Articles

Effect of procalcitonin-guided antibiotic treatment on

mortality in acute respiratory infections: a patient level
meta-analysis
Philipp Schuetz*, Yannick Wirz*, Ramon Sager*, Mirjam Christ-Crain, Daiana Stolz, Michael Tamm, Lila Bouadma, Charles E
Luyt, Michel Wolff, Jean Chastre, Florence Tubach, Kristina B Kristoffersen, Olaf Burkhardt, Tobias Welte, Stefan Schroeder,
Vandack Nobre, Long Wei, Heiner C Bucher, Djillali Annane, Konrad Reinhart, Ann R Falsey, Angela Branche, Pierre Damas,
Maarten Nijsten, Dylan W de Lange, Rodrigo O Deliberato, Carolina F Oliveira, Vera Maravić-Stojković, Alessia Verduri,
Bianca Beghé, Bin Cao, Yahya Shehabi, Jens-Ulrik S Jensen, Caspar Corti, Jos A H van Oers, Albertus Beishuizen, Armand R J
Girbes, Evelien de Jong, Matthias Briel*, Beat Mueller
Summary Background In February, 2017, the US Food and Drug Administration approved the blood infection marker
procalcitonin for guiding antibiotic therapy in patients with acute respiratory infections. This meta-analysis of patient
data from 26 randomised controlled trials was designed to assess safety of procalcitonin-guided treatment in patients with
acute respiratory infections from different clinical settings.
Methods Based on a prespecified Cochrane protocol, we did a systematic literature search on the Cochrane Central
Register of Controlled Trials, MEDLINE, and Embase, and pooled individual patient data from trials in which patients
with respiratory infections were randomly assigned to receive antibiotics based on procalcitonin concentrations
(procalcitonin-guided group) or control. The coprimary endpoints were 30-day mortality and setting-specific treatment
failure. Secondary endpoints were antibiotic use, length of stay, and antibiotic side-effects.
Findings We identified 990 records from the literature search, of which 71 articles were assessed for eligibility after
exclusion of 919 records. We collected data on 6708 patients from 26 eligible trials in 12 countries. Mortality at 30 days
was significantly lower in procalcitonin-guided patients than in control patients (286 [9%] deaths in 3336 procalcitonin-
guided patients vs 336 [10%] in 3372 controls; adjusted odds ratio [OR] 0·83 [95% CI 0·70 to 0·99], p=0·037). This
mortality benefit was similar across subgroups by setting and type of infection (p
interactions
Lancet Infect Dis 2018; 18: 95–107
Published Online October 13, 2017 http://dx.doi.org/10.1016/ S1473-3099(17)30592-3
See Comment page 11
*Contributed equally as first authors
Medical University Department, Kantonsspital Aarau, Aarau, Switzerland (Prof P Schuetz MD, Y Wirz MD, R Sager
MD, Prof B Mueller MD); Faculty of Medicine, University of Basel, Basel, Switzerland (Prof P Schuetz, >0·05), although
mortality was very low in primary care and in patients with acute bronchitis. Procalcitonin guidance was also associated
with a
M Christ-Crain MD, D Stolz MD, M Tamm MD, H C Bucher MD, M Briel MD, Prof B Mueller); 2·4-day reduction in
antibiotic exposure (5·7 vs 8·1 days [95% CI –2·71 to –2·15], p<0·0001) and a reduction in
Divis
ion of Endocrinology, antibiotic-related side-effects (16% vs 22%, adjusted OR 0·68 [95% CI 0·57 to 0·82], p<0·0001).
Diabetology and Clinical Nutrition (M Christ-Crain),
Interpretation Use of procalcitonin to guide antibiotic treatment in patients with acute respiratory infections reduces
Clinic of Pneumology and Pulmonary Cell Research antibiotic exposure and side-effects, and improves survival.
Widespread implementation of procalcitonin protocols in
(D
Stolz, M Tamm), and Basel patients with acute respiratory infections thus has the potential to improve antibiotic
management with positive effects on clinical outcomes and on the current threat of increasing antibiotic multiresistance.
 Institute for Clinical Epidemiology and Biostatistics (H C Bucher, M Briel), University Hospital Basel, Basel, Funding
National Institute for Health Research.
Switzerland; Service de Réanimation Médicale, Introduction The US Food and Drug Administration approved
Université Paris 7-Denis-Diderot, Assistance Publique Hôpitaux de Paris the blood infection biomarker procalcitonin for
the
(AP-
HP), Paris, France purpose of guiding antibiotic therapy in the context of acute respiratory infections and sepsis in February,
2017.1 Procalcitonin is a calcitonin-related gene product
(L Bouadma MD, M Wolff MD); Service de Réanimation Médicale, Université Paris 6-Pierre-et-Marie-Curie, Paris,
expressed by human epithelial cells in response to
Fran
ce (C E Luyt MD, bacterial infections and is conversely downregulated during viral infections.2,3 Study findings have shown
that procalcitonin concentrations fall rapidly during recovery
J Chastre MD); Département de Biostatistique, Santé publique et Information médicale, AP-HP, Hôpital from acute bacterial
infections.4 As a surrogate marker of host response to bacterial infections, procalcitonin has therefore been proposed as an
adjunct to traditional
Pitié-Salpêtrière, Université Pierre et Marie Curie, Sorbonne Universités, Paris, France (Prof F Tubach MD); clinical and
diagnostic parameters in helping to manage
Depa
rtment of Infectious patients presenting with clinical symptoms suggestive of systemic infections and to guide antibiotic
prescribing practices.5
Diseases, Aarhus University Hospital, Aarhus, Denmark (K B Kristoffersen MD); Acute respiratory tract illnesses are one of the
leading causes of adult hospital admissions and death worldwide, and are associated with antibiotic overuse.6 Although more
than 40% of respiratory infections have a viral cause, imprecise bacterial diagnostics and provider concerns about co-infection
prompt antibiotic prescription in most cases.7 Several trials have reported significant reductions in antibiotic exposure, when
procalcitonin was used to guide decisions about initiation of antibiotics in low-risk patients (eg, patients with a clinical syndrome
of bronchitis in the emergency department) and duration of treatment in high-risk patients (eg, in patients with pneumonia).8
However, although one trial9 found a reduction in mortality associated with procalcitonin-guided antibiotic stewardship in the
intensive care unit (ICU), conclusive evidence on the safety of this approach across clinical settings and different types of
respiratory infections has
www.thelancet.com/infection Vol 18 January 2018 95
Articles
Department of Pulmonary Medicine, Medizinische
Research in context Hochschule Hannover, Member of the German Center of Lung
Evidence before this study
Research, Hannover, Germany
Use of the blood infection marker procalcitonin has gained (O Burkhardt MD, Prof T Welte MD); Department of
Anesthesiology and Intensive Care Medicine,
much attention in the past 10 years as adjunct to clinical judgment in discriminating viral and bacterial infections and guiding
both prescription and duration of antibiotic therapy. Krankenhaus Dueren, Dueren,
Several individual trials showed positive effects with a Germany (Prof S Schroeder MD);
Department of Intensive Care, Hospital das Clinicas da Universidade Federal de Minas
reduction of antibiotic exposure in patients with respiratory infections. Yet, there is ongoing concern about safety of this
approach regarding mortality. Previous meta-analyses reported Gerais, Belo Horizonte, Brazil
no significant effect on mortality, but confidence intervals were (V Nobre MD); Department of
Internal and Geriatric Medicine, Shanghai Jiao Tong University, Affiliated Sixth People’s
large and harm could thus not be excluded. Based on a protocol previously published in the published in the Cochrane Library,
we did a systematic literature search on the Cochrane Central Hospital, Shanghai, China
Register of Controlled Trials (CENTRAL; January, 2017, issue 1), (L Wei MD); Critical Care
Department, Hôpital Raymond Poincaré, AP-HP, Faculty of Health Science Simone Veil,
MEDLINE (1966 to February, 2017), and Embase (1980 to February, 2017). We searched PubMed using the search terms
“Calcitonin”, “Procalcitonin”, “ProCT” and UVSQ–University Paris Saclay,
“Anti-Bacterial Agents”, “antibiotic”, “Antibiotics”, Garches, France (D Annane MD);
Department of Anesthesiology and Intensive Care Medicine, Jena University Hospital, Jena,
“antibacterial, “anti-bacterial”, “amoxicillin, “penicillin”, “ampicillin”, “cotrimoxazole”, “chloramphenicol”, “trimethoprim”,
“sulphamethoxazole”, “tmp smx” and Germany (Prof K Reinhart MD);
“Biomarkers”, “Marker”, “Level”, “levels”, Guide”, “Guidance” Department of Medicine, University of
Rochester, Rochester, NY, USA (A R Falsey MD,
For other data sources, we used similar key search terms as above. Individual patient data were collected from eligible
randomised controlled trials that assessed adults with a clinical diagnosis of upper or lower acute respiratory tract infection.
Added value of this study This study showed substantial relative and absolute reductions in antibiotic use in patients with
respiratory infections managed by procalcitonin protocols compared with patients in the control group. Although such reductions
were found in previous research, importantly in this large cohort of patients, we also found an improvement in clinical outcomes,
namely a reduction in 30-day mortality and antibiotic side-effects. This analysis is the first report, to our knowledge, that shows
clinical benefits beyond antibiotic reductions with the use of procalcitonin protocols.
Implications of all the available evidence This report integrates most of the available evidence on procalcitonin in patients with
acute respiratory infection from randomised trials. Given the positive results regarding antibiotic reduction and improvements in
clinical outcomes, this report strengthens the rationale to use procalcitonin to and “randomised controlled trial”, “controlled
clinical trial”,
support antibiotic stewardship decisions in patients with acute
“randomized”, “randomised”, “placebo”, “drug therapy”,
respiratory infections
. “randomly”, “trial”, “groups”, but not
“animals”, “not humans”. Prof A Branche MD); Department of General Intensive Care, University Hospital of Liege,
Domaine Universitaire de Liège, Liège, Belgium (P Damas MD);
been impeded by insufficient statistical power in most individual trials.9 Moreover, previous meta-analyses10–12 concluded that
 although procalcitonin use was effective at University Medical Centre,
reducing antibiotic exposure, results about the effect of University of Groningen, Groningen,
Netherlands (M Nijsten MD); University Medical Center Utrecht and
procalcitonin-guided antibiotic stewardship on clinical outcomes were inconclusive. These meta-analyses, however, were based
on aggregate data rather than University of Utrecht, Utrecht, Netherlands (D W de Lange MD); Critical Care Unit, Hospital
Israelita Albert
individual patient-level data, restricting the ability to harmonise outcome definitions and to assess differences between
subgroups, and also had a more narrow focus Einstein, São Paulo, Brazil
and thus only included a limited number of trials. (R O Deliberato MD); Department of Internal
Medicine, School of Medicine, Universidade Federal de Minas
We therefore did a search and meta-analysis of individual patient data from 26 randomised-controlled trials,9,13–37 based on a
prespecified Cochrane protocol,38,39 Gerais, Belo Horizonte, Brazil
to comprehensively and definitively assess the safety of (C F Oliveira MD); Immunology
Laboratory, Dedinje Cardiovascular Institute, Belgrade, Serbia
using procalcitonin to guide antibiotic decisions in patients with respiratory illnesses from different clinical settings and with
different types of respiratory infections. (V Maravić-Stojković MD);
This analysis is an update of a previous meta-analysis Section of Respiratory Medicine, Department of
Medical and Surgical Sciences,
published in 2012,39 and an extended version of this review will be published in the Cochrane Library.
University Polyclinic of Modena, University of Modena
Methods and Reggio Emilia, Modena, Italy (A Verduri MD, B Beghé MD); Center for Respiratory
Diseases,
Search strategy and selection criteria For this systematic review and meta-analysis, trial selection and data collection was done
based on a protocol Department of Pulmonary and
published in the Cochrane Library and the report prepared Critical Care Medicine,
according to PRISMA individual patient data guidelines.40,41
96 www.thelancet.com/infection Vol 18 January 2018
We collected individual patient data from eligible randomised controlled trials that assessed adults with a clinical diagnosis of
upper or lower acute respiratory tract infection including community-acquired pneumonia, hospital-acquired pneumonia,
ventilator-associated pneumonia, and exacerbation of COPD and bronchitis. We excluded paediatric trials and trials that did not
use procalcitonin for guiding initiation and duration of antibiotic treatment from the analysis.
The search strategy for this review was updated in Feb 10, 2017, in collaboration with Cochrane, and done in all databases
from the date of their inception to Feb 10, 2017. All references were screened for eligibility. The databases searched were the
Cochrane Central Register of Controlled Trials (CENTRAL; Feb 10, 2017, issue 1), MEDLINE (1966 to Feb 10, 2017), and
Embase (1980 to Feb 10, 2017). There were no language or publication restrictions.
Two reviewers (YW and RS) independently assessed trial eligibility based on titles, abstracts, full-text reports, and further
information from investigators as needed. Study protocols, case-report forms, and unedited databases containing individual
patient data were requested from investigators of all eligible trials. Data from each trial were first checked against reported results
and queries were resolved with the principal investigator, trial data manager, or statistician. Data were assessed in a consistent
manner across all trials with standard definitions and parameters
Articles
and thus mortality and adverse outcome rates differed
to-treat principle—analysing patients according to the slightly
from previous reports. In accordance with the
groups to which they were randomly assigned. We Cochrane
method, we used GRADE system42 to assess risk
excluded patients who withdrew consent and assumed no for
selection bias, performance bias, detection bias,
events for the few patients lost to follow-up before day 30 attrition
bias, reporting bias, and other bias.
after randomisation. Censoring was used for patients with a follow-up shorter than 30 days for time-to-event Data analysis
analyses. We included all patients with an acute respiratory
Prespecified sensitivity analyses were done for the infection
randomly assigned to a procalcitonin-guided
quality indicators allocation concealment, blinded care group or a
control group in the analysis. There were
outcome assessment, follow-up time, and protocol two
prespecified primary endpoints: all-cause mortality
adherence (<70% vs ≥70%). We evaluated heterogeneity within 30
days of randomisation and treatment failure
of disease severity across the patient population with within 30
days of randomisation. For trials with a shorter
prespecified analyses stratified by clinical setting and follow-up
period, we used the available information (eg,
diagnosis. We tested for subgroup effects by adding treatment
failure at the time of hospital discharge).
interaction terms to the model. Finally, heterogeneity Definitions
of treatment failure varied by and were
and inconsistency was further assessed in a meta- specific for each
clinical setting. For the primary care
analysis of aggregate data from all eligible trials using I2 setting,
we defined treatment failure as death, hospital
and Cochran’s Q test.45 All statistical analyses were done
admission, infection-specific complications (eg,
using Stata (version 9.2) and Review Manager empyema for lower
respiratory tract infection, or
(version 5.3). meningitis for upper respiratory tract infection),
recurrent or worsening infection and patients reporting
Role of the funding source any symptoms of an ongoing
respiratory infection (eg,
The funder of the study had no role in study design, data fever,
cough, or dyspnoea) at 30-day follow-up. Recurrent
collection, data analysis, data interpretation, or writing of or
worsening infection was defined as receiving another
 the report. The corresponding author had full access to course of
antibiotics in patients in whom antibiotics were
all the data in the study and had final responsibility for
discontinued, or increasing antibiotic dose or frequency
the decision to submit for publication. in patients already receiving
therapy for the same index infection. For patients initially evaluated in the emergency department or hospital, but not ICU setting,
we defined treatment failure as death, subsequent ICU admission, hospital re-admission after index hospital discharge, infection-
associated complications (eg, empyema or acute respiratory distress syndrome), and recurrent or worsening infection within 30
days of follow-up. In the ICU setting, we defined treatment failure as death within 30 days of follow-up and recurrent
71 articles assessed for eligibility
or worsening infection.
Secondary endpoints were antibiotic use defined as initiation of antibiotics, duration of antibiotics in days, and total exposure to
antibiotics (total number of antibiotic days divided by total number of patients). Exploratory analyses of other clinical outcomes
included length of hospital stay, ICU admission, length of ICU stay, antibiotic side-effects (appendix p 8), and number of days
with restricted activities of daily living within 14 days of randomisation.
For the coprimary endpoints (mortality and treatment failure), we calculated odds ratios (ORs) and 95% CIs
4 datasets not received using
multivariable hierarchical logistic regression.43,44 Variables in the multivariate analysis were treatment
China-Japan Friendship Hospital, Beijing, China (B Cao MD); Critical Care and Peri-operative Medicine, Monash
Health, Melbourne, VIC, Australia (Prof Y Shehabi PhD); School of Clinical Sciences, Faculty of Medicine Nursing and
Health Sciences, Monash University, Melbourne, VIC, Australia (Y Shehabi); Centre of Excellence for Health, Immunity
and Infections, Department of Infectious Diseases and Rheumatology, Finsencentret, Rigshospitalet, University of
Copenhagen, Copenhagen, Denmark (J-U S Jensen MD); Department of Respiratory Medicine, Copenhagen University
Hospital Bispebjerg, Copenhagen, Denmark (C Corti MD); Elisabeth Tweesteden Hospital, Tilburg, Netherlands (J A H
van Oers MD); Medisch Spectrum Twente, Enschede, Netherlands (A Beishuizen MD); VUmc University Medical Center,
Amsterdam, Netherlands (A R J Girbes MD, E de Jong MD); and Department of Health Research Methods, Evidence,
and Impact, 990 records identified through database searching
McMaster University, of Cochrane Central Register of Controlled Trials,
MEDLINE, and Embase
Hamilton, ON, Canada (M Briel)
Correspondence to: Prof Philipp Schuetz, University
919 records excluded based on review of
titles and abstracts
Department of Medicine, Kantonsspital Aarau, CH-5001 Aarau, Switzerland [email protected]
32 RCTs included in aggregate data analysis
(9909 participants)
2 datasets with no identifiable respiratory
infection patients
group, age, sex, and type of infection. To control for variability within and between trials, we added a trial variable to the model
as a random effect. Linear regression
39 articles excluded
1 did not use procalcitonin 2 reviews 2 paediatric studies 2 editorials 26 non-randomised trials 6 duplicate publications
See Online for appendix
26 RCTS included in final patient data analysis
 (6708 participants with acute respiratory infections) models were fitted for continuous endpoints and logistic regression models
were fitted for binary secondary
Figure 1: Study selection endpoints. Analyses were done following
the intention-
RCT=randomised controlled trial.
www.thelancet.com/infection Vol 18 January 2018 97
Articles
Country Setting, type of trial Clinical diagnosis Type of procalcitonin algorithm and procalcitonin cutoffs used
Bloos et al
Germany ICU, multicentre Severe sepsis or septic shock Discontinuation at day 4, 7, and 10; recommendation
against antibiotic: (2016)13
<1·0 μg/L or >50% drop to previous
value Bouadma et al
France ICU, multicentre Suspected bacterial infections during ICU stay (2010)14
without prior antibiotic (<24 h)
98 www.thelancet.com/infection Vol 18 January 2018
Initiation and duration; recommendation against antibiotic: <0·5 μg/L (<0·25 μg/L); recommendation for antibiotic: >0·5 μg/L
(>1·0 μg/L) Branche et al (2015)15
USA ED, medical ward, single centre Lower acute respiratory infection Initiation and duration; recommendation against
antibiotic: <0·25 μg/L
(<0·1 μg/L); recommendation for
antibiotic: >0·25 μg/L (>0·5 μg/L) Briel et al (2008)16 Switzerland Primary care, multicentre Upper and lower acute respiratory
infection Initiation and duration; recommendation against antibiotic: <0·25 μg/L
(<0·1 μg/L); recommendation for
antibiotic: >0·25 μg/L (>0·5 μg/L) Burkhardt et al (2010)17
Germany Primary care, multicentre Upper and lower acute respiratory infection Initiation; recommendation against antibiotic:
<0·25 μg/L;
recommendation for antibiotic: >0·25
μg/L Christ-Crain et al (2004)18
Switzerland ED, single centre Lower acute respiratory infection with x-ray
confirmation
Initiation; recommendation against antibiotic: <0·25 μg/L (<0·1 μg/L); recommendation for antibiotic: >0·25 μg/L (>0·5 μg/L)
Christ-Crain et al (2006)19
Switzerland ED, medical ward, single centre Community-acquired pneumonia with x-ray
confirmation
Initiation and duration; recommendation against antibiotic: <0·25 μg/L (<0·1 μg/L); recommendation for antibiotic: >0·25 μg/L
(>0·5 μg/L) Corti et al (2016)20
Denmark ED, single centre Acute exacerbation of COPD Initiation and duration; recommendation against antibiotic <0·25 μg/L
(0·15 μg/L)/80% decrease,
recommendation for antibiotic >0·25 μg/L de Jong et al (2016)9
Netherlands ICU, multicentre Critically ill patients with presumed infection Duration; recommendation against antibiotic: <0·5
μg/L or >80% drop
Deliberato et al (2013)21
Brazil ICU, single centre Septic patients with proven bacterial infection Duration; recommendation against antibiotic: <0·5 μg/L
or >90% drop
Hochreiter et al (2009)22
Germany Surgical ICU, single centre Suspected bacterial infections and >1 systemic
inflammatory response syndrome criteria
Duration; recommendation against antibiotic: <1 μg/L or >65% drop over 3 days Kristoffersen et al (2009)23
Denmark ED, medical ward, multicentre Lower acute respiratory infection without x-ray
confirmation
Initiation and duration; recommendation against antibiotic: <0·25 μg/L; recommendation for antibiotic: >0·25 μg/L (>0·5 μg/L)
Layios et al (2012)24
Belgium ICU, single centre Suspected infection Initiation; recommendation against antibiotic: <0·5 μg/L (<0·25 μg/L);
recommendation for antibiotic: >0·5 μg/L
(>1·0 μg/L) Long et al (2009)26
China ED, outpatients, single centre Community-acquired pneumonia with x-ray
confirmation
Initiation and duration; recommendation against antibiotic: <0·25 μg/L; recommendation for antibiotic: >0·25 μg/L Long et al
(2011)25
China ED, outpatients, single centre Community-acquired pneumonia with x-ray
confirmation
Initiation and duration; recommendation against antibiotic: <0·25 μg/L; recommendation for antibiotic: >0·25 μg/L Long et al
(2014)27
China ED, single centre Severe acute exacerbation of asthma Initiation; recommendation against antibiotic: <0·25 μg/L (<0·1
μg/L);
recommendation for antibiotic: >0·25
μg/L Maravić-Stojković et al (2011)28
Serbia ICU surgical, single centre Infection after open heart surgery Initiation; recommendation for antibiotic: >0·5 μg/L
Nobre et al (2008)29
Switzerland ICU, single centre Suspected severe sepsis or septic shock Duration; recommendation against antibiotic: <0·5 μg/L
(<0·25 μg/L) or
>90% drop; recommendation for
antibiotic: >0·5 μg/L (>1·0 μg/L) Oliveira et al (2013)30
Brazil ICU, two-centre Severe sepsis or septic shock Discontinuation; initial <1·0 μg/L: recommendation against antibiotic:
0·1 μg/L at day 4; initial >1·0 μg/L:
recommendation against: >90% drop Schroeder et al (2009)31
Germany Surgical ICU, single centre Severe sepsis after abdominal surgery Duration; recommendation against antibiotic: <1
μg/L or >65% drop over
3 days Schuetz et al (2009)32
Switzerland ED, medical ward, multicentre Lower acute respiratory infection with x-ray
confirmation
Initiation and duration; recommendation against antibiotic: <0·25 μg/L (<0·1 μg/L); recommendation for antibiotic: >0·25 μg/L
(>0·5 μg/L) Shehabi et al (2014)33
Australia ICU, multicentre Suspected sepsis, undifferentiated infections Duration; recommendation against antibiotic: <0·25 μg/L
(<0·1 μg/L) or
>90% drop Stolz et al (2007)34
Switzerland ED, medical ward, single centre Exacerbated COPD Initiation and duration; recommendation against antibiotic:
<0·25 μg/L
(<0·1 μg/L); recommendation for
antibiotic: >0·25 μg/L (>0·5 μg/L) Stolz et al (2009)35
Switzerland, USA
ICU, multicentre Ventilator-associated pneumonia when
intubated for >48 h
Duration; recommendation against antibiotic: <0·5 μg/L (<0·25 μg/L) or >80% drop; recommendation for antibiotic: >0·5 μg/L
(>1·0 μg/L) Verduri et al (2015)36
Italy ED, medical ward, multicentre Acute exacerbation of COPD Initiation; recommendation against antibiotic:<0·1 μg/L;
recommendation for antibiotic: >0·25
μg/L Wang et al (2016)37
China ICU, single centre Acute exacerbation of COPD All patients had initial procalcitonin <0·1 μg/L; antibiotic-group treated
with
antibiotic for at least 3 days, control group no antibiotic in the first 10 days
Recommendation relates to initiation or cessation of antibiotics. ICU=intensive care unit. ED=emergency department.
COPD=chronic obstructive pulmonary disease.
Table 1: Characteristics of included trials
Results We identified 990 records from the literature search, of which 71 articles were assessed for eligibility after exclusion of
919 records (figure 1). Data from
6708 individual patients were obtained and included in the meta-analysis of 26 eligible trials. We excluded two trials in which
patients did not have confirmed respiratory infections, and patient data were unavailable from four
Articles
additional trials. Trials were done in 12 countries: Australia, Belgium, Brazil, China, Denmark, France, Germany, Italy, the
Netherlands, Serbia, Switzerland,
Control
Procalcitonin (n=3372)
group (n=3336)
and the USA (table 1, appendix p 1). There were two primary care trials with patients with upper respiratory tract infections and
lower respiratory tract infection (n=1008), 11 trials from emergency departments and medical wards with patients with lower
respiratory
Age, years Sex Men Women Clinical setting
61·2 (18·4) 60·7 (18·8)
1910 (57%) 1898 (57%) 1462 (43%) 1438 (43%)
tract infection (n=3253), and 13 trials from ICUs with
Primary care 501 (15%) 507 (15%) patients who were septic
because of lower respiratory
Emergency department 1638 (49%) 1615 (48%) tract
infections (n=2447). Procalcitonin-based algorithms
ICU 1233 (37%) 1214 (36%) used in the different trials were
similar in concept and
Primary diagnosis recommended initiation or continuation of
antibiotic therapy based on procalcitonin cutoff levels. Adherence
Total upper acute respiratory
280 (8%) 292 (9%) infection
to algorithms was variable, ranging from 44% to 100% (appendix p 3). Quality of trials according to GRADE was moderate to
high (appendix p 6). Caregivers and patients were blinded to the intervention in most of the trials, but half of trials did not have a
blinded outcome assessment.
Common cold Rhino-sinusitis, otitis Pharyngitis, tonsillitis Total lower acute respiratory infection
156 (5%) 149 (4%) 67 (2%) 73 (2%) 46 (1%) 61 (2%) 3092 (92%) 3044 (91%)
There was no evidence of publication bias based on inspection of the funnel plot (appendix p 7).
Baseline characteristics of individual patients were similar in procalcitonin and control groups (table 2). Most patients were
recruited in the emergency department or the ICU. Community-acquired pneumonia was the most frequent diagnosis in more
than 40% of patients (table 2). There were 286 deaths within 30 days in 3336 procalcitonin-
Community-acquired pneumonia 1468 (44%) 1442 (43%) Hospital-acquired pneumonia 262 (8%) 243 (7%) Ventilator-
associated pneumonia 186 (6%) 194 (6%) Acute bronchitis 287 (9%) 257 (8%) Exacerbation of COPD 631 (19%) 621 (19%)
Exacerbation of asthma 127 (4%) 143 (4%) Other lower acute respiratory infection
guided patients (9%) compared with 336 deaths in 3372 controls (10%), resulting in a significantly lower mortality associated
with procalcitonin-guided therapy (adjusted OR 0·83 [95% CI 0·70–0·99], p=0·037; table 3). This effect was consistent across
clinical settings (no significant difference due to subgroup effect), although mortality could not be estimated in primary care trials
in which only one death was reported in a control patient. The effects on mortality were also consistent among different types of
infections (no significant difference for each interaction), excluding patients with bronchitis for whom mortality could not be
assessed (table 3).
Treatment failure in procalcitonin-guided patients was numerically lower than control patients, but not significantly different
(23·0% vs 24·9%; adjusted OR 0·90 [95% CI 0·80–1·01], p=0·068). These results were similar among subgroups by clinical
setting and type of respiratory infection (p
interactions
www.thelancet.com/infection Vol 18 January 2018 99
131 (4%) 144 (4%)
Procalcitonin dose on enrolment
Data available 2590 (77%) 3171 (95%) <0·1 μg/L 921 (36%) 981 (31%) 0·1–0·25 μg/L 521 (20%) 608 (19%) >0·25–0·5 μg/L
308 (12%) 383 (12%) >0·5–2·0 μg/L 358 (14%) 520 (16%) >2·0 μg/L 482 (19%) 679 (21%)
Data are mean (SD) or n (%). ICU=intensive care unit. COPD=chronic obstructive pulmonary disease.
Table 2: Baseline characteristics of included patients
failure associated with procalcitonin-guided treatment (0·90 [0·81–0·99]). Heterogeneity for both endpoints was low suggesting
similar effects among subgroups (I2=0% for both).
Procalcitonin guidance was associated with a reduction >0·05; table 3). Mortality
and treatment
in total antibiotic exposure (mean 5·7 days vs 8·1 days in failure
results were also not significantly different from
the control group, adjusted regression coefficient the main analysis
in the sensitivity analysis based on the
–2·43 days [95% CI –2·71 to 2·15], p<0·0001; table 4, main
quality indicators of trials with no evidence of effect
figure 3). Fewer patients in the procalcitonin group were
modification (appendix p 5).
prescribed antibiotics than in the control group and, in As an
additional sensitivity analysis, a meta-analysis of
patients for whom antibiotics were prescribed, duration the
aggregate results of all 32 eligible trials was done and
of therapy was shorter in procalcitonin-guided patients. included
the six trials initially excluded from the
The effect on antibiotic use differed by clinical setting. In
individual patient data analysis (figure 2). The point
the primary care setting, lower antibiotic exposure was estimate for
mortality was similar to the individual
mainly due to lower initial prescription rates in patient data
analysis, but was not significant (OR 0·89
procalcitonin-guided patients than control patients [95% CI 0·78–
1·01]). The aggregate analysis of treatment
(p
interaction failure showed a significant reduction in risk of
treatment
<0·0001). Similarly, lower antibiotic exposure due to lower prescription rates was found in selected
Articles
Control (n=3372) Procalcitonin group
Adjusted OR (95% CI)*, p value p
interaction (n=3336)
Overall 30-day mortality 336 (10%) 286 (9%) 0·83 (0·7 to 0·99), p=0·037 ·· Treatment failure 841 (25%) 768 (23%) 0·90 (0·80
to 1·01), p=0·068 ·· Length of ICU stay, days 13·3 (16·0) 13·7 (17·2) 0·39 (–0·81 to 1·58), p=0·524 ·· Length of hospital stay,
days 13·7 (20·6) 13·4 (18·4) –0·19 (–0·96 to 0·58), p=0·626 ·· Antibiotic-related side-effects 336/1521 (22%) 247/1513 (16%)
0·68 (0·57 to 0·82), p<0·0001 ·· Setting-specific outcomes Primary care 501 507 ·· ·· 30-day mortality 1 (<1%) 0 (0) ·· ··
Treatment failure 164 (33%) 159 (31%) 0·96 (0·73 to 1·25), p=0·751 0·715 Days with restricted activities 8·9 (4·2) 8·9 (4·1)
0·07 (–0·44 to 0·59), p=0·777 ·· Antibiotic-related side-effects 128/498 (26%) 102/506 (20%) 0·65 (0·46 to 0·91), p=0·012
0·596 Emergency department 1638 1615 ·· ··
30-day mortality 62 (4%) 57 (4%) 0·91 (0·63 to 1·33), p=0·635 0·546 Treatment failure 292 (18%) 259 (16%) 0·87 (0·72 to
1·05), p=0·141 0·807 Length of hospital stay, days 8·2 (10·5) 8·1 (7·5) –0·14 (–0·73 to 0·44), p=0·631 0·684 Antibiotic-related
side-effects 208/1023 (20%) 145/1007 (14%) 0·66 (0·52 to 0·83), p=0·001 0·596 Intensive care unit 1233 1214 ·· ··
30-day mortality 273 (22%) 229 (19%) 0·84 (0·69 to 1·02), p=0·081 0·619 Length of ICU stay, days 14·8 (16·2) 15·3 (17·5)
0·56 (–0·82 to 1·93), p=0·427 0·849 Length of hospital stay, days 26·3 (26·9) 25·8 (23·9) –0·33 (–2·28 to 1·62), p=0·739 0·641
Disease-specific outcomes Community-acquired pneumonia 1468 1442 ·· ··
30-day mortality 206 (14%) 175 (12%) 0·82 (0·66 to 1·03), p=0·083 0·958 Treatment failure 385 (26%) 317 (22%) 0·78 (0·66 to
0·93), p=0·005 0·052 Length of ICU stay, days 10·5 (10·3) 11·9 (13·3) 1·45 (0·15 to 2·75), p=0·029 0·119 Length of hospital
stay, days 13·3 (15·7) 13·9 (16·1) 0·74 (–0·25 to 1·73), p=0·143 0·094 Antibiotic-related side-effects 186/671 (28%) 127/666
(19%) 0·62 (0·48 to 0·8), p<0·0001 0·227 Exacerbation of COPD 631 621 ·· ··
30-day mortality 24 (4%) 19 (3%) 0·80 (0·43 to 1·48), p=0·472 0·847 Treatment failure 110 (17%) 104 (17%) 0·94 (0·7 to
1·27), p=0·704 0·676 Length of hospital stay, days 9·3 (13·9) 8·4 (7·2) –0·6 (–1·84 to 0·64), p=0·342 0·658 Antibiotic-related
side-effects 30/274 (11%) 29/275 (11%) 0·93 (0·53 to 1·63), p=0·805 0·198 Acute bronchitis 287 257 ·· ·· 30-day mortality 0 (0)
2 (1%) ·· ·· Treatment failure 55 (19%) 52 (20%) 1·11 (0·72 to 1·7), p=0·643 0·4 Length of hospital stay, days 2·6 (5·7) 2·2
(4·7) –0·21 (–0·9 to 0·48), p=0·556 0·97 Antibiotic-related side-effects 54/250 (22%) 39/226 (17%) 0·77 (0·49 to 1·22),
p=0·263 0·657 Ventilator-associated pneumonia 186 194 ·· ··
30-day mortality 29 (16%) 23 (12%) 0·75 (0·41 to 1·39), p=0·366 0·644 Treatment failure 51 (27%) 44 (23%) 0·78 (0·48 to
1·28), p=0·332 0·522 Length of ICU stay, days 23·5 (20·5) 21·8 (19·1) –1·74 (–5·64 to 2·17), p=0·383 0·441 Length of hospital
stay, days 33·8 (27·6) 32·0 (23·1) –2·14 (–7·04 to 2·75), p=0·391 0·448
Data are n, mean (SD), or n (%), unless otherwise specified. OR=odds ratio. ICU=intensive care unit. COPD=chronic obstructive
pulmonary disease. *Multivariable hierarchical regression with outcome of interest as dependent variable; age and respiratory
tract infection diagnosis as independent variables; and trial as a random effect.
Table 3: Clinical endpoints overall and stratified by setting and diagnosis
infections such as acute bronchitis. Lower antibiotic prescription rates and shorter duration of antibiotic therapy in patients
contributed to the lower overall exposure in the emergency department setting. In the ICU setting and in patients diagnosed with
community-
100 www.thelancet.com/infection Vol 18 January 2018
acquired pneumonia, the lower exposure was mainly explained by shorter treatment durations.
There was a significant reduction in antibiotic-related side-effects in procalcitonin-guided patients (16% vs 22%; adjusted OR
0·68 [95% CI 0·57–0·82], p<0·0001). This
Articles
A
Procalcitonin algorithm
Control algorithm Odds ratio (95% CI) Events
Events Total Primary care trials Briel (2008) Burkhardt (2010)
Subtotal (95% CI) Total events Heterogeneity: not applicable Test for overall effect: Z=0·69, p=0·49
Emergency department trials Christ-Crain (2004) Christ-Crain (2006) Stolz (2007) Kristoffersen (2009) Schuetz (2009) Long
(2009) Long (2011) Tang (2013) Long (2014) Ogasawara (2014) Verduri (2015) Branche (2015) Corti (2016) Lima (2016)
Subtotal (95% CI) Total events Heterogeneity: I2=0%, p=0·90 Test for overall effect: Z=0·16, p=0·88
Intensive care unit trials Nobre (2008) Stolz (2009) Schroeder (2009) Hochreiter (2009) Bouadma (2010) Maravić-Stojković
(2011) Layios (2012) Ananne (2013) Ding (2013) Deliberato (2013) Oliveira (2013) Shehabi (2014) Najafi (2015) de Jong
(2016) Bloos (2016) Wang (2016) Subtotal (95% CI) Total events Heterogeneity: I2=0%, p=0·89 Test for overall effect: Z=1·92,
p=0·06
Total (95% CI) Total events Heterogeneity: I2=0%, p=0·98 Test for overall effect: Z=1·86, p=0·06 Test for subgroup
differences: I2=0%, p=0·70
(Figure 2 continues on next page)
www.thelancet.com/infection Vol 18 January 2018 101
Weight Total
0
232
0·2%
0·32 (0·01–7·98) 0
275
Not estimable 507
0·2%
0·32 (0·01–7·98) 0
4
124
0·7%
1·29 (0·28–5·88) 18
151
3·4%
0·89 (0·45–1·75) 3
102
0·5%
1·58 (0·26–9·63) 2
103
0·3%
2·10 (0·19–23·51) 34
671
6·5%
1·06 (0·65–1·73) 0
63
Not estimable 0
77
Not estimable 1
132
0·3%
0·50 (0·04–5·58) 0
90
Not estimable 5
48
1·2%
0·44 (0·14–1·41) 3
88
0·5%
1·55 (0·25–9·53) 1
151
0·3%
0·49 (0·04–5·46) 1
62
0·3%
0·46 (0·04–5·20) 4
30
0·5%
2·23 (0·38–13·20) 1892
14·3%
0·97 (0·70–1·36) 76
8
39
1·3%
1·03 (0·34–3·09) 8
51
1·6%
0·59 (0·22–1·59) 3
14
0·5%
0·91 (0·15–5·58) 15
57
2·2%
0·99 (0·43–2·32) 65
307
10·4%
1·05 (0·71–1·55) 3
102
0·6%
1·01 (0·20–5·13) 56
258
8·7%
1·04 (0·68–1·58) 7
30
1·2%
0·55 (0·17–1·72) 21
33
1·6%
1·31 (0·49–3·48) 2
42
0·5%
0·44 (0·08–2·54) 16
49
2·1%
0·97 (0·41–2·29) 30
196
4·9%
1·20 (0·68–2·11) 5
30
0·8%
1·30 (0·31–5·40) 149
761
27·0%
0·73 (0·57–0·93) 140
552
21·6%
0·88 (0·68–1·16) 2
96
0·6%
 0·38 (0·07–2·02) 2617
85·6%
0·88 (0·77–1·00) 530
5016
100·0%
0·89 (0·78–1·01) 606
0·005 0·1 1 10 200
Favours procalcitonin algorithm Favours control algorithm 1 0 1
3 20 2 1 33 0 0 2 0 10 2 2 2 2
79
8 12 3 14 64 3 53 10 20 4 15 26 4 196 149 5
586
666
226 275 501
119 151 106 107 688 64 79 133 90 48 90 149 58 31 1913
40 50 13 53 314 103 251 28 35 39 45 198 30 785 537 95 2616
5030
outcome was only assessed in primary care and emergency department trials (six trials). There was no evidence of subgroup
effects (p
interactions
6708 patients is the first report to describe significant and relevant improvements in clinical outcomes and >0·05; table 3).
specifically a decreased risk for mortality for patients with Length of
hospital stay (adjusted regression coefficient
acute respiratory infections, when procalcitonin was used –0·19
days [95% CI –0·96 to 0·58], p=0·626) and ICU stay
to guide antibiotic treatment decisions. This effect was (0·39 days
[–0·81 to 1·58], p=0·524) were similar in the
consistent across clinical settings and types of infections,
procalcitonin and control groups and across setting-specific
and proved to be robust in different sensitivity analyses. and
disease-specific subgroups (p
interactions
>0·05; table 3).
For primary care and acute bronchitis patients, however, mortality was very low and the effect of procalcitonin Discussion
could not be reliably assessed. In line with previous To our
knowledge, this systematic review and individual
research, procalcitonin use was also associated with a patient data
meta-analysis of 26 randomised trials and
reduction in exposure to antibiotics mainly by reduced
Articles
B
Procalcitonin algorithm
Control algorithm
Weight Odds ratio (95% CI) Events
Total
Events Total Primary care trials Briel (2008)
73
232
68
6·0%
1
·07 (0·72–1·59) Burkhardt (2010)
86
275
96
7·5%
0
·85 (0·59–1·21) Subtotal (95% CI)
507
13·6%
0
·94 (0·72–1·22) Total events
159
164 Heterogeneity: I2=0%, p=0·40 Test for overall effect: Z=0·46,
p=0·64
Emergency department trials Christ-Crain (2004)
10
124
8
1·0%
1
·22 (0·46–3·20) Christ-Crain (2006)
36
151
56
3·8%
0
·53 (0·32–0·87) Stolz (2007)
13
102
15
1·5%
0
·89 (0·40–1·97) Schuetz (2009)
103
671
130
11·8%
0
·78 (0·59–1·03) Long(2009)
4
63
6
0·6%
0
·66 (0·18–2·44) Kristoffersen (2009)
8
103
6
0·8%
1
·42 (0·47–4·24) Long (2011)
8
77
7
0·8%
1
·19 (0·41–3·47) Tang (2013)
6
132
10
0·9%
0
·59 (0·21–1·66) Long (2014)
7
90
9
0·9%
0
·76 (0·27–2·13) Ogasawara (2014)
12
48
18
1·2%
0
·56 (0·23–1·33) Branche (2015)
3
151
5
0·5%
0
·58 (0·14–2·49) Verduri (2015)
19
88
12
1·5%
1
·79 (0·81–3·95) Corti (2016)
22
62
15
1·5%
1
·58 (0·72–3·46) Lima (2016)
6
30
4
0·5%
1
·69 (0·42–6·70) Subtotal (95% CI)
1892
27·4%
0
·85 (0·69–1·05) Total events
257
301 Heterogeneity: I2=8%, p=0·37 Test for overall effect: Z=1·53,
p=0·13
Intensive care unit trials Nobre (2008)
9
39
9
0·9%
1
·03 (0·36–2·96) Schroeder (2009)
3
14
3
0·3%
0
·91 (0·15–5·58) Stolz (2009)
8
51
12
1·0%
0
·59 (0·22–1·59) Hochreiter (2009)
15
57
14
1·3%
0
·99 (0·43–2·32) Bouadma (2010)
85
307
80
7·5%
1
·12 (0·78–1·60) Maravić-Stojković (2011)
40
102
41
3·0%
0
·98 (0·56–1·71) Layios (2012)
56
258
53
5·3%
1
·04 (0·68–1·58) Oliveira (2013)
19
49
16
1·4%
1
·15 (0·50–2·65) Ding (2013)
21
33
20
1·0%
1
·31 (0·49–3·48) Deliberato (2013)
4
42
5
0·5%
0
·72 (0·18–2·88) Ananne (2013)
7
30
10
0·7%
0
·55 (0·17–1·72) Shehabi (2014)
36
196
38
3·7%
0
·95 (0·57–1·57) Najafi (2015)
5
30
4
0·5%
1
·30 (0·31–5·40) Bloos (2016)
140
552
149
13·2%
0
·88 (0·68–1·16) de Jong (2016)
187
761
219
18·5%
0
·84 (0·67–1·06) Wang (2016)
2
96
5
0·3%
0
·38 (0·07–2·02) Subtotal (95% CI)
2617
59·0%
0
·92 (0·81–1·05) Total events
637
678 Heterogeneity: I2=0%, p=0·98 Test for overall effect: Z=1·28,
p=0·20
Total (95% CI)
5016
100·0%
0
·90 (0·81–0·99) Total events
1053
1143 Heterogeneity: I2=0%, p=0·89
0·1 0·2 0·5 1 2 5 10 Test for overall
effect: Z=2·16, p=0·03 Test for subgroup differences: I2=0%, p=0·78
Favours procalcitonin algorithm Favours control algorithm 102 www.thelancet.com/infection Vol 18 January 2018 226 275 501
119 151 106 688 64 107 79 133 90 48 149 90 58 31 1913
40 13 50 53 314 103 251 45 35 39 28 198 30 537 785 95 2616
5030
Figure 2: Forest plot showing overall mortality (A) and treatment failure (B) at 30 days from aggregate data meta-
analysis Odds ratios calculated with a random-effects Mantel-Haenszel test.
antibiotic prescription in low-risk settings and low-risk patients and shorter duration and earlier discontinuation of antibiotics in
high-risk patients. Procalcitonin use also resulted in reduced antibiotic side-effects, but did not have an effect on length of ICU or
hospital stay.
Acute respiratory infections are caused by bacteria, viruses, and other causes and are often treated with antibiotic
therapy.6,7,46 Although early initiation of antibiotic therapy reduces morbidity associated with bacterial infections, overuse of
antibiotics in patients with viral bronchitis and prolonged use in patients
with bacterial infection and sepsis has contributed to the development of multidrug-resistant bacterial pathogens.47,48 Reduction
of antibiotic use without increasing the risk for adverse patient outcomes is an international priority. In the past 10 years, the
infection blood biomarker procalcitonin has been proposed as an adjunct to clinical judgment and traditional clinical parameters
to guide antibiotic prescribing practices in patients with acute respiratory infections. Procalcitonin measurements increase within
6–12 h of infection in response to pro-inflammatory mediator release after
Articles
Control (n=3372) Procalcitonin group
Adjusted OR or difference (95% CI), (n=3336)
p value*
www.thelancet.com/infection Vol 18 January 2018 103
p
interaction
Overall Initiation of antibiotics 2894 (86%) 2351 (70%) 0·27 (0·24 to 0·32), p<0·0001 ·· Duration of antibiotics, days† 9·4 (6·2)
8·0 (6·5) –1·83 (–2·15 to –1·5), p<0·0001 ·· Total exposure of antibiotics, days‡ 8·1 (6·6) 5·7 (6·6) –2·43 (–2·71 to –2·15),
p<0·0001 ·· Setting-specific outcomes Primary care 501 507 ·· ··
Initiation of antibiotics 316 (63%) 116 (23%) 0·13 (0·09 to 0·18), p<0·0001 <0·0001 Duration of antibiotics, days† 7·3 (2·5) 7·0
(2·8) –0·52 (–1·07 to 0·04), p=0·068 0·064 Total exposure of antibiotics, days‡ 4·6 (4·1) 1·6 (3·2) –3·02 (–3·45 to –2·58),
p<0·0001 0·101 Emergency department 1638 1615 ·· ··
Initiation of antibiotics 1354 (83%) 1119 (69%) 0·49 (0·41 to 0·58), p<0·0001 <0·0001 Duration of antibiotics, days† 9·8 (5·4)
7·3 (5·1) –2·45 (–2·86 to –2·05), p<0·0001 <0·0001 Total exposure of antibiotics, days‡ 8·2 (6·2) 5·2 (5·4) –3·02 (–3·41 to –
2·62), p<0·0001 <0·0001 Intensive care unit 1233 1214 ·· ··
Initiation of antibiotics 1224 (99%) 1116 (92%) 0·02 (0·01 to 0·05), p<0·0001 <0·0001 Duration of antibiotics, days† 9·5 (7·4)
8·8 (7·8) –1·23 (–1·82 to –0·65), p<0·0001 <0·0001 Total exposure of antibiotics, days‡ 9·5 (7·4) 8·1 (7·9) –1·44 (–1·99 to –
0·88), p<0·0001 <0·0001 Disease-specific outcomes Community-acquired pneumonia 1468 1442 ·· ··
Initiation of antibiotics 1455 (99%) 1340 (93%) 0·08 (0·04 to 0·15), p<0·0001 <0·0001 Duration of antibiotics, days† 10·5 (6·2)
8·0 (5·7) –2·45 (–2·87 to –2·02), p<0·0001 <0·0001 Total exposure of antibiotics, days‡ 10·4 (6·2) 7·5 (5·9 –2·94 (–3·38 to –
2·5), p<0·0001 0·004 Exacerbation of COPD 631 621 ·· ··
Initiation of antibiotics 453 (72%) 266 (43%) 0·29 (0·23 to 0·36), p<0·0001 0·017 Duration of antibiotics, days† 7·4 (5·3) 7·2
(6·7) –1·15 (–2 to –0·31), p=0·007 0·003 Total exposure of antibiotics, days‡ 5·3 (5·6) 3·1 (5·6) –2·22 (–2·83 to –1·6),
p<0·0001 0·506 Acute bronchitis 287 257 ·· ··
Initiation of antibiotics 189 (66%) 68 (26%) 0·18 (0·12 to 0·26), p<0·0001 <0·0001 Duration of antibiotics, days† 7·1 (3·0) 6·4
(3·5) –0·35 (–1·15 to 0·45), p=0·393 0·359 Total exposure of antibiotics, days‡ 4·7 (4·2) 1·7 (3·3) –2·95 (–3·59 to –2·31),
p<0·0001 0·33 Ventilator-associated pneumonia 186 194 ·· ·· Initiation of antibiotics 186 (100%) 193 (100%) ·· ·· Duration of
antibiotics, days† 13·1 (7·9) 10·8 (8·7) –2·22 (–3·8 to –0·65), p=0·006 0·253 Total exposure of antibiotics, days‡ 13·1 (7·9)
10·8 (8·7) –2·45 (–4·09 to –0·82), p=0·003 0·786
Data are n, mean (SD), or n (%), unless otherwise specified. OR=odds ratio. COPD=chronic obstructive pulmonary disease.
*Multivariable hierarchical model adjusted for age and diagnosis and trial as a random effect. †Total days of antibiotic therapy in
patients in whom antibiotics were initiated. ‡Total days of antibiotic therapy in all randomly assigned patients.
Table 4: Antibiotic treatment overall and stratified by setting and diagnosis
bacterial invasion, are highest in patients who have bacteraemia, and correlate with disease severity and clinical outcome of
patients with infection.49,50 Unlike other inflammatory markers, procalcitonin release is blocked by cytokines, which
characterise the typical immune response to viral infections (interferon γ).51 Procalcitonin is therefore more specific for bacterial
infections than C-reactive protein or white cell count.52–54 Procalcitonin concentrations rapidly fall by about 50% each day
during resolution of infection and are therefore useful in monitoring the clinical course and supporting decisions to discontinue
antibiotic treatment. However, an important impediment to the evaluation and validation of any sepsis marker has been the
absence
of a reliable reference standard for bacterial infection, particularly for respiratory infections. For procalcitonin, sensitivities and
specificities of around 80% have been reported in previous observational studies using blood culture as the reference
standard.55,56 To increase sensitivity and specificity of procalcitonin, existing algorithms use a variety of cutoff points in
conjunction with clinical criteria to guide antibiotic prescription.10 Although observational research does not permit
measurements of the true diagnostic accuracy of procalcitonin, interventional research is helpful to understand the clinical effect
of such algorithms. Several studies have now compared antibiotic use and clinical outcomes of acute respiratory infections in
patients
Articles
A 100
86%
B Control group
10 Procalcitonin group
9
8·14
kinetics over 72 h to be a strong and independent predictor of mortality. Early identification of non-responders to antibiotic and
other medical treatment might also help to prevent adverse events. Secondly, increased risk for 82% 80
71%
71% ) % ( s c
63%
8
treatment failure in control patients might be related to
5·71
prolonged antibiotic exposure and risk for secondary complications and re-admission to hospital.58,59 In our analysis,
procalcitonin-guided care also correlated with
51%
lower risk for antibiotic side-effects, which can be associated with both treatment failure and mortality. Third, in sick patients
with evidence of an acute respiratory tract illness, a lower-than-expected procalcitonin concen- P
tration might direct clinicians to look for alternative explanations of these symptoms (eg, pulmonary embolism or heart failure).
Finally, as detailed in current sepsis guidelines, several observational studies have now
0
Overall
reported lower mortality and treatment failure risk associated with early antibiotic de-escalation in patients
Figure 3: Antibiotic use (A) Proportions of patients on antibiotics. (B) Mean duration of antibiotic use.
)syad(sc
7
itoi
60
58%
itoib
6bitna
itnan
5nostneit
40
36%
41%
oemitn
4
a
23%
27%
aeM
3
20
16%
18%
11%
13%
8%
2
1
0
Day 0
Day 2 Day 4 Day 6 Day 8 Day 10 Day 12 Day 14
with sepsis than in patients with no de-escalation.60,61
Although most research showing the benefit of procalcitonin treatment algorithms has focused on patients in ICUs and
emergency departments, the value managed with and without procalcitonin protocols. In
of procalcitonin in primary care is still incompletely these trials,
procalcitonin algorithms were paired with
understood. Observational outpatient data have been
recommendations for or against antibiotic initiation and
largely inconclusive on the added value of procalcitonin
continuation based on clinical stability and at specific
to diagnose infection compared with other markers and
procalcitonin levels or procalcitonin kinetics.10 Although
clinical parameters. In our analysis, which included in most trials
such a strategy proved to be effective in
two non-inferiority primary care trials with 1008 patients, reducing
antibiotic use, the safety of this approach has
procalcitonin had a strong effect on antibiotic use in remained an
ongoing concern.
primary care patients. Moreover, resolution of illness as Our analysis
of the use of procalcitonin-guided care in
measured by days with restricted activities of daily living a large
aggregate patient population from different trials
was similar between procalcitonin and control patients and
countries did not conclusively reveal any associated
suggesting that patients not treated with antibiotics in harm, and
importantly showed significantly reduced
 the procalcitonin group did not need these drugs. mortality with
the use of procalcitonin treatment
For the primary care setting, it could be argued that any
algorithms. These results are consistent with the largest
intervention using other biomarkers or clinical parameters trial of
patients in ICUs, which also reported reduced
could reduce antibiotic use in a low-risk setting with high mortality
associated with procalcitonin-guided care.9 The
rates of overprescription. Studies comparing procalcitonin relative
mortality reduction was 14% (ie, from 10·0% to
and C-reactive protein have reported low correlation of 8·6%), and
was highest in ICU trials (15%) and in
these markers suggesting that important differences exist, patients
with community-acquired pneumonia (13%)
which could lead to different recommendations on and ventilator-
associated pneumonia (23%). For patients
antibiotic use in individual patients.62,63 However, head-to- in
emergency departments, the relative mortality
head studies assessing the clinical effect of procalcitonin reduction
was still 8%, whereas no effect could be
compared with C-reactive protein guided treatment estimated for
primary care because of the low number of
algorithms are still needed. Finally, one strategy that has events.
These results were also confirmed in an aggregate
had an impact in the inpatient setting is early provider data meta-
analysis including 32 eligible trials.
notification of procalcitonin results and this strategy Procalcitonin
thus seems to have the most clinical benefit
might improve the performance of procalcitonin-guided in high-
risk patient populations and no demonstrable
treatment algorithms if used in future studies done in safety
concerns in low-risk groups.
primary care settings. There are several possible explanations for the
positive
The strengths of this meta-analysis include a predefined effects
of procalcitonin-guided antibiotic treatment on
study protocol, a comprehensive search and retrieval of all
mortality in patients with acute respiratory infections.
relevant trials, and a network that permitted inclusion of First,
procalcitonin provides additional prognostic
individual patient data from most eligible trials. We also
information in the assessment of patients, which
standardised outcome definitions across trials and did influences
decisions about site-of-care and timing of
appropriate subgroup and sensitivity analyses, thereby
discharge.29 A large US study57 found procalcitonin
overcoming the limitation of previous meta-analyses with
104 www.thelancet.com/infection Vol 18 January 2018
Articles
aggregated data to make definitive conclusions. However,
Pharmaceuticals. YW, RS, LB, KBK, TW, VN, LW, DA, KR,
ABr, PD, MN, our study still had limitations. First, adherence to the procalcitonin algorithm was varied among the studies
ROD, CFO, VM-S, AV, BB, BC, JAHvO, ABe, ARJG, and EdJ declare no competing interests.
ranging from 44% to 100%. However, a sensitivity analysis found similar effects in trials with high and low adherence. Second,
we limited our analysis to immunocompetent adults with acute respiratory infections, thereby reducing
Acknowledgments The National Institute for Health Research (NIHR) provided a research grant for this review update. The 2012
review was supported by unrestricted research grants from Thermo-Fisher Scientific, the Gottfried and Julia Bangerter-Rhyner-
Foundation, the Swiss Foundation generalisability of our conclusions to other patient populations. Third, our patient population
was substantially heterogeneous with regard to clinical setting and type of respiratory infection. This heterogeneity also limits
generalisability of results particularly for the main
for Grants in Biology and Medicine (SSMBS, PASMP3–127684/1), and Santésuisse. We thank all researchers and patients
involved in the individual trials for sharing their data.
References 1 US Food and Drug Administration. FDA press release. FDA clears test to help manage antibiotic treatment for
lower respiratory tract endpoint, mortality, and specifically whether or not procalcitonin guidance results in a reduction in
mortality in the primary care setting remains unproven. Finally, we
infections and sepsis. Feb 23, 2017. https://www.fda.gov/ NewsEvents/Newsroom/PressAnnouncements/ucm543160.htm
(accessed Oct 2, 2017). 2 Schuetz P, Albrich W, Mueller B. Procalcitonin for diagnosis of did not do a cost-effectiveness analysis
because cost data were not available in most trials. Future studies should investigate cost-effectiveness of procalcitonin-guided
care.56
infection and guide to antibiotic decisions: past, present and future. BMC Med 2011; 9: 107. 3 Schuetz P, Christ-Crain M, Muller
B. Procalcitonin and other
biomarkers to improve assessment and antibiotic stewardship in
In conclusion, procalcitonin-guided antibiotic
infections—hope for hype? Swiss Med Wkly 2009; 139: 318–
26. treatment in patients with acute respiratory infections effectively reduced antibiotic exposure and antibiotic
4 Charles PE, Tinel C, Barbar S, et al. Procalcitonin kinetics within
the first days of sepsis: relationship with the appropriateness of antibiotic therapy and the outcome. Crit Care 2009; 13: R38. side-
effects while improving mortality. These findings
5 Sager R, Kutz A, Mueller B, Schuetz P. Procalcitonin-guided
were conserved across all clinical settings and clinical presentations of acute respiratory infections although mortality could not
be estimated in patients in primary
diagnosis and antibiotic stewardship revisited. BMC Med 2017; 15: 15. 6 Musher DM, Thorner AR. Community-acquired
pneumonia.
N Engl J Med 2014; 371: 1619–28. 7 Silverman M, Povitz M, Sontrop JM, et al. Antibiotic prescribing for care and in those
with bronchitis. When embedded in clinical algorithms, the use of procalcitonin has the potential to inform and improve care of
patients with acute respiratory infections by reducing antibiotic
nonbacterial acute upper respiratory infections in elderly persons. Ann Intern Med 2017; 166: 765–74. 8 Schuetz P, Albrich W,
Christ-Crain M, Chastre J, Mueller B.
Procalcitonin for guidance of antibiotic therapy. Expert Rev Anti Infect Ther 2010; 8: 575–87. exposure and the associated risk
of developing subsequent antibiotic resistance and more importantly improving clinical outcomes. These findings have
9 de Jong E, van Oers JA, Beishuizen A, et al. Efficacy and safety of procalcitonin guidance in reducing the duration of antibiotic
treatment in critically ill patients: a randomised, controlled, open-label trial. Lancet Infect Dis 2016; 16: 819–27. substantial
clinical and public health implications.
10 Schuetz P, Chiappa V, Briel M, Greenwald JL. Procalcitonin
algorithms for antibiotic therapy decisions: a systematic review
of Contributors
randomized controlled trials and recommendations for clinical
PS, MB, HCB, and BM conceived of the study and wrote the initial
algorithms. Arch Intern Med 2011; 171: 1322–31. protocol.
PS, YW, and RS did the literature search and all analysis for
11 Wu MH, Lin CC, Huang SL, et al. Can procalcitonin tests aid
in this report and wrote the first draft. All authors shared trial data, gave
identifying bacterial infections associated with influenza
crucial feedback on the manuscript, and approved the final version. PS,
pneumonia? A systematic review and meta-analysis. MB, and
BM oversaw the study.
Influenza Other Respir Viruses 2012; 7: 349–55.
Declaration of interests PS, MC-C, and BM have received support from Thermo-Fisher and bioMérieux to attend meetings and
fulfilled speaking engagements. BM has served as a consultant for and received research support from Thermo-Fisher. HCB and
MB have received research support from
12 Berg P, Lindhardt BO. The role of procalcitonin in adult patients with community-acquired pneumonia—a systematic review.
Dan Med J 2012; 59: A4357. 13 Bloos F, Trips E, Nierhaus A, et al. Effect of sodium selenite
administration and procalcitonin-guided therapy on mortality in patients with severe sepsis or septic shock: a randomized clinical
Thermo-Fisher for a previous meta-analysis regarding procalcitonin.
trial. JAMA Intern Med 2016; 176: 1266–76. DWdL’s hospital
received financial support for the randomisation tool by
14 Bouadma L, Luyt CE, Tubach F, et al. Use of procalcitonin to
reduce ThermoFisher. DS, OB, and MT have received research support from
patients’ exposure to antibiotics in intensive care units
Thermo-Fisher. TW and SS have received lecture fees and research
(PRORATA trial): a multicentre randomised controlled trial.
support from Thermo-Fisher. CEL has received lecture fees from Brahms
Lancet 2010; 375: 463–74. and Merck Sharp & Dohme-
Chibret. JC has received consulting and
15 Branche AR, Walsh EE, Vargas R, et al. Serum procalcitonin
lecture fees from Pfizer, Brahms, Wyeth, Johnson & Johnson,
measurement and viral testing to guide antibiotic use for
Nektar-Bayer, and Arpida. MW has received consulting and lectures fees from Merck Sharp & Dohme-Chibret, Janssen Cilag,
Gilead, Astellas,
respiratory infections in hospitalized adults: a randomized controlled trial. J Infect Dis 2015; 212: 1692–700. Sanofi, and
Thermo-Fisher. FT’s institution received funds from Brahms. CC has received an unrestricted grant of €2000 from Thermo-
Fisher Scientific, and non-financial support from bioMérieux for the ProToCOLD study. YS has received unrestricted research
grants from Thermo-Fisher, bioMérieux, Orion Pharma, and Pfizer. ARF has served on advisory boards for Novavax, Hologic,
Gilead, and MedImmune; and has received research funding from AstraZeneca, Sanofi Pasteur, GlaxoSmithKline, and ADMA
Biologics. J-USJ declares that he was invited to the European Respiratory Society meeting 2016 by Roche
16 Briel M, Schuetz P, Mueller B, et al. Procalcitonin-guided antibiotic
use vs a standard approach for acute respiratory tract infections in primary care. Arch Intern Med 2008; 168: 2000–07. 17
Burkhardt O, Ewig S, Haagen U, et al. Procalcitonin guidance and reduction of antibiotic use in acute respiratory tract infection.
Eur Respir J 2010; 36: 601–07. 18 Christ-Crain M, Jaccard-Stolz D, Bingisser R, et al. Effect of
procalcitonin-guided treatment on antibiotic use and outcome in lower respiratory tract infections: cluster-randomised, single-
blinded intervention trial. Lancet 2004; 363: 600–07.
www.thelancet.com/infection Vol 18 January 2018 105
Articles
19 Christ-Crain M, Stolz D, Bingisser R, et al. Procalcitonin
guidance of antibiotic therapy in community-acquired pneumonia: a randomized trial. Am J Respir Crit Care Med 2006; 174: 84–
93. 20 Corti C, Fally M, Fabricius-Bjerre A, et al. Point-of-care procalcitonin test to reduce antibiotic exposure in patients
hospitalized with acute exacerbation of COPD. Int J Chron Obstruct Pulmon Dis 2016; 11: 1381–89. 21 Deliberato RO, Marra
AR, Sanches PR, et al. Clinical and economic
impact of procalcitonin to shorten antimicrobial therapy in septic patients with proven bacterial infection in an intensive care
setting. Diagn Microbiol Infect Dis 2013; 76: 266–71. 22 Hochreiter M, Kohler T, Schweiger AM, et al. Procalcitonin to guide
duration of antibiotic therapy in intensive care patients: a randomized prospective controlled trial. Crit Care 2009; 13: R83. 23
Kristoffersen KB, Sogaard OS, Wejse C, et al. Antibiotic
treatment interruption of suspected lower respiratory tract infections based on a single procalcitonin measurement at hospital
admission—a randomized trial. Clin Microbiol Infect 2009; 15: 481–7. 24 Layios N, Lambermont B, Canivet JL, et al.
Procalcitonin usefulness
for the initiation of antibiotic treatment in intensive care unit patients. Crit Care Med 2012; 40: 2304–9. 25 Long W, Deng X,
Zhang Y, Lu G, Xie J, Tang J.
Procalcitonin guidance for reduction of antibiotic use in low-risk outpatients with community-acquired pneumonia. Respirology
2011; 16: 819–24. 26 Long W, Deng XQ, Tang JG, et al. The value of serum procalcitonin in treatment of community acquired
pneumonia in outpatient. Chin J Intern Med 2009; 48: 216–19 (in Chinese). 27 Long W, Li LJ, Huang GZ, et al. Procalcitonin
guidance for
reduction of antibiotic use in patients hospitalized with severe acute exacerbations of asthma: a randomized controlled study with
12-month follow-up. Crit Care 2014; 18: 471. 28 Maravić-Stojković V, Lausević-Vuk L, Jović M, Ranković A,
Borzanović M, Marinković J. Procalcitonin-based therapeutic strategy to reduce antibiotic use in patients after cardiac surgery: a
randomized controlled trial. Srp Arh Celok Lek 2011; 139: 736–42. 29 Nobre V, Harbarth S, Graf JD, Rohner P, Pugin J. Use of
procalcitonin to shorten antibiotic treatment duration in septic patients: a randomized trial. Am J Respir Crit Care Med 2008; 177:
498–505. 30 Oliveira CF, Botoni FA, Oliveira CR, et al. Procalcitonin versus C-reactive protein for guiding antibiotic therapy in
sepsis: a randomized trial. Crit Care Med 2013; 41: 2336–43. 31 Schroeder S, Hochreiter M, Koehler T, et al. Procalcitonin
(PCT)-guided algorithm reduces length of antibiotic treatment in surgical intensive care patients with severe sepsis: results of a
prospective randomized study. Langenbecks Arch Surg 2009; 394: 221–26. 32 Schuetz P, Christ-Crain M, Thomann R, et al.
Effect of
procalcitonin-based guidelines vs standard guidelines on antibiotic use in lower respiratory tract infections: the ProHOSP
randomized controlled trial. JAMA 2009; 302: 1059–66. 33 Shehabi Y, Sterba M, Garrett PM, et al. Procalcitonin algorithm in
critically ill adults with undifferentiated infection or suspected sepsis. A randomized controlled trial. Am J Respir Crit Care Med
2014; 190: 1102–10. 34 Stolz D, Christ-Crain M, Bingisser R, et al. Antibiotic treatment of exacerbations of COPD: a
randomized, controlled trial comparing procalcitonin-guidance with standard therapy. Chest 2007; 131: 9–19. 35 Stolz D,
Smyrnios N, Eggimann P, et al. Procalcitonin for reduced
antibiotic exposure in ventilator-associated pneumonia: a randomised study. Eur Respir J 2009; 34: 1364–75. 36 Verduri A,
Luppi F, D’Amico R, et al. Antibiotic treatment of severe
exacerbations of chronic obstructive pulmonary disease with procalcitonin: a randomized noninferiority trial. PLoS One 2015;
10: e0118241. 37 Wang JX, Zhang SM, Li XH, Zhang Y, Xu ZY, Cao B. Acute
exacerbations of chronic obstructive pulmonary disease with low serum procalcitonin values do not benefit from antibiotic
treatment: a prospective randomized controlled trial. Int J Infect Dis 2016; 48: 40–45.
106 www.thelancet.com/infection Vol 18 January 2018
38 Schuetz P, Briel M, Christ-Crain M, et al. Procalcitonin to initiate or
withhold antibiotics in acute respiratory tract infections (protocol). Cochrane Database Syst Rev 2008; 4: CD007498. 39 Schuetz
P, Muller B, Christ-Crain M, et al. Procalcitonin to initiate or discontinue antibiotics in acute respiratory tract infections.
Cochrane Database Syst Rev 2012; 9: CD007498. 40 Stewart LA, Clarke M, Rovers M, et al. Preferred Reporting Items
for Systematic Review and Meta-Analyses of individual participant data: the PRISMA-IPD Statement. JAMA 2015; 313: 1657–
65. 41 Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of
studies that evaluate health care interventions: explanation and elaboration. J Clin Epidemiol 2009; 62: e1–34. 42 Guyatt GH,
Oxman AD, Vist GE, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations.
BMJ 2008; 336: 924–26. 43 Thompson SG, Turner RM, Warn DE. Multilevel models for
meta-analysis, and their application to absolute risk differences. Stat Methods Med Res 2001; 10: 375–92. 44 Turner RM, Omar
RZ, Yang M, Goldstein H, Thompson SG.
A multilevel model framework for meta-analysis of clinical trials with binary outcomes. Stat Med 2000; 19: 3417–32. 45 Higgins
JP, Thompson SG, Deeks JJ, Altman DG. Measuring
inconsistency in meta-analyses. BMJ 2003; 327: 557–60. 46 Mizgerd JP. Acute lower respiratory tract infection.
N Engl J Med 2008; 358: 716–27. 47 Gonzales R, Steiner JF, Sande MA. Antibiotic prescribing for adults with colds, upper
respiratory tract infections, and bronchitis by ambulatory care physicians. JAMA 1997; 278: 901–04. 48 Lawrence KL, Kollef
MH. Antimicrobial stewardship in the
intensive care unit: advances and obstacles. Am J Respir Crit Care Med 2009; 179: 434–38. 49 Kutz A, Briel M, Christ-Crain M,
et al. Prognostic value of
procalcitonin in respiratory tract infections across clinical settings. Crit Care 2015; 19: 74. 50 Schuetz P, Hausfater P, Amin D, et
al. Biomarkers from distinct
biological pathways improve early risk stratification in medical emergency patients: the multinational, prospective, observational
TRIAGE study. Crit Care 2015; 19: 377. 51 Linscheid P, Seboek D, Schaer DJ, Zulewski H, Keller U, Muller B.
Expression and secretion of procalcitonin and calcitonin gene-related peptide by adherent monocytes and by macrophage-
activated adipocytes. Crit Care Med 2004; 32: 1715–21. 52 Christ-Crain M, Muller B. Procalcitonin in bacterial infections—
hype, hope, more or less? Swiss Med Wkly 2005; 135: 451–60. 53 Christ-Crain M, Muller B. Biomarkers in respiratory tract
infections: diagnostic guides to antibiotic prescription, prognostic markers and mediators. Eur Respir J 2007; 30: 556–73. 54
Linscheid P, Seboek D, Zulewski H, Keller U, Muller B.
Autocrine/paracrine role of inflammation-mediated calcitonin gene-related peptide and adrenomedullin expression in human
adipose tissue. Endocrinology 2005; 146: 2699–708. 55 Wacker C, Prkno A, Brunkhorst FM, Schlattmann P.
Procalcitonin as a diagnostic marker for sepsis: a systematic review and meta-analysis. Lancet Infect Dis 2013; 13: 426–35. 56
Muller F, Christ-Crain M, Bregenzer T, et al. Procalcitonin levels
predict bacteremia in patients with community-acquired pneumonia: a prospective cohort trial. Chest 2010; 138: 121–29. 57
Schuetz P, Birkhahn R, Sherwin R, et al. Serial procalcitonin
predicts mortality in severe sepsis patients: results from the Multicenter Procalcitonin MOnitoring SEpsis (MOSES) Study. Crit
Care Med 2017; 45: 781–89. 58 Roberts RR, Hota B, Ahmad I, et al. Hospital and societal costs of
antimicrobial-resistant infections in a Chicago teaching hospital: implications for antibiotic stewardship. Clin Infect Dis 2009; 49:
1175–84. 59 Classen DC, Jaser L, Budnitz DS. Adverse drug events among
hospitalized Medicare patients: epidemiology and national estimates from a new approach to surveillance. Jt Comm J Qual
Patient Saf 2010; 36: 12–21. 60 Garnacho-Montero J, Gutierrez-Pizarraya A, Escoresca-Ortega A,
et al. De-escalation of empirical therapy is associated with lower mortality in patients with severe sepsis and septic shock.
Intensive Care Med 2014; 40: 32–40.
Articles
61 Schuetz P, Mueller B. Biomarker-guided de-escalation of empirical
therapy is associated with lower risk for adverse outcomes. Intensive Care Med 2014; 40: 141. 62 Meili M, Kutz A, Briel M, et
al. Infection biomarkers in primary
care patients with acute respiratory tract infections-comparison of procalcitonin and C-reactive protein. BMC Pulm Med 2016;
16: 43.
www.thelancet.com/infection Vol 18 January 2018 107
63 Meili M, Muller B, Kulkarni P, Schutz P. Management of patients
with respiratory infections in primary care: procalcitonin, C-reactive protein or both? Expert Rev Respir Med 2015; 9: 587–601.