Background
Health and economic burden
Definition and diagnosis
Community-onset sepsis
Methods
Study inclusion criteria
-
Study setting: population- or hospital-based studies considering COS separately
-
Exposure/risk factor: any patient characteristic or clinical parameter (e.g., age, sex, co-morbidity, heart rate, body temperature) associated with a risk of COS
-
Outcome: annual incidence (per population/hospital admission) of COS; risk factors for COS.
Search strategy
Study selection
Data extraction
Quality assessment
Data analysis and synthesis
Results
Literature search and included studies
Study and population characteristics
Study ID year [country] | Study characteristics | Population characteristics | Case identification and data source | Exposure and follow-up |
---|---|---|---|---|
Esteban 2007 [46] Spain |
Design: prospective |
Case definition and criteria used: ACCP/SCCM definitions μ
| Case report forms |
Exposure: Not defined |
Study setting: ICU, hospital ward |
Criteria for COS: 83% sepsis patients had a community-acquired infection (no other details given) |
FU: March 1, 2003 to June 30, 2003 | ||
Geographic scope: city/municipal |
Inclusion criteria: Consecutively admitted patients (age≥18 years) |
FU duration: 4 months | ||
Recruitment years: March 1, 2003 to June 30, 2003 |
Description (n sample size): 15,852 (includes n=702 non-severe sepsis cases) | |||
Ginde 2013 [47] The USA |
Design: retrospective |
Case definition and criteria used: Severe sepsis as concurrent bacterial or fungal infection plus acute organ dysfunction using ICD-9-CM codes (995.92) for infection and acute organ dysfunction | Annual survey of ED visits conducted by the National Centre for Health Statistics (associated with U.S. non-institutional, general and short stay hospitals) |
Exposure: Age, nursing home residence |
Study setting: nursing residence, community-living, ED | ||||
FU: 2005 to 2009 | ||||
Criteria for COS: patients admitted to ED for severe sepsis |
FU duration: 5 years | |||
Geographic scope: Nationwide |
Inclusion criteria: adults (≥18 years) who visited ED (with or without infection other than sepsis) | |||
Recruitment years: 2005–2009 |
Description (n sample size): a nationally representative survey of all ED visits of adults (age ≥ 18 years) in 2005–2009 (n=87,500,000) | |||
Harrison 2006 [11] The UK |
Design: retrospective |
Case definition and criteria used: severe sepsis - infection plus ≥3 SIRS criteria and ≥1 organ dysfunction during the 24-hour period | The Case Mix Programme Database containing data on demographics, case mix, outcome and activity for admissions. Patient data were abstracted by trained data collectors according to precise rules and definitions |
Exposure: not defined |
Study setting: ICU and combined ICU/HDU |
FU: December 2001 to January 2005 | |||
Criteria for COS: as having severe sepsis during the first 24 hours following admission to the critical care unit | ||||
Geographic scope: Nationwide (England, Wales, and Northern Ireland) |
FU duration: 2 years (2002-2004) | |||
Inclusion criteria: adults (≥16 years) presenting with severe sepsis within 24 hours of admission to critical care | ||||
Description (n sample size): n=59,388 (2002), n=59,527 (2003), n=24,905 (2004) | ||||
Recruitment years: December 1995 to January 2005 | ||||
Henriksen 2015a [42] Denmark |
Design: Prospective |
Case definition and criteria used: the ACCP/SCCM criteriaμ
| The hospitals’ patient administrative database and electronic patient records which were manually reviewed. |
Exposure: age, sex |
Study setting: medical ED, ICU, community-based |
Criteria for COS: Patient had an infection at arrival or developed an infection during the first 48 hours after admission. |
FU: September 1, 2010 to August 31, 2011 | ||
FU duration: 12 months | ||||
Geographic scope: city/municipal | ||||
Inclusion criteria: adults (≥15 years) admitted to the medical ED or directly to the medical ICU | ||||
Description (n sample size): A population-based survey of all adult | ||||
Recruitment years: September 1, 2010, to August 31, 2011 | patients admitted to the medical ED at Odense University Hospital, Denmark (n=235,598); n=8,358 hospitalisations | |||
Husak 2010 [48] Canada |
Design: retrospective |
Case definition and criteria used: Severe sepsis (including septic shock) - sepsis complicated by organ dysfunction in at least one of the six organ systems and the ICD codes | Hospital discharge abstract database |
Exposure: Not defined |
Study setting: hospital ward, ED, ICU |
FU: April 1, 2004 to March 31, 2009 | |||
Geographic scope: Nationwide |
Criteria for COS: the majority (79%) of patients with sepsis were admitted via the EDs, while 12.4% were admitted directly through hospitals |
FU duration: 5 years | ||
Recruitment years: April 1, 2004 to March 31, 2009 | ||||
Inclusion criteria: NR | ||||
Description (n sample size): sepsis cases n=26,803 (2004-2005), n=30,587 (2008-2009), n for non-cases (all non-sepsis hospitalizations): NR | ||||
Nygard 2014 [49] Norway |
Design: prospective |
Case definition and criteria used: ACCP/SCCMμ and SCCM/ESICM/ACCP/ATS/SIS | Clinical data were registered prospectively until hospital discharge or in-hospital death using predefined case report forms. Information was collected from medical records, patient charts, and the intensive care electronic monitoring system |
Exposure: not defined |
Study setting: ED, ICU, HDU, and combined ICU/HDU |
FU: January 2008 to December 2008 | |||
Criteria for COS: if they developed severe sepsis within 24 hours of admission to the primary institution | ||||
FU duration: 12 months | ||||
Geographic scope: city/municipal |
Inclusion criteria: patients > 15 years of age hospitalized due to COS, including patients transferred from affiliated hospitals, if they developed severe sepsis within 24 hours of admission to the primary institution | |||
Recruitment years: January 2008 - December 2008 | ||||
Description (n sample size): total N hospitalisations (NR); n=220 severe sepsis cases; non-cases (n): 350,000 population | ||||
Page 2015 [50] The USA |
Design: retrospective |
Case definition and criteria used: severe sepsis was defined using the methodology of Angus et al. identifying hospitalizations with the presence of ICD-9 discharge diagnoses for both a serious infection and organ dysfunction | Hospital discharge data from the UHC representing 300 academic and community hospitals across 42 states. Using medical record review, coders assigned discharge diagnoses for each hospitalization |
Exposure: not defined |
Study setting: NR |
FU: January 1, 2012 to December 31, 2012 | |||
Geographic scope: Nationwide | ||||
FU duration: 12 months | ||||
Recruitment years: January to December 2012 |
Criteria for COS: Hospitalizations with an infection present at admission were subdivided into healthcare-associated (admitted from nursing facility, receiving home healthcare, or were on haemodialysis prior to admission), community acquired (from the community). Those discharged without infections at admission were categorized as hospital-acquired | |||
Inclusion criteria: all hospital discharges from January 1, 2012 to December 31, 2012 | ||||
Description (n sample size): n=3,355,753 hospitalisations; non-cases n (population)=NR | ||||
Wang 2012 [36] REGARDS-sepsis cohort study The USA |
Design: prospective |
Case definition and criteria used: An infection plus ≥ 2 SIRS criteriaμ
| Structured interviews, in-home visits, lab results, monitoring every 6 months, medical and hospital admission records (clinical and lab data) |
Exposure: anthropometric, socio-demographic, dietary, and life-style factors, chronic conditions, and statin use |
Study setting: Community-based, hospital ward, ED |
Criteria for COS: cohort of community-dwelling individuals. The study focused on individuals presenting to the hospital or ED with community-acquired sepsis | |||
Geographic scope: Nationwide |
Inclusion criteria: NR | |||
Description (n sample size): Community-dwelling ≥45 years old individuals in the US (n=30,239) | ||||
FU: 5 February, 2003 to 31 December, 2012 | ||||
Recruitment years: 2003-2007 | ||||
FU duration: ≥ 7 years | ||||
Seymour 2012 [51] The USA |
Design: retrospective |
Case definition and criteria used: Hospitalization with severe sepsis using the ICD-9-CM (995.92 and 785.52). Criteria for COS: pre-hospital severe sepsis | EMS reports computerized database including dispatch, demographic, clinical, and transport data for each incident. EMS data were linked to hospital discharge records |
Exposure: not defined |
Study setting: pre-hospital EMS |
FU: 2000 to 2009 | |||
FU duration: 10 years | ||||
Geographic scope: regional (King County in Washington State) |
Inclusion criteria: all adult EMS encounters involving non-trauma, non–cardiac arrest patients transported from a scene to a receiving hospital by ground ambulance | |||
Recruitment years: 2000-2009 |
Description (n sample size): all 2000-2009 EMS encounters in the area of 1.2 million residents (n=407,176) | |||
Wang 2007 [52] The USA |
Design: retrospective cohort study |
Case definition and criteria used: the criteria by ACCP/SCCMμ and SCCM/ESICM/ACCP/ATS/SIS; ICD-9 codes 990.90 and 995.92 as sepsis and severe sepsis, respectively | The study used the 2001–2004 NHAMCS public use data set which is a national sample of ED and outpatient visits at hospitals across the US |
Exposure: not defined |
FU: 2001 to 2004 | ||||
Study setting: ED |
FU duration: 4 years | |||
Geographic scope: Nationwide |
Criteria for COS: patients with severe sepsis presenting to ED | |||
Inclusion criteria: adults (≥18 years) presenting to EDs during 2001-2004 (with or without severe sepsis) | ||||
Recruitment years: 2001-2004 | ||||
Description (n sample size): n=113,123 ED visits (sample-based 4 years); 4-year national estimate n=331,531,000 ED visits; annual national estimate n=82,883,000 ED visits |
Study ID year [country] | Study characteristics | Population characteristics | Identification and data source |
---|---|---|---|
Henriksen 2015b [43] Denmark |
Study setting: medical ED, ICU, community-based |
Case definition: sepsis (non-severe), severe sepsis, or septic shock according to the ACCP/SCCM criteria. | Data electronically extracted from the patient’s records and validated by trained data. All admissions were manually reviewed. Predefined risk factors retrieved from several population-based registers |
Control definition: all adults (≥15 years) with residence in the hospital catchment-area (N = 235,598) during the study period who had not been hospitalized up to 7 days before the index date. | |||
Geographic scope: city/municipal | |||
Recruitment years: September 1, 2010 to August 31, 2011 |
Criteria for COS: Patient had an infection at arrival within the first 48 hours after admission | ||
Inclusion criteria (cases): All adults (≥15 years) admitted to the medical ICU or ED. | |||
Inclusion criteria (controls): see controls definition and criteria | |||
Exclusion criteria: cases with a prior hospitalization up to 7 days before the current admission. Patients transferred from other hospitals, patients residing outside the hospitals catchment-area at the time of admission and patients who were unidentified throughout the entire course of admission | |||
N cases: 1,713 sepsis of any severity (n=621 non-severe sepsis, n=1,071 severe sepsis, and n=21 septic shock) | |||
N controls: 227,054 | |||
Jovanovich 2014 [44] The USA |
Study setting: tertiary-level care centres and small clinics and hospitals |
Case definition: adults hospitalized for sepsis or severe sepsis; ICD-9 codes (995.91, 995.92) | Electronic health and administrative data |
Control definition: randomly selected adult patients without sepsis diagnosis admitted within the same time period and matched 1:1 with cases by age, sex, race, and season of 25(OH)D measurement | |||
Geographic scope: Inter-State (Utah and Idaho) |
Criteria for COS: community-living adults | ||
Inclusion criteria (cases): NR | |||
Inclusion criteria (controls): NR | |||
Recruitment years: 1 January 2008 and 31 December 2010 |
Exclusion criteria: NR | ||
N cases: 211 | |||
N controls: 211 | |||
Legras 2009 [45] France |
Study setting: ICU, hospital ward |
Case definition: the ACCP/SCCM criteria was used (severe sepsis or septic shock) | Medical histories and data on previous prescriptions obtained from relatives and general practitioner. NSAID use was quantified by listing all the drugs taken during the observation period, and standard interviews were conducted by physicians |
Geographic scope: regional |
Control definition: Participants admitted to hospital for mild bacterial infection (without severe sepsis or septic shock) | ||
Recruitment years: February 2004 to November 2005 |
Criteria for COS: Community-acquired (NR) | ||
Inclusion criteria (cases): Participants >15 years admitted to an ICU with community-onset severe sepsis or septic shock | |||
Inclusion criteria (controls): Participants admitted to hospital for mild bacterial community-acquired (non-sepsis) infection | |||
Exclusion criteria: chronic kidney failure (creatinine clearance <30 ml/min), pregnancy, nosocomial infection, or congenital/acquired immunosuppression (defined as the presence of metastatic neoplasia, haemopathy, aplasia before the onset of sepsis), AIDS (independently of CD4+ T-cell count) and chronic administration of immunosuppressive treatments | |||
N cases: n=152 (n=34 sever sepsis; n=118 septic shock) | |||
N controls: 152 | |||
Wang 2013c [39] REGARDS-sepsis case-control study The USA |
Study setting: community-based, hospital, ED |
Case definition: laboratory confirmed infection plus ≥ 2 SIRS criteria | Structured interviews, in-home visits, lab results, monitoring every 6 months, medical and hospital admission records (clinical and lab data); blood samples collected from fasting subjects at their homes |
Control definition: hospitalized for a serious infection (but did not meet sepsis criteria) | |||
Criteria for COS: cohort of community-dwelling individuals. Presentation to the hospital consisted of the time of Emergency Department triage or admission to inpatient unit (for participants admitted directly to the hospital). To allow for acute changes in the participant's condition during early hospitalization, we used vital signs and laboratory test results for the initial 28 h of hospitalization. Our study focused on individuals presenting to the hospital or ED with community-acquired sepsis. We did not include “hospital-acquired” sepsis developing at later points of hospitalization | |||
Geographic scope: Nationwide | |||
Recruitment years: January 2003- October 2007 | |||
Inclusion criteria (cases): Patients hospitalized for sepsis during the observation period were eligible. | |||
Inclusion criteria (controls): Individuals with serious infection who did not experience a hospitalization for sepsis, matching for age (±5 years), sex, and time epoch | |||
Exclusion criteria: individuals hospitalized for conditions unrelated to infections | |||
N cases: 162 | |||
N controls: 162 |
Cohort studies
Case-control studies
Quality of included studies
Cohort studies
Case-control studies
Review outcomes
Incidence of COS
Study ID country | Cohort and study characteristics | Study design and duration of follow-up | Type of sepsis | Incidence – overall (total cohort) | Methodological quality (high, acceptable, low) | ||
---|---|---|---|---|---|---|---|
N of cases per 100,000 population per year [95% CI] | CIP % per hospitalisation or ED visit per year [95% CI] | IDR (N of cases per 100,000 p-y) [95% CI] | |||||
Esteban 2007 [46] Spain |
Geographic scope
| Prospective cohort study | Non-severe | 367 [352, 384] | 13.28% [12.76, 13.81] | NR | Acceptable quality |
City/municipal | Severe | 104 [96, 113] | 3.76% [3.47, 4.06] | NR | |||
FU: 4 months | |||||||
Setting
| Septic shock | 31 [27, 36] | 1.11% [0.95, 1.28] | NR | |||
ICU, hospital ward | |||||||
Cohort denominator
| |||||||
N (population)=573,149 | |||||||
N (hospitalisations)=15,852 | |||||||
Ginde 2013 [47] The USA |
Geographic scope
| Retrospective cohort study | Severe | NR | 0.40% [0.39, 0.41] | NR | Low quality |
Nationwide | |||||||
Setting
| FU: 5 years | ||||||
ED | |||||||
Cohort denominator
| |||||||
N (population)=NR | |||||||
N (all ED visits)=87,500,000 | |||||||
Harrison 2006 [11] The UK |
Geographic scope
| Retrospective cohort study | Severe | 66 [NR]μ
| 27.87% [27.52, 28.24] | NR | Acceptable quality |
Nationwide | |||||||
Setting
| FU: 2 years | ||||||
ICU, combination of ICU with HDU | |||||||
Cohort denominator
| |||||||
N (population)=NR | |||||||
N (hospitalisations)=59,527 | |||||||
Henriksen 2015a [42] Denmark |
Geographic scope
| Prospective cohort study | All | 727 [693, 762] | NR | 731 [697, 767] | Acceptable quality |
City/municipal | Non-severe | 264 [243, 285] | NR | 265 [245, 287] | |||
Setting
| FU: 1 year | Severe | 455 [428, 482] | NR | 457 [430, 485] | ||
ED, ICU | Septic shock | 9 [6, 13] | NR | 9 [6, 14] | |||
Cohort denominator
| |||||||
N (population )=235,598 | |||||||
N (hospitalisations)=8,358 | |||||||
Husak 2010 [48] Canada |
Geographic scope
| Retrospective cohort study | All | 103 [NR]μ
| NR | NR | Low quality |
Nationwide | Non-severe | 64 [NR]μ
| NR | NR | |||
Setting
| FU: 5 years | Severe | 40 [NR]μ
| NR | NR | ||
ED, ICU, hospital ward | |||||||
Cohort denominator
| |||||||
N (population or hospitalisations)=NR | |||||||
Nygard 2014 [49] Norway |
Geographic scope
| Prospective cohort study | Severe | 50 [NR]μ
| 0.22% [NR]μ
| NR | Low quality |
City/municipal | |||||||
Setting
| FU: 1 year | ||||||
ED, ICU, HDU, combination of ICU with HDU | |||||||
Cohort denominator
| |||||||
N (population or hospitalisations)=NR | |||||||
Page 2015 [50] The USA |
Geographic scope
| Retrospective cohort study | Severe | NR | CA-SS | NR | Low quality |
Nationwide | 5.75% [5.72, 5.77] | ||||||
Setting
| FU: 1 year | HCA-SS | |||||
NR | 2.37% [2.35, 2.38] | ||||||
Cohort denominator
| |||||||
N (population)=NR | |||||||
N (hospitalisations)=3,355,753 | |||||||
Wang 2012 [36] REGARDS-sepsis cohort study 2012-2015β The USA |
Geographic scope
| Prospective cohort study | Non-severe | 514 [489, 539] | NA | 800 [760, 840] | Acceptable quality |
Nationwide | |||||||
Setting
| FU: 9-10 years | ||||||
Hospital ward, ED | |||||||
Cohort denominator
| |||||||
N (population)=30,239 | |||||||
N (hospitalisations)=NR | |||||||
Seymour 2012 [51]The USA |
Geographic scope
| Retrospective cohort study | Severe | NR | Entire 10-year cohort | NR | Low quality |
Regional (within-State) | 3.25% [3.20, 3.31] | ||||||
Setting
| FU: 10 years | One-year cohort | |||||
Pre-hospital emergency medical services | 4.93% [4.73, 5.13] | ||||||
Cohort denominator
| |||||||
N (population)= NR | |||||||
N (emergency encounters)=407,176 | |||||||
Wang 2007 [52] The USA |
Geographic scope
| Retrospective cohort study | Severe | NR | 0.69% [0.61, 0.77] | NR | Low quality |
Nationwide | |||||||
Setting
| FU: 4 years | ||||||
ED | |||||||
Cohort denominator
| |||||||
N (population)=NR | |||||||
N (ED visits)=82,883,000 |
Associations between various factors and the occurrence of sepsis
Study ID country | Geographic scope and setting | Study design Sample size N | Type of sepsis | Risk factor (reference and exposure groups) | Summary measure of association (exposure vs. reference group) 95% CI | Covariates adjusted for | Methodological quality |
---|---|---|---|---|---|---|---|
Age (years)
| |||||||
Ginde 2013 [47] The USA | Nationwide | Retrospective cohort study | Severe | <65 | Ref 1.00 | sex and race/ethnicity | Low quality |
ED | ≥65 | OR=1.00 (0.52, 1.90) | |||||
N (cohort baseline – all ED visits)= 87,500,000 | |||||||
Wang 2012 [36] REGARDS-sepsis cohort study β2012-2015 The USA | Nationwide | Prospective cohort study | Non-severe | 45-54 | Ref 1.00 | Not adjusted (crude) | Acceptable quality |
Hospital ward, ED | 55-64 | HRR=1.44 (1.04, 2.00) | |||||
N (cohort baseline)=30,239 | 65-74 | HRR=2.29 (1.66, 3.16) | |||||
75≤ | HRR=3.87 (2.80, 5.35) | ||||||
Henriksen 2015b [43] Denmark | City/municipal | Case-control study | All | 15-64 | Ref 1.00 | Sex, alcoholism- related conditions, comorbidity, and immunosuppression | Low quality |
ED, ICU | N (cases)=1,713 | 65-84 | OR=3.09 (2.75, 3.48) | ||||
N (controls)=227,054 | ≥85 | OR=6.02 (5.09, 7.12) | |||||
Non-severe | 15-64 | Ref 1.00 | See above | ||||
65-84 | OR=2.15 (1.78, 2.60) | ||||||
≥85 | OR=3.66 (2.74, 4.88) | ||||||
Severe | 15-64 | Ref 1.00 | See above | ||||
65-84 | OR=3.93 (3.39, 4.56) | ||||||
≥85 | OR=7.84 (6.38, 9.63) | ||||||
Sex
| |||||||
Ginde 2013 [47] The USA | Nationwide | Retrospective cohort study | Severe | Female | Ref 1.00 | Age and race/ethnicity | Low quality |
ED | Male | OR=1.13 (0.62, 2.00) | |||||
N (cohort baseline– all ED visits)= 87,500,000 | |||||||
Wang 2012 [36] REGARDS-sepsis cohort study β 2012-2015 The USA | Nationwide Hospital ward, ED | Prospective cohort study analysis | Non-severe | Female | Ref 1.00 | Not adjusted (crude) | Acceptable quality |
N (cohort baseline)=30,239 | Male | HRR=1.30 (1.15, 1.48) | |||||
Henriksen 2015b [43] Denmark | City/municipal | Case-control study | All | Female | Ref 1.00 | Age, alcoholism- related conditions, comorbidity, and immunosuppression | Low quality |
ED, ICU | N (cases)=1,713 | Male | OR=1.01 (0.91, 1.11) | ||||
N (controls)=227,054 | |||||||
Non-severe | Female | Ref 1.00 | See above | ||||
Male | OR=0.89 (0.76, 1.05) | ||||||
Severe | Female | Ref 1.00 | See above | ||||
Male | OR=1.07 (0.95, 1.22) | ||||||
Race/ethnicity
| |||||||
Ginde 2013 [47] The USA | Nationwide | Retrospective cohort study | Severe | Non-Hispanic White | Ref 1.00 | Age and sex | Low quality |
Non-Hispanic Black | OR=1.30 (0.62, 2.60) | ||||||
ED | N (cohort baseline– all ED visits)= 87,500,000 | Hispanic | OR=0.63 (0.23, 1.70) | ||||
Other | OR=2.40 (0.87, 6.50) | ||||||
Wang 2012 [36] REGARDS-sepsis cohort study β 2012-2015 The USA | Nationwide | Prospective cohort study analysis | Non-severe | Black | Ref 1.00 | sex, age, geographic region, education level, income, tobacco, alcohol use, baseline chronic medical conditions, biomarkers | Acceptable quality |
Hospital ward, ED | N (cohort baseline)=30,239 | White | HRR=1.56 (1.38, 1.75) | ||||
Education
| |||||||
Wang 2012 [36] REGARDS-sepsis cohort study β 2012-2015 The USA | Nationwide | Prospective cohort study analysis | Non-severe | ≥College | Ref 1.00 | Not adjusted (crude) | Acceptable quality |
Hospital ward, ED | N (cohort baseline)=30,239 | Some college | HRR=1.41 (1.19, 1.67) | ||||
High school | HRR=1.52 (1.28, 1.80) | ||||||
<High school | HRR=1.88 (1.54, 2.29) | ||||||
Nursing home residence
| |||||||
Ginde 2013 [47] The USA | Nationwide | Retrospective cohort study | Severe | No | Ref 1.00 | Age, sex and race/ethnicity | Low quality |
ED | Yes | OR=2.60 (1.20, 5.60) | |||||
N (cohort baseline)= 87,500,000 |