Background
Methods
Data sources and search strategy
Inclusion criteria
Data extraction, definitions, and outcomes
Quality assessment
Data synthesis and statistical analysis
Results
Study search
Author, year [reference] | Design | Country | Total no. of patients | Mean age, years | Male sex (%) | Location | Overall mortality rate (%) | qSOFA score recorded | Measured mortality | Participant selection | Primary study aim |
---|---|---|---|---|---|---|---|---|---|---|---|
April et al., 2017 [18] | Retrospective, single-center, cohort study | USA | 214 | 68 | 59 | ED | 18.2 | Worst values during ED stay | In-hospital mortality | Suspected infection, admitted to ICU | Comparison of prognostic accuracy of qSOFA and SIRS for predicting in-hospital mortality |
Askim et al., 2017 [19] | Prospective, single-center, observational study | Norway | 1535 | 62 | 53 | ED | 4.4 | ED arrival | 7- and 30-day mortality | Suspected infection | Clinical usefulness of qSOFA to predict severe sepsis and 7- and 30-day mortality |
Chen et al., 2016 [20] | Retrospective, single-center, observational study | China | 1631 | 73 | 59 | ED | 33 | ED arrival | 28-day mortality | Community-acquired pneumonia | Comparison of prognostic performance of qSOFA, CRB-65, and CRB |
Churpek et al., 2017 [21] | Retrospective, single-center, observational study | USA | 30,677 | 58 | 47 | ED, ward | 5.4 | At time of initial suspicion of infection | In-hospital mortality | Suspected infection | Comparison of qSOFA with other commonly used early warning scores for in-hospital mortality |
Donnelly et al., 2017 [22] | Retrospective, multicenter, cohort study | USA | 2593 | 67 | 40 | NA | 11.3 | Worst values within 28 h of hospital admission | 28-day and 1-year mortality | Suspected infection | Incidence and long-term outcomes of patients diagnosed with sepsis and septic shock |
Dorsett et al., 2017 [23] | Retrospective, single-center, observational study | USA | 152 | NA | NA | ED | NA | Prehospital, upon ED arrival, and during ED stay | NA | Suspected infection | Prehospital qSOFA score in early identification of patients with severe sepsis or septic shock |
Finkelsztein et al., 2017 [24] | Prospective, single-center, cohort study | USA | 151 | 64 | 55 | ED, ward | 19 | Within 8 h before ICU admission | In-hospital mortality | Suspected infection, admitted to medical ICU | Comparison of discriminatory capacity of qSOFA vs. SIRS criteria for predicting in-hospital mortality and ICU-free days |
Forward et al., 2017 [25] | Retrospective, single-center, observational study | Australia | 162 | NA | NA | Non-ICU | 15.5 | Within 24 h of deterioration | In-hospital mortality | Suspected infection | Comparison of prognostic performance of qSOFA, SIRS, and SK criteria |
Freund et al., 2017 [26] | Prospective, multicenter, cohort study | Europe | 879 | 67 | 53 | ED | 8.4 | Worst values during ED stay | In-hospital mortality | Suspected infection | Validation of qSOFA as mortality predictor comparing SIRS with SOFA |
Giamarellos-Bourboulis et al., 2017 [27] | Retrospective, multicenter, cohort study | Greece | 3436 | NA | NA | ED, ward | 25.2 | Initial values measured during admission to ED | In-hospital mortality | Suspected or confirmed infection | Sensitivity of qSOFA for early assessment of mortality and organ dysfunction |
Henning et al., 2017 [28] | Post hoc analysis | USA | 7637 | 58 | 50 | ED | 14.2 | Worst values during ED stay | In-hospital mortality | Suspected infection | Performance of qSOFA predicting in-hospital mortality |
Huson et al., 2017 [29] | Retrospective, single-center, observational study | Gabon | 329 | 34 | 38 | Non-ICU | 4.5 | At time of initial suspicion of infection | In-hospital mortality | Suspected infection | Predictive value of qSOFA score for mortality |
Hwang et al., 2017 [30] | Retrospective, single-center, cohort study | South Korea | 1395 | 65 | 56 | ED | 15 | ED arrival and within 3, 6, and 24 h | In-hospital and 28-day mortality | Severe sepsis or septic shock | Diagnostic performance of positive qSOFA score for predicting 28-day mortality among critically ill patients with sepsis |
Kim et al., 2017 [31] | Retrospective, single-center, observational study | South Korea | 615 | 54 | 33 | Non-ICU | 3.2 | At time of initial suspicion of infection | 28-day mortality | Neutropenic fever | Predictive performance of qSOFA as screening tool for sepsis, mortality, and ICU admission |
Kolditz et al., 2017 [32] | Retrospective, multicenter, observational study | Germany | 9327 | 64 | 56 | Non-ICU | 3.0 | At time of initial suspicion of infection | 30-day mortality | Community-acquired pneumonia | Comparison of qSOFA and CRB-65 for risk prediction |
Mellhammar et al., 2017 [33] | Retrospective population-based study | Sweden | 339 | NA | NA | Non-ICU | NA | Within ± 12 h from initiation of antibiotic therapy | NA | Suspected infection | Incidence of sepsis with organ dysfunction |
Park et al., 2017 [34] | Retrospective, single-center, observational study | South Korea | 1009 | 67 | 45 | ED | 15.8 | ED arrival | In-hospital mortality | Suspected infection | Comparison of performance of qSOFA and SIRS to predict development of organ failure |
Peake et al., 2017 [35] | Post hoc analysis | Australia | 1591 | 63 | 60 | ED | 18.7 | Worst values during ED stay | 90-day mortality | Early septic shock | Exploration of utility and potential effects of new Sepsis-3 definitions |
Quinten et al., 2017 [36] | Prospective, single-center, observational study | The Netherlands | 193 | 60 | 56 | ED | 4.1 | Initial values measured during admission to ED | In-hospital, 28-day, and 6-month mortality | Suspected or confirmed infection | Comparison of predictive performance of qSOFA, CIS, and PIRO score for ICU admission |
Ranzani et al., 2017 [37] | Retrospective, two-center, cohort study | Spain | 6874 | 66 | 62 | ED | 6.4 | ED arrival | In-hospital mortality | Community-acquired pneumonia | Comparison of predictive performance of SIRS, qSOFA, CRB, mSOFA, and CURB-65 for in-hospital mortality |
Seymour et al., 2016 [6] | Retrospective, multicenter, cohort study (in the UPMC validation cohort) | USA | 66,522 | 61 | 43 | ED, ward | 2.8 | At time of initial suspicion of infection | In-hospital mortality | Suspected infection | Comparison of performance of qSOFA, SIRS, SOFA, and MODS score to predict sepsis |
Wang et al., 2016 [38] | Retrospective, single-center, observational study | China | 477 | 73 | 62 | ED | 27.4 | ED arrival | 28-day mortality | Suspected infection | Performance of qSOFA for predicting mortality and ICU admission |
Williams et al., 2017 [39] | Retrospective, single-center, observational study | Australia | 8871 | 49 | 51 | ED | 8.7 | Worst values during ED stay | 30-day and 1-year mortality | Suspected infection | Comparison of diagnostic accuracy of SIRS and qSOFA for organ dysfunction and mortality |
Diagnostic accuracy for in-hospital mortality using positive qSOFA scores and SIRS criteria
Meta-regression for positive qSOFA scores in predicting in-hospital mortality
Sensitivity | Specificity | |||||
---|---|---|---|---|---|---|
Variable | No. of studies | No. of patients | Adjusted (95% CI) | P value | Adjusted (95% CI) | P value |
Study design | ||||||
Prospective | 4 | 2759 | 0.59 (0.32–0.86) | 0.68 | 0.81 (0.64–0.99) | 0.30 |
Retrospective | 16 | 141,778 | 0.49 (0.36–0.62) | 0.84 (0.76–0.91) | ||
Study location | ||||||
USA | 6 | 107,795 | 0.69 (0.53–0.85) | 0.13 | 0.70 (0.52–0.87) | <0.01 |
Other countries | 14 | 36,742 | 0.42 (0.30–0.54) | 0.87 (0.81–0.93) | ||
No. of patients | ||||||
≥ 1500 | 10 | 5424 | 0.39 (0.25–0.54) | 0.13 | 0.89 (0.83–0.95) | 0.84 |
< 1500 | 10 | 139,113 | 0.62 (0.47–0.77) | 0.74 (0.62–0.86) | ||
Overall mortality, % | ||||||
≥ 10% | 10 | 18,715 | 0.54 (0.38–0.70) | 0.68 | 0.77 (0.66–0.89) | 0.01 |
< 10% | 10 | 125,822 | 0.47 (0.31–0.63) | 0.88 (0.80–0.95) | ||
Location of enrollment | ||||||
Only ED | 11 | 30,725 | 0.47 (0.31–0.63) | 0.59 | 0.85 (0.76–0.94) | 0.27 |
Other non-ICU | 9 | 113,812 | 0.55 (0.38–0.72) | 0.81 (0.69–0.92) | ||
Timing of the qSOFA score measurement | ||||||
At time of initial suspicion of infection | 13 | 124,030 | 0.39 (0.28–0.51) | 0.01 | 0.88 (0.82–0.94) | 0.95 |
Worst values | 7 | 20,507 | 0.71 (0.57–0.85) | 0.72 (0.56–0.87) | ||
Disease severity | ||||||
Suspected or confirmed infection | 17 | 142,776 | 0.46 (0.34–0.57) | 0.18 | 0.87 (0.82–0.92) | <0.01 |
Sepsis or septic shock | 3 | 1761 | 0.74 (0.53–0.96) | 0.49 (0.23–0.76) | ||
Source of infection | ||||||
Suspected or confirmed infection | 16 | 126,080 | 0.58 (0.47–0.69) | 0.08 | 0.79 (0.70–0.87) | 0.66 |
Specific infectious entity (community-acquired pneumonia or neutropenic fever) | 4 | 18,457 | 0.24 (0.08–0.40) | 0.94 (0.89–1.00) |
Covariates | Coefficient | SE | RDOR (95% CI)a | P valueb |
---|---|---|---|---|
Study location | 0.04 | 0.15 | 1.04 (0.75–1.44) | 0.79 |
Overall mortality ≥ 10% | −0.34 | 0.14 | 0.71 (0.53–0.96) | 0.03 |
Timing of qSOFA score measurement | −0.53 | 0.159 | 0.59 (0.43–0.81) | <0.01 |
Disease severity | −0.44 | 0.26 | 0.64 (0.37–1.12) | 0.11 |