SOFA score of 10, the patient will have a 50% probability of developing OSS.
Figure 2 illustrates the statistically significant association (coefficient −1.05, P < 0.001) between the presence of septic shock and patients' probability of developing AKI. When a patient has NSS, the patient would have an 11.9% probability of developing AKI. However, the presence of septic shock is associated with 45% chance of developing AKI.
This study illustrates that up to 13% of the patients with SSTI developed septic shock and will eventually require vasopressors within 24 h after leaving the ED. These patients with OSS, despite being normotensive, have clinical indicators consistent with shock and organ dysfunction similar to those who required vasopressors while in the ED; however, they were treated similar to those who had no septic shock. We also defined a few clinical factors among patients with SSTIs that were associated with higher risk of developing OSS, which is the need for vasopressors within 24 h of arrival at our quaternary medical center.
Our study also identified a few early clinical factors between the patients who had OSS versus NSS. Patients with OSS, despite having relatively normal blood pressure, presented with elevated serum lactate levels, high SOFA score, and relatively high shock index, which were similar to those who had ESS and received norepinephrine in the ED. However, the shock index for those who had OSS still increased during the ED stay, while the shock index from those who had NSS or ESS decreased. Although further studies are needed to confirm our observations, our findings illustrated the importance of early recognition and treatment of sepsis among patients with SSTI. Both the ESS and OSS groups had the same mortality rate and the same rate of AKI, while the ESS group required less debridement and shorter hospital length of stay. Our observation is similar to a previous studies comparing patients with ESS and late septic shock. Roman-Marchant et al. reported that patients who had ESS showed early organ dysfunction, yet they were associated with slightly better mortality and shorter hospital stay, when compared to patients who developed late septic shock. ESS patients in this study received more crystalloid fluids and earlier vasopressors and thus may have had earlier correction of hypoperfusion and inflammatory burden when compared with OSS patients. Patients with occult presentations of shock, where recurrent hypotension may not prompt more aggressive care, would have more prolonged time of hypoperfusion or organ dysfunction before recognition and treatment. A previous study showed that a 5-min period of hypotension was associated with 18% likelihood of developing AKI or 30% higher likelihood of myocardial infarction.
Although more studies are needed to support our observations, our findings provide further evidence to suggest early interventions for patients who were at risk of developing OSS, even when their blood pressures were still normal. In our study population, patients with OSS presented to EDs with similar clinical manifestations and serum markers as patients with ESS, but they received much less fluid resuscitation, and they received vasopressors at a much later time. Therefore, for patients with elevated serum lactate and SOFA score, clinicians should have low threshold to initiate vasopressor. There has been evidence that early vasopressor administration improves mortality in patients with septic shock. In addition, initiating vasopressor at low dose via peripheral venous catheters has been shown to be safe. Hence, clinicians, especially busy emergency clinicians, do not have to consider central venous access immediately.
Critically ill patients who developed AKI would be associated with worse outcome. With a small increase in serum creatinine according to the KDIGO criteria of Stage 1 AKI, patients will develop 31% higher likelihood of mortality at 28 days (hazard ratio 1.31, 95% CI 1.0–1.7, P = 0.042). Our study suggested that patients who had soft tissue infection and did not develop septic shock were not associated with high probability of AKI. Once a patient developed septic shock, the patient had 45% probability of developing AKI. Therefore, while waiting for definitive surgical treatment for soft tissue infection, clinicians should consider organ dysfunction as a marker or patients needing further resuscitation and at significant probability of needing vasopressors.
Our study has several limitations. We performed a retrospective, chart review to assess patients' clinical care; therefore, possible charting error or incomplete charts limit our data validity. As such, we could not ascertain the rationale why patients were placed on vasopressors or not. In addition, we used serum creatinine at ED presentation as the baseline for our criteria of AKI; therefore, we might have underestimated the rate of AKI as patients with severe SSTI might have presented to the ED with already elevated creatinine levels. Furthermore, since most of our patients were transferred from different locations, we did not have access to their records after they were discharged from our hospital, so we could not ascertain whether their AKI resolved. There were a small number of patients with either ESS or OSS, which may cause a wide 95% CI for certain independent variables in our multivariable logistic regressions. Furthermore, we did not use mortality as an outcome because the mortality rate of our patient population was low, which was likely due to the conditions of our patients' disease states. Once they received surgical debridement to remove their sources of infection, their septic condition was likely reversed and improved.
There were approximately 13% of patients with SSTI that developed OSS. Patients with SSTI and OSS with early evidence of organ dysfunction and need for vasopressors within 24 h had similar serum lactate and SOFA scores as those with ESS, but they required more debridement and had longer hospital stays. Higher SOFA score at presentation was associated with higher probability of developing OSS, while having NSS was associated with lower likelihood of developing AKI. Further studies are necessary to confirm our observations.
Research quality and ethics statement
This study was approved by the Institutional Review Board and was determined not to require Ethics review. The authors followed applicable EQUATOR Network (https://www.equator-network.org/) guidelines during the conduct of this research project.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
We thank Ms. Jennifer Tchai for her contributions.
Appendix 1: Kidney Disease Improving Global Outcome Scale
Stage 1 is defined as an increase in serum creatinine of 0.3 mg/dL from the emergency department (ED) admission level. Stage 2 is defined as a 2–2.9-fold increase, and Stage 3 is defined as a 3-fold increase in serum creatinine from the ED admission level.
1. Tran Z, Cho NY, Verma A, Sanaiha Y, Williamson C, Hadaya J, et al Clinical and financial outcomes of necrotizing soft-tissue infections in safety-net hospitals J Surg Res. 2021;267:124–31
2. Edelsberg J, Taneja C, Zervos M, Haque N, Moore C, Reyes K, et al Trends in US hospital admissions for skin and soft tissue infections Emerg Infect Dis. 2009;15:1516–8
3. Psoinos CM, Flahive JM, Shaw JJ, Li Y, Ng SC, Tseng JF, et al Contemporary trends in necrotizing soft-tissue infections in the United States Surgery. 2013;153:819–27
4. Sartelli M, Guirao X, Hardcastle TC, Kluger Y, Boermeester MA, Raşa K, et al 2018 WSES/SIS-E consensus conference: Recommendations for the management of skin and soft-tissue infections World J Emerg Surg. 2018;13:58
5. Gelbard RB, Ferrada P, Yeh DD, Williams BH, Loor M, Yon J, et al Optimal timing of initial debridement for necrotizing soft tissue infection: A Practice Management Guideline from the Eastern Association for the Surgery of Trauma J Trauma Acute Care Surg. 2018;85:208–14
6. Kobayashi L, Konstantinidis A, Shackelford S, Chan LS, Talving P, Inaba K, et al Necrotizing soft tissue infections: delayed surgical treatment is associated with increased number of surgical debridements and morbidity Trauma Inj Infect Crit Care. 2011;71:1400–5
7. Fernando SM, Tran A, Cheng W, Rochwerg B, Kyeremanteng K, Seely AJ, et al Necrotizing soft tissue infection: Diagnostic accuracy of physical examination, imaging, and LRINEC score: A systematic review and meta-analysis Ann Surg. 2019;269:58–65
8. April MD, Long B. What is the accuracy of physical examination, imaging, and the LRINEC score for the diagnosis of necrotizing soft tissue infection? Ann Emerg Med. 2019;73:22–4
9. Huang CT, Tsai YJ, Tsai PR, Yu CJ, Ko WJ. Severe sepsis and septic shock: Timing of septic shock onset matters Shock. 2016;45:518–24
10. Bekker MA, Rai S, Arbous MS, Georgousopoulou EN, Pilcher DV, van Haren FM. Annual prevalence, characteristics, and outcomes of intensive care patients with skin or soft tissue infections in Australia and New Zealand: A retrospective cohort study between 2006-2017 Aust Crit Care. 2021;34:403–10
11. Bulger EM, May A, Bernard A, Cohn S, Evans DC, Henry S, et al Impact and progression of organ dysfunction in patients with necrotizing soft tissue infections: A multicenter study Surg Infect (Larchmt). 2015;16:694–701
12. Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, Bauer M, et al The third international consensus definitions for sepsis and septic shock (Sepsis-3) JAMA. 2016;315:801–10
13. Li Y, Li H, Zhang D. Timing of norepinephrine initiation in patients with septic shock: A systematic review and meta-analysis Crit Care. 2020;24:488
14. Roman-Marchant O, Orellana-Jimenez CEA, De Backer D, Melot C, Vincent JL. Septic shock of early or late onset: Does it matter? Chest. 2004;126:173–8
15. Walsh M, Devereaux PJ, Garg AX, Kurz A, Turan A, Rodseth RN, et al Relationship between intraoperative mean arterial pressure and clinical outcomes after noncardiac surgery Anesthesiology. 2013;119:507–15
16. Tran QK, Mester G, Bzhilyanskaya V, Afridi LZ, Andhavarapu S, Alam Z, et al Complication of vasopressor infusion through peripheral venous catheter: A systematic review and meta-analysis Am J Emerg Med. 2020;38:2434–43
17. Linder A, Fjell C, Levin A, Walley KR, Russell JA, Boyd JH. Small acute increases in serum creatinine are associated with decreased long-term survival in the critically ill Am J Respir Crit Care Med. 2014;189:1075–81