Utility of qSOFA score in identifying patients at risk for poor outcome in Staphylococcus aureus bacteremia

Sepsis is a significant cause of critical illness worldwide with an increasing incidence afflicting over 500,000 cases/year [1] and mortality rates up to 80% [2, 3]. Improved understanding of the pathobiology of sepsis has highlighted the heterogeneity in the host immune response in sepsis, prompting a need to redefine and update the clinical criteria to characterize the syndrome [4–7].

Sepsis is currently defined as a life-threatening acute organ dysfunction secondary to a dysregulated host response to infection (Sepsis-3) [6]. Previously, sepsis was defined as meeting ≥2 systemic inflammatory response syndrome (SIRS) criteria plus suspected infection. However, the SIRS definition was criticized for its lack of sensitivity; a recent study showed that 1 in 8 patients admitted to the intensive care unit (ICU) with severe sepsis did not meet the requisite minimum of 2 SIRS criteria to fulfill the sepsis definition [8]. Thus, to improve the identification of patients at risk for clinical deterioration from infection, Sepsis-3 recommends a quick scoring system, quick sequential organ failure assessment score (qSOFA), comprised of 3 elements assessed at the bedside (altered mental status, respiratory rate, and systolic blood pressure) and without the need for laboratory tests. Patients who show evidence of 2 out of 3 elements are considered positive for risk of clinical deterioration and therefore should prompt clinicians for investigation of organ dysfunction, initiation or escalation of therapy, and consideration of transfer of care to the ICU.

Staphylococcus aureus is the most common bacterial cause of sepsis in ICU patients [9]. S. aureus bacteremia (SAB) affects an estimated 50/100,000 population, with an overall mortality rate of 19–57% in adults [10]. Despite receipt of the standard treatment, persistent bacteremia occurs in one-third of patients beyond 7 days [11, 12], which can lead to metastatic complications, relapse, prolonged hospitalization, and increased mortality [11, 13–16].

Since Sepsis-3, the prognostic utility of qSOFA for in-hospital mortality has been prospectively validated in patients presenting to the emergency department with suspected infection [17], as well as retrospectively evaluated [18–26] but only a small number of evaluated patients had documented bacteremia. Thus, our study aims to evaluate non-ICU patients presenting with sepsis who later were confirmed to have S. aureus bacteremia for 1) differentiating clinical characteristics based on qSOFA score and 2) comparison of the predictive performance of qSOFA to SIRS criteria and Pitt bacteremia score (PBS) as a prognostic tool to identify those at high risk for poor outcome.

This was a 4-year retrospective cohort study conducted at three university-affiliated medical centers (600-bed community teaching, 600-bed county teaching, and 400-bed academic) in Los Angeles County, California. The study protocol was approved by the institutional review boards at each study site (University of Southern California and Quorum Review) and as the study was retrospective, informed consent was waived. Microbiology reports were screened for all patients with at least one positive blood culture for S. aureus between 2012 and 2016. Inclusion criteria were: 1) age ≥ 18 years, 2) receipt of ≥48 h of effective anti-staphylococcal therapy, 3) initiation of effective anti-staphylococcal therapy within 48 h from the time the first positive blood culture was drawn, 4) monomicrobial growth in the blood culture, and 5) not admitted to the ICU prior to or at onset of bacteremia. Patients with receipt of ≤48 h of effective therapy were excluded as assessment of therapy on mortality outcome was unlikely to be contributory to the antibiotic therapy given.

Medical charts were reviewed for relevant demographics including comorbidities, laboratory, radiographic data, surgical and antimicrobial management, and clinical data. Extracted data were entered into a secure database, the Research Electronic Data Capture software hosted by University of Southern California.

Sepsis-3 introduces qSOFA as a simple bedside tool for screening patients with suspicion of infection who are at increased risk for clinical deterioration [6]. Our study aimed to retrospectively evaluate the predictive performance of qSOFA scoring system in the setting of S. aureus bacteremia which has not been adequately studied previously.

Our findings are consistent with those from prior studies involving other infectious syndromes. Freund et al. prospectively evaluated the prognostic accuracy of qSOFA to predict poor outcomes in patients presenting to the emergency department with suspicion of infection. Similar to our study, they reported 24% of their study population had qSOFA scores ≥2 and that qSOFA had improved predictive performance compared to SIRS with AUROC of 0.73 (95% CI 0.68–0.77) for ICU admission, 0.71 (95% CI 0.66–0.76) for duration in ICU > 72 h, and 0.8 (95% CI 0.74–0.85) for in-hospital mortality [17]. In our study, those with initial qSOFA score of ≥2 were 8 times more likely to have prolonged ICU length of stay (> 72 h) and a fatal outcome when compared to patients with a qSOFA score < 2. We used another severity of illness score, PBS previously validated in SAB to predict mortality [14]. PBS scoring system uses 5 variables, 3 of which overlap with qSOFA (hypotension, mental status, and respiratory status) while the other 2 are temperature and evidence of cardiac arrest; the latter requires laboratory testing. Although similar variables are used in the PBS, parameters such as fever and altered mental status are stratified such that more points are assigned as the measured value is increasingly abnormal, whereas with qSOFA the points are assigned with a simple yes or no. The AUROC was comparable between qSOFA and PBS for each of the predefined outcomes. Considering that both PBS ≥2 and qSOFA ≥2 were shown to significantly predict the three clinical outcomes by multivariable logistic regression analysis, qSOFA would be a more practical tool for use at the bedside given its ease of use compared to PBS.

The goal of the qSOFA scoring system is to identify those at increased risk for clinical deterioration [29], and qSOFA score ≥ 2 had good negative predictive value and specificity in predicting prolonged ICU stay and 30-day all-cause mortality in patients with SAB, similar to prior studies [17, 21, 26]. Inversely, SIRS criteria has high sensitivity but was associated with high rate of false positives. Although qSOFA may lack sensitivity compared to SIRS in capturing patients who died, there were only 9 deaths in the low qSOFA group, of which 7 patients had non-SAB related deaths, including withdrawal of care. As SIRS criteria was criticized for lacking specificity which resulted in over-prescribing of antimicrobials, the increased specificity with qSOFA criteria supports its use to be a favored screening system to identify patients most likely to have poor outcome and therefore needing higher level of care.

MSSA was the predominant pathogen overall. Nonetheless, a significantly higher proportion of patients in the high qSOFA group (42%) was infected with MRSA. Notably, among patients with high qSOFA score, clinical outcomes were worse in those infected with MRSA compared to those with MSSA bacteremia: longer length of hospital stay and microbial persistence beyond 4 days of effective therapy. This is consistent with prior literature showing an association of worse outcomes with MRSA compared to MSSA bacteremia [30]. Despite the outcome difference observed between MSSA and MRSA-infected patients in our study, there was no significant difference seen in MRSA infected patients with high qSOFA scores in terms of initial PBS score, the proportions with high risk sources of SAB (MSSA 25% vs MRSA 34%, p = 0.48), and time to initiation of effective therapy compared to MSSA infected patients with high qSOFA scores. It is possible that inherent difference in antimicrobial efficacy and/or indirect immunomodulatory effects between agents used to treat MRSA and those (beta-lactams) available to treat MSSA bacteremia contributed to the outcome difference [30]. Taken together, results from our study adds to the existing data [17, 18, 20] in support of the use of qSOFA as a bedside tool in identifying non-ICU patients presenting with sepsis (including those later confirmed to have bacteremia due to S. aureus) for early and aggressive management.

Importantly, we found that in patients with persistent bacteremia, a high qSOFA score at initial presentation was associated with a 9 times higher risk of death than those with a low qSOFA. This finding is consistent with our previously published study in that a lack of early response in S. aureus bacteremia was the strongest predictor of treatment failure in a multivariable logistic regression model [31]. A trend towards higher risk of death was observed among those who continued on the same treatment despite lack of early response compared to those who switched to alternative therapy (mortality 38% vs 10%, p = 0.13). In addition, every additional day of persistently positive blood culture with S. aureus, significantly increases the patient’s risk for mortality [32]. Thus, early recognition of high-risk patients using qSOFA could prompt clinicians to make a timely change in management including performing source control procedures earlier in the course to avoid negative outcomes. As the availability of rapid diagnostic technology has significantly shortened time to organism identification for bloodstream infection cases, early recognition of patients at high risk of poor outcome could allow antimicrobial stewardship teams to increase their vigilance of aiding clinicians to obtain early source control, screen for metastatic complications, obtain ID consultation, and initiate optimal antimicrobial therapy as soon as resistance information is known.

Our study has several limitations. As qSOFA scores were not used to prospectively guide therapeutic management in this study, we could not control for all confounding variables that may have affected the clinical outcomes. It is notable that similar proportion of patients between the high and low qSOFA groups received infectious disease consultation and source control procedure. While this study was conducted in the same geographic region, the number of patients included in this study was relatively large and included diverse patient populations such as the elderly and the medically underserved younger populations (age range 19–89 years). Our study population had a higher proportion of male over female patients which is consistent with prior epidemiologic studies, which found SAB to occur more frequently in male patients [33, 34]. We evaluated mortality within 30-days from onset of SAB, which is a common endpoint used in SAB studies while others employed in-hospital mortality as an endpoint in the sepsis literature. Only one patient in our study died past 30 days (died on day 46) after the onset of SAB. Thus, using an alternative definition of mortality did not affect our main analysis. Although at initial presentation of sepsis, the diagnosis of S. aureus bacteremia would not have been known, our study demonstrated that a high qSOFA score measured at initial presentation of sepsis predicted poor outcomes in this large cohort of patients who had confirmed S. aureus bacteremia later. Thus, if prospectively applied, qSOFA would likely demonstrate prognostic capability in patients with sepsis due to a variety of infection types, including bacteremia.

qSOFA is a simple 3-variable bedside tool that is more specific than SIRS and simpler to calculate than PBS in identifying septic patients at initial presentation (later confirmed to have S. aureus bacteremia) who are at high risk for poor outcomes. Patients identified with high qSOFA score should receive aggressive management for infection including possible transfer to higher-level of care in the ICU. Future studies should include prospective evaluation of the utility of qSOFA scoring system in guiding antibiotic selection (e.g. need for MRSA coverage) particularly when combining with use of rapid diagnostic platform to confirm diagnosis of S. aureus bacteremia.