Original Research

Journal of Intensive Care Medicine

1-11
Epidemiology and Changes in Mortality ª The Author(s) 2017
Reprints and permission:
sagepub.com/journalsPermissions.nav
of Sepsis After the Implementation of DOI: 10.1177/0885066617711882
journals.sagepub.com/home/jic
Surviving Sepsis Campaign Guidelines

Rubén Herrán-Monge, MD1, Arturo Muriel-Bombı́n, MD1,

Marta M. Garcı́a-Garcı́a, MD, PhD1, Pedro A. Merino-Garcı́a, MD1,
Miguel Martı́nez-Barrios, MD2, David Andaluz, MD, PhD3,
Juan Carlos Ballesteros, MD4, Ana Marı́a Domı́nguez-Berrot, MD5,
Susana Moradillo-Gonzalez, MD6, Santiago Macı́as, MD7,
Braulio Álvarez-Martı́nez, MD8, M. José Fernández-Calavia, MD9,
Concepción Tarancón, MD10, Jesús Villar, MD, PhD11,12, and
Jesús Blanco, MD, PhD1,11; on behalf of the GRECIA Network

Abstract
Purpose: To determine the epidemiology and outcome of severe sepsis and septic shock after 9 years of the implementation of
the Surviving Sepsis Campaign (SSC) and to build a mortality prediction model. Methods: This is a prospective, multicenter,
observational study performed during a 5-month period in 2011 in a network of 11 intensive care units (ICUs). We compared our
findings with those obtained in the same ICUs in a study conducted in 2002. Results: The current cohort included 262 episodes
of severe sepsis and/or septic shock, and the 2002 cohort included 324. The prevalence was 14% (95% confidence interval: 12.5-
15.7) with no differences to 2002. The population-based incidence was 31 cases/100 000 inhabitants/year. Patients in 2011 had a
significantly lower Acute Physiology and Chronic Health Evaluation II (APACHE II; 21.9 + 6.6 vs 25.5 + 7.07), Logistic Organ
Dysfunction Score (5.6 + 3.2 vs 6.3 + 3.6), and Sequential Organ Failure Assessment (SOFA) scores on day 1 (8 + 3.5 vs 9.6 +
3.7; P < .01). The main source of infection was intraabdominal (32.5%) although microbiologic isolation was possible in 56.7% of
cases. The 2011 cohort had a marked reduction in 48-hour (7% vs 14.8%), ICU (27.2% vs 48.2%), and in-hospital (36.7% vs 54.3%)
mortalities. Most relevant factors associated with death were APACHE II score, age, previous immunosuppression and liver
insufficiency, alcoholism, nosocomial infection, and Delta SOFA score. Conclusion: Although the incidence of sepsis/septic shock
remained unchanged during a 10-year period, the implementation of the SSC guidelines resulted in a marked decrease in the
overall mortality. The lower severity of patients on ICU admission and the reduced early mortality suggest an improvement in
early diagnosis, better initial management, and earlier antibiotic treatment.

1
Intensive Care Unit, Hospital Universitario Rı́o Hortega, Valladolid, Spain
2
Intensive Care Unit, Hospital Universitario de Burgos, Burgos, Spain
3
Intensive Care Unit, Hospital Clı́nico Universitario, Valladolid, Spain
4
Intensive Care Unit, Complejo Hospitalario de Salamanca, Salamanca, Spain
5
Intensive Care Unit, Complejo Hospitalario de León, León, Spain
6
Intensive Care Unit, Hospital Rı́o Carrión, Palencia, Spain
7
Intensive Care Unit, Hospital General de Segovia, Segovia, Spain
8
Intensive Care Unit, Hospital El Bierzo, Ponferrada, León, Spain
9
Intensive Care Unit, Complejo Hospitalario de Soria, Soria, Spain
10
Intensive Care Unit, Hospital Virgen de la Concha, Zamora, Spain
11
CIBER de Enfermedades Respiratorias, Instituto de Salud Carlos III, Madrid, Spain
12
Multidisciplinary Organ Dysfunction Evaluation Research Network (MODERN), Hospital Universitario Dr Negrin, Las Palmas de Gran Canaria, Spain

Received December 25, 2016. Received revised April 10, 2017. Accepted for publication May 5, 2017.

Corresponding Author:
Rubén Herrán-Monge, Intensive Care Unit, Hospital Universitario Rı́o Hortega, Dulzaina, 2, 47012 Valladolid, Spain.
Email: ruben.herran.monge@gmail.com
2 Journal of Intensive Care Medicine XX(X)
Keywords
severe sepsis, septic shock, organ failure, mortality, epidemiology, Surviving Sepsis Campaign
Introduction
Sepsis is one of the leading causes of admission to intensive
care unit (ICU). Treatment of patients with sepsis implies a
major economic burden for national health services.1-4 Sepsis
is the main cause of acute respiratory distress syndrome and
multi-organ dysfunction in critically ill patients.5 Several epide-
miologic studies have shown a high prevalence in hospitalized
patients. During the year 2011, 57 000 patients with severe
sepsis (SS) were admitted to 124 US academic hospitals, of
whom 69% required ICU admission.6 Prevalence of sepsis varies
among ICUs and according to sepsis definitions used, ranging
from 6% to 30%.2,7 More than 50% of patients with sepsis
develop SS and 25% septic shock (SSh), representing the 15%
of all ICU admissions.8 Population-based incidence of sepsis is
estimated in 300 cases/100 000 inhabitants/year in the United
States3 and 367 cases/100 000 inhabitants/year in Spain, with an
ICU incidence of 44 cases/100 000 inhabitants/year.9
At the beginning of the last decade, sepsis was the second
cause of mortality in noncardiologic ICUs and the 10th global
cause of death in the United States1 with a hospital mortality of
47%.10 This prompted the joint development by several inter-
national scientific societies of the global initiative “Surviving
Sepsis Campaign” (SSC) in 2002. Surviving Sepsis Campaign
was aimed at reducing mortality from SS and SSh by 25% in 5
years by setting recommendations grouped into “bundles”
based on the early recognition, early use of antibiotics, early
goal-directed therapy (EGDT) resuscitation protocol, and early
supportive care in the ICU. The campaign also promoted the
development of educational campaign to improving diagnosis
and use of adequate treatment, educating health-care profes-
sionals, and indeed, building awareness about sepsis. The cam-
paign continues improving over time and the last update edition
was published in 2017.11-14
After the launch of the first SSC guidelines11 in 2004 and
the implementation of the resuscitation and management bun-
dles (with updates in 2008 and 2012),12,13 many observational
studies reported a reduction in mortality over time. 15-26
In 2012, Levy et al27 reported a raw mortality of 41.1% in
European ICUs and 28.3% in US ICUs, with adjusted mortal-
ities of 32.3% and 31.3%, respectively. In 2002, we conducted
an epidemiologic study in a network of Spanish ICUs. We
reported a prevalence of SS of 12.4%, an incidence of 25
cases/100 000 population/year, a mortality of 54.3%, and a
median hospital length of stay (LOS) of 25 days.28 More than
25% of patients died within the first 48 hours, suggesting a
delay in diagnosis, initial resuscitation, and empirical antibiotic
administration. Following that study, SSC was launched and
new treatment and clinical management pathways were
implemented.11,20,29,30 In a nationwide multicenter study in
Spain, Ferrer et al16 reported a mortality reduction after an
educational campaign for implementing SSC guidelines.
A new study is necessary to update data on sepsis-associated
mortality and the impact of clinical guidelines and recommen-
dations in the treatment of SS and SSh that has been in place
until now following the introduction of SSC. Although new
sepsis definition has been developed recently,31 it is unclear
how it may affect the epidemiology and mortality of this
syndrome.
Methods
Objectives
The main goals of our study are (1) to determine the ICU
prevalence and population-based incidence of SS and SSh
admitted to the ICU; (2) to determine the early (48 hours), ICU,
and in-hospital mortalities of SS and SSh of those patients; (3)
to identify prognostic factors related to death that could be used
for designing a prediction model; and (4) to compare the
changes in epidemiology, severity, and outcomes of patients
with sepsis with data from 9 years earlier. The secondary goals
are to know (1) the evolution of organ dysfunction, (2) the
types of infection, and (3) the frequency of other factors asso-
ciated with the development and outcome of sepsis.
Study Design and Definitions
This was an observational, multicenter, prospective study con-
ducted during a 5-month period (February to June, 2011) in 11
medical/surgical ICUs of 10 teaching hospitals in the Spanish
region of Castilla y León (Supplemental Table S1). The study
was approved by the ethics committee of the coordinating cen-
ter (Hospital Universitario Rı́o Hortega, Valladolid, Spain) and
by the institutional review boards of participating hospitals.
Signed informed consent by the patient or his next of kin was
required for participation.
Sepsis was defined according to the 2001 International Con-
sensus Conference criteria.32 Severe sepsis was defined as sep-
sis plus the presence of at least 1 organ dysfunction,
hypoperfusion, or hypotension. Sepsis shock was defined as
sepsis plus hypotension refractory to volume infusion and the
need for vasoactive drugs. All the sepsis episodes as a cause of
admission in the ICU and those developed while patients were
hospitalized for other causes were recorded. Daily screening
for SS and SSh was performed in all patients from the time of
ICU admission. Patients younger than 18 years and those
admitted for ischemic coronary heart disease or cardiac rhythm
disorders were excluded.
Infection was defined according to the Center for Disease
Control and Prevention criteria33 and microbiology identified
or suspected by leukocytes in sterile fluids, perforation of
abdominal viscera, radiographic criteria in chest X-ray for
pneumonia plus purulent secretions, or by high suspicion of
Herrán-Monge et al 3

infection based in clinical findings. Episodes were also classi-

fied according to the mode of acquisition (community, hospital,
or intra-ICU), diagnosis (suspicion, clinical or image findings,
surgical or microbiology identification), infective agents if they
were identified, and organ location. Adequate empiric antibio-
tic treatment was considered according to antibiogram when it
was available.

Data Collection
Data on SS/SSh episodes were collected in standardized data
forms (SDFs) by the local manager physician in every partici-
pating ICU and then sent to the coordinating center for being
included in a database designed for the study. Every datum was
checked by the medical staff of the coordinating center
according to the standardized ranges of physiological and bio-
logical variables, in order to correct extreme variables if it was
necessary before data analysis.
Demographics and clinical variables were recorded. Previ-
ous health state defined by McCabe index34 and comorbidities
was registered. Patients were classified into different cate-
gories according to the patient location at the time of diag-
nosis and source of infection. Severity and organ failure
scores were assessed using Acute Physiology and Chronic
Health Evaluation II (APACHE II) score,35 Logistic Organ
Dysfunction (LOD) System score 36 on day 0 (D0), and
Sequential Organ Failure Assessment (SOFA) score.37 Organ
Figure 1. Episodes of severe sepsis recorded in the patients admitted
dysfunction was defined as score 1 or 2, and organ failure as 3 to the intensive care units (ICUs).
or 4 using the SOFA score. The evolution of organ failure was
determined by sequential measurement of SOFA score on Mann-Whitney U test for nonparametric continuous vari-
days 0 to 7 (D0-D7), 14 (D14), and 28 (D28). We also calcu- ables, and Pearson w2 test for categorical variables. Multi-
lated Delta SOFA D0-D3 and D1-D3. Day zero (D0) was variate logistic regression was performed to determine
defined as the date at the time of diagnosis until 0:00 AM of differences in mortality between 2 cohorts and to determine
the next day. Neurological status was defined using Glasgow independent risk factors associated with in-hospital mortality
Coma Scale score. We also recorded ICU and hospital LOS with variables that reached statistically significant difference
and mortality at ICU discharge, at 28 days, and at hospital or P values around .1. A prognostic model was built by logis-
discharge. A complete list of definitions of variables is tic regression using “backward stepwise” and “all subsets”
available in Supplementary Appendix.32,38 methodology after adjusting for possible confounders. The
risk factors independently associated with mortality are
expressed with odds ratio (OR) and 95% CI. All analysis were
Data Analysis performed with STATA 11.1 (STATA Co, College Station,
Every hospital in the Spanish Public Health System assists a Texas).
health area with a well-known demography. We analyzed data
relative to a geographical area of 8 administrative zones cared Results
by the 10 hospitals involved in the study. The total population
older than 18 years was obtained from the census of the
Prevalence and Incidence
“Instituto Nacional de Estadı́stica.”38 A total of 2916 patients were admitted to all participating ICUs
Continuous variables are expressed as mean (standard during the study period; 1042 (35.7%) were excluded and 262
deviation [SD]) or median (interquartile range [IQR]) and com- episodes of SS/SSh were identified (incidence 14%, 95% CI:
pared using t test and Mann-Whitney U test, as appropriate. 12.5-15.7), with no differences when compared to 2002
Categorical variables are expressed as frequencies and percen- (12.4%, 95% CI: 11.1-13.6; P ¼ .1). The number of episodes
tages with 95% confidence interval (CI) and analyzed by the w2 finally analyzed was 231 in 229 patients: 92.2% were SS as
test for comparison. All tests were 2-tailed with a significance cause of ICU admission and the remaining 7.8% were diag-
level of P < .05. The relationship between risk factors and death nosed during ICU stay (Figure 1).
was first examined by univariate analysis using the 2-sample The population older than 18 years in the geographical area
unpaired Student t test for parametric continuous variables, of participating hospitals in 2011 was 2 025 248 inhabitants
4 Journal of Intensive Care Medicine XX(X)

Table 1. Clinical Characteristics of Both Cohorts and Comparison Results.

Clinical And Demographic Characteristics 2011 229 Patients 231 Episodes 2002 311 Patients 324 Episodes P Value

Age, years, mean (SD) 66.87 (14.22) 63.9 (14.61) .02

Sex, male, n (%) 153 (66.2) 208 (66.9) .87
McCabe index, n (%)
0 64 (27.8) 81 (26) .64
1 101 (44) 131 (42.1) .67
2 56 (24.34) 63 (20.25) .25
3 9 (4) 36 (11.57) <.01
Patient’s category, n (%)
Medical 152 (66.09) 179 (57.56) .04
Urgent surgery 54 (23.48) 93 (29.9) .09
Scheduled surgery 15 (6.52) 23 (7.4) .64
Traumatic 9 (3.91) 16 (5.14) .5
Origin prior ICU admission, n (%)
Medical ward 84 (37) 100 (32.15) .24
Surgical ward 40 (17.62) 67 (21.54) .26
Operating room 21 (9.25) 51 (16.4) .016
Emergency 69 (30.4) 67 (21.54) .02
Other ICU 4 (1.76) 10 (3.22) .3
Others 9 (3.96) 16 (5.14) .52
Comorbidities, n (%)
AIDS 2 (0.86) 4 (1.3) .64
immunosuppression 50 (21.64) 41 (13.18) <.01
Cancer 15 (6.5) 17 (5.46) .61
Heart failure 15 (6.5) 15 (4.82) .4
Renal failure 15 (6.5) 23 (7.39) .68
Chronic respiratory insufficiency 28 (12.12) 46 (14.79) .37
Chronic liver insufficiency 7 (0.3) 14 (4.5) .38
Diabetes 27 (11.68) 22 (7.1) .063
Alcohol abuse 16 (6.92) 25 (8.03) .63
Hypercoagulability 1 (0.43) 2 (0.64) .74
Increased bleeding risk 31 (13.42) 31 (9.97) .21
Abbreviations: ICU, intensive care unit; SD, standard deviation.

where 46% lived in urban areas.39 The estimated cumulative Table 2. Severity Scales and Organ Failure.
population incidence of SS in ICU was 31 cases/100 000
2011 2002 P
inhabitants/year. Severity Scales (n ¼ 231) (n ¼ 324) Value
The demographic and clinical characteristics of patients are
shown in Table 1. In the 2011 cohort, patients were older (63.9 APACHE II, mean (SD) 21.9 (6.67) 25.5 (7.07) <.01
+ 14.2 years) than those from 2002. There was a significantly LOD, mean (SD) 5.6 (3.28) 6.39 (3.65) .01
higher number of medical patients (66% vs 57.6%, P ¼ .04) SOFA on day 1, mean (SD) 8.06 (3.58) 9.66 (3.71) <.01
Number of organ failures, n (%)
and more episodes admitted from the emergency department 0 29 (12.9) 11 (3.5) <.001
(30.4% vs 21.5%, P ¼ .02). Previous health status assessed by 1 68 (30.2) 96 (30.9) .87
McCabe index and comorbidities was similar in both cohorts, 2 86 (38.2) 97 (31.2) .09
but there were more immunosuppressed patients in 2011 3 32 (14.2) 61 (19.7) .1
(21.6% vs 13.2%, P < .01). Patients were less sicker at the time 4 9 (4) 25 (8) .058
of diagnosis, as assessed by lower severity scores (Table 2). 5 1 (0.4) 21 (6.8) <.001
Organ failurea, n (%)
Neurologic 16 (7.1) 37 (11.9) .06
Infection Characteristics Respiratory 98 (43.5) 233 (74.9) <.01
Cardiovascular 176 (78.2) 180 (57.8) <.01
The diagnosis of infection was made by clinical suspicion plus Renal 46 (20.4) 124 (39.9) <.01
compatible radiological images in 26% (95% CI: 20.3-31.6) Hematological 31 (13.7) 69 (22.2) .01
and by surgical findings in 10.8% (95% CI: 6.8-14.8). The Liver 11 (4.9) 40 (12.7) <.01
59.2% (95% CI: 53-65.6) of infections were community
Abbreviations: APACHE II, Acute Physiology and Chronic Health Evaluation II
acquired. Main locations were abdominal (32.5%, 95% CI: Score; LOD, Logistic Organ Dysfunction Score; SD, standard deviation; SOFA,
26.4-38.5) and respiratory (29.9%, 95% CI: 24-35.8), followed Sequential Organ Failure Assessment.
a
by urinary tract (15.2%, 95% CI: 10.5-19.8). There were less Defined as 3 or 4 points in the SOFA score.
Herrán-Monge et al 5

Table 3. Infection Characteristics and Microbiological Isolates.

Infection Characteristics 2011 Cohort (n ¼ 231) 2002 Cohort (n ¼ 324) P Value

Acquisition mode, n (%)

Community 136 (59.13) 167 (53.7) .1
Hospital 76 (33.04) 93 (29.9) .22
Intra-ICU 18 (7.83) 51 (16.4) <.01
Source of infection, n (%)
Abdominal 75 (32.5) 101 (31.5) .1
Lung 69 (29.9) 138 (44.8) <.01
Surgical wound 2 (0.87) 10 (3.1) .06
Urinary tract 35 (15.2) 20 (6.2) <.01
Bloodstream/catheter–related 15 (6.5) 20 (4.6) .97
Soft tissues 12 (5.2) 8 (3.1) .1
Others 23 (10) 14 (4.5) .01
Microbiological isolates (n ¼ 164) (n ¼ 258) P Value
Gram-negative bacteria, n (%) 85 (51.8) 131 (50.8) .15
Gram-negative bacilli (N) 82 129
Escherichia coli (%) 48.8 37.2
Pseudomonas aeuruginosa (%) 8.5 20.9
Acinetobacter baumannii (%) 8.5 10.9
Legionella pneumophila (%) – 7.8
Klebsiella pneumoniae (%) 9.8 3.1
Proteus mirabilis (%) 6.1 3.1
Others gram negative (N) 3 2
Gram-positive bacteria, n (%) 61 (37.2) 112 (43.4) <.01
Gram-positive cocci (N) 52 104
Staphylococcus aureus (%) 20 32.7
Streptococcus pneumonia (%) 17.3 21.2
Enterococcus spp (%) 17.3 9.6
Other Gram positive (N) 9 8
Fungi 14 (8.5) 15 (5.8) .2
Candida albicans 64.3 66.7
Other 2 (0.8) 4 (2.4)
Abbreviation: ICU, intensive care unit.

Gram-positive microorganism identified (31.8% vs 48.5%, P ¼
.002), less respiratory infections (29.9% vs 44.4%, P < .001),
and more urinary tract infections (15.2% vs 6.4%, P < .001) in
the 2011 cohort. There were less ICU-acquired infections
(7.8% vs 16.4%, P ¼ .003).
Microbiological isolation was obtained only in 131 episodes
(56.7%, 95% CI: 50.3-63.1) with 164 microbiological isolates:
50% (95% CI: 42.3-57.7) Gram-negative bacilli, 31.7% (95% CI:
24.6-38.8) Gram-positive cocci, 8.5% (95% CI: 4.3-12.8) fungi,
and 8.5% (95% CI: 4.3-12.8) other isolates. In 131 episodes with
positive microbiological isolates and antibiogram, the empirical
antibiotic treatment was classified as adequate in 80.2% (95% CI:
73.3-87) and wrong in 13.7% (95% CI: 7.8-19.6; Table 3).
Outcome
Median (IQR) of hospital LOS was 25 days (16-43), and ICU
LOS was 8 days (4-18), similar to the figures in 2002. There
were no differences regarding survivors and nonsurvivors, but
survivors from the 2011 cohort stayed fewer days in the ICU
(7 days vs 12 days, P ¼ .0012) and in the hospital (7 days vs
12 days, P ¼ .0012). However, nonsurvivors had longer hos-
pital stays (26.5 days vs 15 days, P ¼ 0.001; Table 4).
Overall, ICU mortality was 27.2% (95% CI: 21.3-32.8)
and in-hospital mortality was 36.7% (95% CI: 30-42.9),
markedly lower than in 2002 (48.2% and 54.3%, respectively,
P < .001) with an absolute reduction of 17.6% and relative
risk reduction of 32.5% (95% CI: 19.7-43.8). Mortality at
28 days was 28% (95% CI: 22.1-33.8). Also, a marked
reduction in early mortality was found when compared
to 2002: 7% (95% CI: 3.7-10.3) vs 14.8% (95% CI: 310.8-
18.7; P ¼ .003). In the multivariate analysis, 2011 cohort
had lower risk of death than 2002 cohort, both crude as
adjusted by cohort, age, APACHE II, location of infection,
SOFA score, and the number of organ failures at admission
on ICU. These differences reached statistical significance in
ICU mortality and near significance in in-hospital mortality
(Table 4).
Differences in the course of hospital mortality in both
cohorts are shown in Figure 2 and cumulative percentage of
mortality on days 0 to 14 and 28 is shown in Figure 3.
Organ Dysfunction
According to the SOFA score, 68.4% of patients had 1 or 2
organ failures at the time of study entry and 18.7% had 3. In
6 Journal of Intensive Care Medicine XX(X)

Table 4. Outcome.

Main Outcomes 2011; (N ¼ 229) 2002; (N ¼ 311) P

Median (IQR) Median (IQR)

Length of stay, days

ICU global 8 (4-18) 10 (4-20) .26
ICU survivors 7 (4-13) 12 (5-23) <.01
ICU nonsurvivors 12 (4-23) 8 (3-18) .13
Hospital global 25 (16-43) 24 (11-44) .13
Hospital survivors 25 (16-51) 35 (22-59) <.01
Hospital nonsurvivors 26.5 (13-38.5) 15 (7-30) <.01
Mortality N (%; 95% CI) N (%; 95% CI)
48 hours 16 (7%; 3.3-9.8) 46 (14.8%; 10.8-18.7) <.01
Day 28 64 (28%; 22.1-33.8) 149 (47.9%; 42.4-53.5) <.01
ICU 62 (27.2%; 21.3-32.8) 150 (48.2%; 42.7-53.8) <.01
Hospital 83 (36.7%; 30-42.5) 169 (54.3%; 48.8-59.9) <.01
Multivariate Analysis Odds ratio 95% CI
Mortality (crude effect)a
ICU 0.35 0.28-0.58 <.01
Hospital 0.48 0.33-0.69 <.01
ICU mortality (adjusted)b
Cohort 2011 0.64 0.42-0.99 .04
Age 0.99 0.97-0.99 .02
APACHE II 0.94 0.90-0.97 <.01
Location of infection 1.33 1.17-1.53 <.01
Organ failure >2 0.88 0.54-1.44 .6
SOFA score >10 0.28 0.17-0.45 <.01
Hospital mortality (adjusted)b
Cohort 2011 0.75 0.49-1.13 .16
Age 0.98 0.96-0.99 <.01
APACHE II 0.94 0.90-0.97 <.01
Location of infection 1.26 1.11-1.42 <.01
Organ failure >2c 0.86 0.53-1.40 .5
SOFA score >10d 0.33 0.20-0.52 <.01
Abbreviations: APACHE II, Acute Physiology and Chronic Health Evaluation II; CI, confidence interval; ICU, intensive care unit; IQR, interquartile range; SOFA,
Sequential Organ Failure Assessment.
a
Crude effect in mortality of cohort 2011 versus 2002. The 2011 cohort has a “protective effect” in mortality regarding 2002 cohort.
b
Mortality adjusted by several factors by multivariate analysis. The 2011 cohort ICU mortality is lower than 2002 cohort with statistical significance and near to
significance in in-hospital mortality. Location of infection didn’t affect the results.
c
Number of organ failures >2 at ICU admission.
d
SOFA score >10 on day 1.

the 2011 cohort, more patients had no organ dysfunction
(12.9% vs 3.5%, P < .01), and there were fewer patients with
4 organ failures (4% vs 8%, P ¼ .05) when compared to 2002
(Table 2). We observed a reduction in the mean SOFA score
from 8.04 (95% CI: 1.06-1.75) on D0 to 4.86 (95% CI: 1.06-
1.75) on D14, and 3.59 (95% CI: 1.06-1.75) on D28. Mean
Delta SOFA (D3-D1) was 1.41 + 2.42. The mean SOFA score
in nonsurvivors was higher (although not significant) and it
remained higher during the first 15 days of evolution.
Risk Factors
The variables associated with death identified in the univariate
analysis are shown in Supplemental Table S2. We made several
prognostic models by logistic regression including age and
severity scales, organ failures, and comorbidities. In the uni-
variate analysis, a category of patients did not reach statistical
significance individually, but we recoded it in medical versus
nonmedical patients (grouping surgical and trauma) and
included it in the models getting results: the risk of death for
nonmedical patients was more than twice that of medical
patients (OR: 2.2). Finally, we selected the model with the
greatest discriminative capacity and a moderate number of
factors to classify the patients. The strongest factors indepen-
dently associated with hospital mortality were APACHE II
score, age, prior immunosuppression, liver insufficiency and
chronic alcohol abuse, nosocomial acquisition of infection,
nonmedical category, and Delta SOFA (D3-D1). However,
although CIs included the 1, its association with mortality
remains uncertain (Table 5). Our prediction model shows an
area under the receiver operating curve of 0.85, with higher
specificity (89.86%) and negative predictive value (85.7%). It
has moderate sensitivity (69.8%) and positive predictive value
(77.2%; Figure 4).
Herrán-Monge et al 7

Figure 2. Mortality evolution in both cohorts. Cumulative mortality from the time of sepsis diagnosis.

Figure 3. Time course of mortality in nonsurvivors. Figure 4. Receiver operating curve of the prognostic model and its
characteristics.

Table 5. Risk Factors Independently Associated With Mortality. Discussion

Risk Factors OR 95% CI The main findings of this study are (1) incidence of SS was
slightly higher than in 2002; (2) although overall mortality
Age 1.05 1.02-1.08 remains high, we observed a significant absolute reduction in
Chronic alcohol abuse 6.35 1.64-24.63
ICU and hospital mortalities; (3) patients with SS required a
Nosocomial acquisition of infection 4.84 2.23-10.48
Previous state of immunosuppression 3.32 1.24-8.9 large ICU and hospital LOS; and (4) risk factors associated
Chronic liver failure 8.94 1.25-64.1 with fatal hospital outcome are APACHE II score, age, immu-
Delta SOFA (3-1) 1.28 1.08-1.51 nosuppression state, previous liver insufficiency and alcohol-
APACHE II Score 1.10 1.03-1.17 ism abuse, nosocomial acquisition, and Delta SOFA (D3-D1).
Not medical category (surgical and trauma) 2.2 0.95-5.04 The unification of diagnostic criteria for sepsis, SS, and
Abbreviations: APACHE II, Acute Physiology and Chronic Health Evaluation II; SSh and the associated multiple system organ failure32 made
CI, confidence interval; OR, odds ratio; SOFA, Sequential Organ Failure possible the evaluation of the epidemiology of sepsis and its
Assessment. progression in the last few years. In Spain, sepsis prevalence
8 Journal of Intensive Care Medicine XX(X)
remains high, and the slight increase observed from 2002 is in
agreement with the predictions made by Angus et al.3 This
progression has been attributed to an increase in the use of
immunosuppressive drugs, malnutrition, alcoholism, cancer,
diabetes, the use of more invasive procedures, and aging of
the ICU patient population.1,2,3 The population-based inci-
dence of SS/SSh in this study is lower than that previously
reported in developed countries, where the estimated inci-
dence is 50 to 100 cases/100 000 inhabitants/year.1,2,40 This
difference could be justified by population differences, dif-
ferent possibilities of access to hospital care, and less SS
detection. In our case, the incidence refers to SS admitted to
ICUs. We did not include patients with withdrawal of life
support. More than half of our study population lived in rural
areas 38 with a low index of immigrant population, less
accessibility to hospital care, and likely with lower rate of
SS detection previous to ICU admission.
With respect to 2002, there were no changes in hospital and
ICU LOS. However, LOS in nonsurvivors was higher than in
survivors, a finding that could be explained because the higher
early mortality in 2002 reduced the global stay of nonsurvivors
with respect to those patients who survived. It is plausible that
the lower early mortality and the application of more aggres-
sive initial resuscitation measures in 2011 prolonged the ICU
and hospital stays of patients who finally died. On the other
hand, survivors had lower LOS and this could be attributed to
less severity at the moment of diagnosis and to a more appro-
priate treatment in the early phases, as recommended by the
SSC guidelines. Intensive care unit and hospital mortalities
were lower than in 2002 and similar to those reported in Eur-
opean and North American ICUs by Levy et al.27 This mortal-
ity reduction is also similar to other studies.16,18,21,40 The
relative risk reduction in our study is lower and mortality is
higher than in the recent study by Kaukonen et al,23 although in
that work, only ICU patients with SS in the first 24 hours were
included and episodes developed later were not collected. Mor-
tality in our study differs from multicenter reports published
recently.24-26 Mortality at 60 days in the different groups of
Protocolized Care for Early Septic Shock (ProCESS) study
ranges from 18.9% to 21% and in the The Australasian Resus-
citation in Sepsis Evaluation study (ARISE) study is around
18.6%. In the Protocolised Management in Sepsis (ProMISE)
study, a multicenter study conducted in Europe, the mortality
is around 29.5%. By contrast, 90-day mortality in different
groups of ProCESS ranges between 30.8% and 33.7%, a fatal-
ity rate similar to our study. Differences in outcome could be
explained by differences in the study design. The above stud-
ies were randomized controlled trials where inclusion and
exclusion criteria prevented the inclusion of a high proportion
of consecutive patients with sepsis and affected the true mor-
tality rate of SS/SSh. Our study’s population had a higher
severity, as assessed by mean APACHE II score, than in the
ProCESS, ARISE, and ProMISE trials.
We believe that mortality reduction in our study could be
attributed to an early recognition of sepsis in conjunction with
the implementation of standardized diagnosis criteria and an
early initial management. Patients in our 2011 cohort had lower
disease severity at the time of diagnosis and lower SOFA score,
LOD, and APACHE II than in 2002. This less degree of
severity could explain the mortality reduction from 14.8% to
7% in the first 24 hours of evolution and the lower 28-day
mortality. Differences in mortality also could be due to differ-
ences in case mix (as we can notice in the results), but in our
opinion, this is another consequence of SSC implementation.
With a better initial recognition and management and earlier
admission to ICU, patients could benefit from adequate man-
agement and had less organ failure than in 2002 cohort, without
delay in the transfer to ICU from emergency instead of from
ward or other areas with no protocolized treatment. In multi-
variate analysis, differences in mortality reached statistical sig-
nificance between the 2 cohorts with “protective” ORs for 2011
cohort against 2002 cohort, adjusted by severity, age, organ
failure, and SOFA score (Table 4). We conducted this study
to assess whether there have been any changes in epidemiology
and main outcomes of sepsis attended in ICU from an epide-
miologic point of view. In the study period, the most relevant
change in sepsis management between the 2 cohorts period has
been the SSC implementation.
As reported elsewhere, implementation of SSC guidelines
through educational campaigns has been associated with a sig-
nificant decrease in mortality.16-18 In our previous work, we
demonstrated the association between the number of organ
failures at the moment of diagnosis of SS/SSh and the evolution
toward multi-organ failure with higher risk of death. In the
present study, by assessing the sequential SOFA score37 and
Delta SOFA D3-D1, we found that Delta SOFA D3-D1 was
independently associated with fatal outcome. The utility of
Delta SOFA as a prognostic factor associated with mortality
has been demonstrated in several studies in the literature in
different ways, Delta SOFA 0 to 24, 0 to 48, and 72 hours.41,42
When we choose Delta SOFA D3-D1, we realize the disadvan-
tage of needing to wait until the third day to calculate it, but in
our study population, we had few deaths in the first 48 hours
(16 patients, 7%) after diagnosis. Those were among the most
severe patients on ICU admission, with clear high risk of death,
as suspected by the high values of the severity scores
(APACHE II and SOFA). Thus, we did not focus in these
patients due to its severity. We tried to build a model to identify
those patients who survive the first 48 hours but remained
severely at risk of dying in the hospital.
Abdominal location was the most frequent source of infec-
tion, followed by the lungs. Urinary tract infections were
associated with a significant risk reduction of mortality in
relation to other locations. Despite the reported increase in
recent years of infections caused by Gram-positive microor-
ganisms,7,11,43-46 the majority of infections in our study were
caused by Gram-negative bacilli. Regarding fungi isolates,
the proportion was low. Globally, these results do not differ
from our 2002 cohort and from those described by Vincent
et al in 2009.47
This study has some strengths. First, ICUs and medical staff
were the same in both study periods, making the results
Herrán-Monge et al
comparable. Second, the geographical area and the number of
patients included in the study are representative, and our find-
ings can be generalizable to the rest of Spain. Our study also
has some limitations. First, the absence of external monitoring
by an independent staff could reduce its internal validity. How-
ever, although it was not possible to check the clinical data in
each participating hospitals in a randomized fashion, the SDFs
were revised exhaustively by a trained staff at the coordinating
center in order to reduce possible errors. Second, the effects of
seasonal variability in sepsis are well known.48 The short
period of our study (5 months) could be affected by this issue,
although since our study period covered the end of winter and
the beginning of summer, it could minimize this bias. Third, we
are unaware of what happened to those hospitalized patients
with sepsis who were not admitted into the ICUs and this could
underestimate the global incidence of sepsis in the population.
Our work focuses in the critically ill patients with sepsis and in
our environment, the majority of them are treated in the ICU.
We already know that intensive care, essentially for all dis-
eases, is gradually improving over time, and we think that the
process of care may have changed over the decade, apart from
the introduction of the SSC. In our research group, to get the
best practice and excellence in care is a constant goal, and to
perform quality improvement programs with educational cam-
paigns and implementing treatment protocols in the ICUs have
shown better results. We realize that treatment of patients in the
2002 cohort was not protocolized as it used to be after the SSC
implementation in the 2011 cohort, so comparison of these
issues is not possible in the same way and it could affect the
interpretation of the results. In a previous publication, we ana-
lyzed the management of a 2011 cohort of patients with sepsis
after the implementation of SSC measures, with better results
in sepsis outcomes associated with better accomplishment of
SSC bundles and process of care, even in the long term. We
observed a reduction of in-hospital mortality from 45.3% to
36.7% after 5 years of an educational campaign, with better
compliance with the SSC resuscitation bundle.49 These results
of decreased mortality with better accomplishment of the SSC
measures have been corroborated by Sánchez et al in Spain in a
similar period, with a reduction in in-hospital mortality from
44% to 32%.50 This fact reinforces the conclusion of the ben-
efits of implementing the SSC bundles over time, although we
can’t conclude this as the unique cause of better outcomes.
Other process of care could affect this better results, for exam-
ple, the implementation of early detection programs with rapid
response teams as part of local improvement initiatives in some
hospitals of our group.
Since the completion of this study, several changes have
occurred in the management of patients with sepsis, including
a recommendation against the use of hydroxyethyl starches
for fluid resuscitation 14 and the withdrawal of activated
protein C from the market after prospective recombinant
human activated protein C worldwide evaluation in severe sepsis
and septic shock (PROWESS-SHOCK) study results.13,19
Moreover, the publication of the 3 large clinical trials showing
no superiority of EGDT in initial resuscitation24-26 has affected
9
the last revision of the SSC guidelines.14 Despite these issues,
early recognition, systematic resuscitation, and prompt anti-
biotic administration will continue being the standard practices
in the management of patients with sepsis in the ICU setting.
Recently, a new sepsis definition has been updated.31 New
prospective studies must be conducted to validate this new
approach and to check whether it requires further changes for
helping in reducing the incidence and mortality of sepsis.
In conclusion, although the incidence of sepsis/SSh
remained unchanged during a 10-year period, the implementa-
tion of the SSC guidelines resulted in a marked decrease in the
overall mortality. The lower severity of patients on ICU admis-
sion and the reduced early mortality suggest an improvement in
early diagnosis, better initial management, and earlier antibio-
tic treatment.
Appendix
Members of Grecia Network: Hospital Universitario Rı́o Hor-
tega de Valladolid (SACYL, Spain): Arturo F. Muriel-Bombin,
Rubén Herrán-Monge, Marta M. Garcı́a-Garcı́a, Pedro A.
Merino-Garcı́a, and Jesús Blanco-Varela. Complejo Asisten-
cial Universitario de Burgos (SACYL, Spain): Maria Jesús
López-Pueyo, José Antonio Fernández-Ratero, Miguel
Martı́nez-Barrios, Fernando Callejo-Torre, Sergio Ossa-
Echeverri, and Arroyo M. Complejo Hospitalario de León
(SACYL, Spain): Demetrio Carriedo-Ule, Ana Marı́a Domı́n-
guez-Berrot, and Francisco J. Dı́az-Domı́nguez. Hospital
Clı́nico Universitario de Valladolid (SACYL, Spain): Fran-
cisco Gandı́a-Martı́nez, Rafael Cı́tores.González, and David
Andaluz. Complejo Asistencial Universitario de Palencia
(SACYL, Spain): Susana Moradillo-Gonzalez. Hospital El
Bierzo, Ponferrada (SACYL, Spain): Braulio Álvarez-Martı́nez.
Complejo Asistencial Universitario de Salamanca (SACYL,
Spain): Noelia Albalá, Elena Pérez-Losada, Juan Carlos
Ballesteros, and Marta Paz-Pérez. Complejo Asistencial de
Segovia (SACYL, Spain): Santiago Macı́as, Rafael Pajares, and
Noelia Recio-Garcı́a. Complejo asistencial de Soria (SACYL,
Spain): M. Mar Gobernado-Serrano, Daniel Moreno-Torres, and
M. José Fernández-Calavia. Complejo Asistencial de Zamora
(SACYL, Spain): Concepción Tarancón, Teresa Álvarez, and
Priscila Carcelén.
Acknowledgments
The authors would like to thank the staff members and nursery from
participating hospitals for their help and collaboration.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to
the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for
the research, authorship, and/or publication of this article: This study
was supported by a research grant from the Government of Castilla y
León, Spain (JCYL; BOCYL-D-26072010).
10
Supplemental Material
The online supplemental tables and appendix are available at
http://journals.sagepub.com/doi/suppl/10.1177/0885066617711882.
References
1. Martin GS, Mannino DM, Eaton S, Moss M. The epidemiology of
sepsis in the United States from 1979 through 2000. N Engl J
Med. 2003;348(16):1546-1554.
2. Martin GS. Sepsis, severe sepsis and septic shock: changes in
incidence, pathogens and outcomes. Expert Rev Anti Infect Ther.
2012;10(6):701-706.
3. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo
J, Pinsky MR. Epidemiology of severe sepsis in the United States:
analysis of incidence, outcome, and associated costs of care. Crit
Care Med. 2001;29(7):1303-1310.
4. Iñigo J, Sendra JM, Diaz R, Bouza C, Sarrı́a-Santamera A.
Epidemiologı́a y costes de la sepsis grave en Madrid. Estudio
de altas hospitalarias. Med Intensiva. 2006;30(5):197-203.
5. Villar J, Pérez-Méndez L, Espinosa E, et al. Serum lipopolysac-
charide binding protein levels predict severity of lung injury and
mortality in patients with severe sepsis. PLoS One. 2009;4(8):
e6818.
6. Walkey AJ, Wiener RS. Hospital case volume and outcomes
among patients hospitalized with severe sepsis. Am J Respir Crit
Care Med. 2014;189(5):548-555.
7. Vincent JL, Sakr Y, Sprung CL, et al; Sepsis Occurrence in Acutely
Ill Patients Investigators. Sepsis in European intensive care units:
results of the SOAP study. Crit Care Med. 2006;34(2):344-353.
8. Brun-Buisson C. The epidemiology of the systemic inflammatory
response. Intensive Care Med. 2000;26(supp 1):S64-S74.
9. Esteban A, Frutos-Vivar F, Ferguson ND, et al. Sepsis incidence
and outcome: contrasting the intensive care unit with the hospital
ward. Crit Care Med. 2007;35(5):1284-1289.
10. Alberti C, Brun-Buisson C, Burchardi H, et al. Epidemiology of
sepsis and infection in ICU patients from an international multi-
centre cohort study. Intensive Care Med. 2002;28(2):108-121.
11. Dellinger RP, Carlet JM, Masur H, et al. Surviving Sepsis Cam-
paign guidelines for management of severe sepsis and septic
shock. Intensive Care Med. 2004;30(4):536-555.
12. Dellinger RP, Levy MM, Carlet JM, et al. Surviving Sepsis Cam-
paign: international guidelines for management of severe sepsis
and septic shock: 2008. Crit Care Med. 2008;36(1):296-327.
13. Dellinger RP, Levy MM, Rhodes A, et al. Surviving Sepsis Cam-
paign guidelines committee including the pediatric subgroup. Sur-
viving Sepsis Campaign: international guidelines for management
of severe sepsis and septic shock: 2012. Crit Care Med. 2013;
41(2):580-637.
14. Rhodes A, Evans L, Alhazzani W, et al. Surviving Sepsis Cam-
paign: international guidelines for management of sepsis and sep-
tic shock: 2016. Intensive Care Med. 2017;43(3):304-377. doi:10.
1007/s00134-017-4683-6.
15. Dombrovskiy VY, Martin AA, Sunderram J, Paz HL. Facing the
challenge: decreasing case fatality rates in severe sepsis despite
increasing hospitalizations. Crit Care Med. 2005;33(11):
2555-2562.
Journal of Intensive Care Medicine XX(X)
16. Ferrer R, Artigas A, Levy MM, et al. Improvement in process of
care and outcome after a multicenter severe sepsis educational
program in Spain. JAMA. 2008;299(19):2294-2303.
17. Levy MM, Dellinger RP, Townsend SR, et al. The Surviving
Sepsis Campaign: results of an international guideline-based
performance improvement program targeting severe sepsis. Crit
Care Med. 2010;38(2):367-74.
18. Castellanos-Ortega A, Suberviola B, Garcia-Astudillo LA, et al.
Impact of the Surviving Ssepsis Campaign protocols on hospital
length of stay and mortality in septic shock patients: results of a
three-year follow-up quasi-experimental study. Crit Care Med.
2010;38(4):1036-1043.
19. Ranieri VM, Thompson BT, Barie PS, et al. Drotrecogin alfa
(activated) in adults with septic shock. N Engl J Med. 2012;
366(22):2055-2064.
20. Gaieski DF, Edwards JM, Kallan MJ, Carr BG. Benchmarking the
incidence and mortality of severe sepsis in the United States. Crit
Care Med. 2013;41(5):1167-1174.
21. Stevenson EK, Rubenstein AR, Radin GT, Wiener RS, Walkey
AJ. Two decades of mortality trends among patients with severe
sepsis: a comparative meta-analysis. Crit Care Med. 2014;42(3):
625-631.
22. Ayala-Ramı́rez OH, Domı́nguez-Berjón MF, Esteban-Vasallo
MD. Trends in hospitalizations of patients with sepsis and
factors associated with inpatient mortality in the region of
Madrid, 2003-2011. Eur J Clin Microbiol Infect Dis. 2014;
33(3):411-421.
23. Kaukonen KM, Bailey M, Suzuki S, Pilcher D, Bellomo R. Mor-
tality related to severe sepsis and septic shock among critically ill
patients in Australia and New Zealand, 2000-2012. JAMA. 2014;
311(13):1308-1316.
24. ProCESS Investigators, Yealy DM, Kellum JA, et al. A rando-
mized trial of protocol-based care for early septic shock. N Engl
J Med. 2014;370(18):1683-1693.
25. The ARISE Investigators and the ANZICS Clinical Trials Group.
Goal-directed resuscitation for patients with early septic shock. N
Engl J Med. 2014;371(16):1496-1506.
26. Mouncey PR, Osborn TM, Power GS, et al. Trial of early, goal-
directed resuscitation for septic shock. N Engl J Med. 2015;
372(14):1301-1311.
27. Levy MM, Artigas A, Phillips GS, et al. Outcomes of the
Surviving Sepsis Campaign in intensive care units in the USA
and Europe: a prospective cohort study. Lancet Infect Dis. 2012;
12(12):919-924.
28. Blanco J, Muriel-Bombı́n A, Sagredo V, et al. Incidence, organ
dysfunction and mortality in severe sepsis: a Spanish multicentre
study. Crit Care. 2008;12(6):R158.
29. Bernard GR, Vincent JL, Laterre PF, et al. Efficacy and safety of
recombinant human activated protein C for severe sepsis. N Engl
J Med. 2001;344(10):699-709.
30. Annane D, Sébille V, Charpentier C, et al. Effect of treatment
with low doses of hydrocortisone and fludrocortisone on mortality
in patients with septic shock. JAMA. 2002;288(7):862-871.
31. Singer M, Deutschman CS, Seymour CW, et al. The Third
International Consensus Definitions for Sepsis and Septic Shock
(Sepsis – 3). JAMA. 2016;315(8):801-810.
Herrán-Monge et al
32. Levy MM, Fink MP, Marshall JC, et al. 2001 SCCM/ESICM/
ACCP/ATS/SIS international sepsis definitions conference. Crit
Care Med. 2003;31(4):1250-1256.
33. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM. CDC
definitions for nosocomial infections. Am J Infect Control. 1988;
16(3):128-140.
34. McCabe WR, Jackson GG. Gram negative bacteremia: etiology
and ecology. Arch Intern Med. 1962;110(6):847-852.
35. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE
II: a severity of disease classification system. Crit Care Med.
1985;13(10):818-829.
36. Le Gall JR, Klar J, Lemeshow S, et al. The Logistic Organ Dysfunc-
tion System. A new way to assess organ dysfunction in the intensive
care unit. ICU scoring group. JAMA. 1996;276(10):802-810.
37. Vincent JL, Moreno R, Takala J, et al. The SOFA (Sepsis Related
Organ Failure Assessment) score to describe organ/dysfunction/
failure. On behalf of the Working Group on Sepsis-Related Prob-
lems of the European Society of Intensive Care Medicine. Inten-
sive Care Med. 1996;22(7):707-710.
38. Bernard GR, Artigas A, Brigham KL, et al. The American-
European consensus conference on ARDS: definitions, mechan-
isms, relevant outcomes and clinical trial coordination. Am J
Respir Crit Care Med. 1994;149(3 pt 1):818-824.
39. Instituto Nacional de Estadı́stica. Censo de Población y Vivien-
das. 2011. www.ine.es. Accessed May 26, 2017.
40. Danai P, Martin GS. Epidemiology of sepsis: recent advances.
Curr Infect Dis Rep. 2005;7(5):329-334.
41. Jones AE, Trzechiak S, Kline JA. The Sequential Organ Failure
Assessment score for predicting outcome in patients with severe
sepsis and evidence of hypoperfusion at the time of emergency
department presentation. Crit Care Med. 2009;37(5):1649-1654.
11
42. Lopes Ferrerira F, Peres Bota D, Bross A, Mélot C, Vincent JL.
Serial evaluation of the SOFA score to predict outcome in criti-
cally ill patients. JAMA. 2001;286(14):1754-1758.
43. Mayr FB, Yende S, Angus DC. Epidemiology of severe sepsis.
Virulence. 2014;5(1):4-11.
44. Azkárate I, Sebastián R, Cabarcos E, Choperena G, Pascal M,
Salas E. Registro observacional y prospectivo de sepsis grave/
shock séptico en un hospital terciario de la provincia de
Guipúzcoa. Med Intensiva. 2012;36(4):250-256.
45. Annane D, Aegerter P, Jars-Guincestre MC, Guidet B;
CUB-Réa Network. Current epidemiology of septic shock: the
CUB-Réa Network. Am J Respir Crit Care Med. 2003;168(2):
165-172.
46. Brun-Buisson C, Meshaka P, Pinton P, Vallet B; EPISEPSIS
Study Group. EPISEPSIS: a reappraisal of the epidemiology and
outcome of severe sepsis in French intensive care units. Intensive
Care Med. 2004;30(4):580-588.
47. Vincent JL, Rello J, Marshall J, et al. International study of the
prevalence and outcomes of infection in intensive care units.
JAMA. 2009;302(21):2323-2329.
48. Danai PA, Sinha S, Moss M, Haber MJ, Martin GS. Seasonal
variation in the epidemiology of sepsis. Crit Care Med. 2007;
35(2):410-415.
49. Herrán-Monge R, Muriel-Bombı́n A, Garcı́a-Garcı́a MM, et al.
Mortality reduction and long-term compliance with Surviving
Sepsis Campaign: a nationwide multicenter study. Shock. 2016;
45(6):598-606.
50. Sánchez B, Ferrer R, Suarez D, et al. Declining mortality due to
severe sepsis and septic shock in Spanish intensive care units: a
two-cohort study in 2005 and 2011. Med Intensiva. 2017;41(1):
28-37.