Huang et al. Ann.

Intensive Care (2017) 7:114

DOI 10.1186/s13613-017-0338-6

RESEARCH Open Access

Procalcitonin‑guided antibiotic therapy

in intensive care unit patients: a systematic
review and meta‑analysis
Hui‑Bin Huang1,2, Jin‑Min Peng1, Li Weng1, Chun‑Yao Wang1, Wei Jiang1 and Bin Du1*

Abstract 
Background:  Serum procalcitonin (PCT) concentration is used to guide antibiotic decisions in choice, timing, and
duration of anti-infection therapy to avoid antibiotic overuse. Thus, we performed a systematic review and meta-anal‑
ysis to seek evidence of different PCT-guided antimicrobial strategies for critically ill patients in terms of predefined
clinical outcomes.
Methods:  We searched for relevant studies in PubMed, Embase, Web of Knowledge, and the Cochrane Library up to
25 February 2017. Randomized controlled trials (RCTs) were included if they reported data on any of the predefined
outcomes in adult ICU patients managed with a PCT-guided algorithm or according to standard care. Results were
expressed as risk ratio (RR) or mean difference (MD) with accompanying 95% confidence interval (CI).
Data synthesis:  We included 13 trials enrolling 5136 patients. These studies used PCT in three clinical strategies: ini‑
tiation, discontinuation, or combination of antibiotic initiation and discontinuation strategies. Pooled analysis showed
a PCT-guided antibiotic discontinuation strategy had fewer total days with antibiotics (MD − 1.66 days; 95% CI − 2.36
to − 0.96 days), longer antibiotic-free days (MD 2.26 days; 95% CI 1.40–3.12 days), and lower short-term mortality (RR
0.87; 95% CI 0.76–0.98), without adversely affecting other outcomes. Only few studies reported data on other PCT-
guided strategies for antibiotic therapies, and the pooled results showed no benefit in the predefined outcomes.
Conclusions:  Our meta-analysis produced evidence that among all the PCT-based strategies, only using PCT for
antibiotic discontinuation can reduce both antibiotic exposure and short-term mortality in a critical care setting.
Keywords:  Procalcitonin, Antibiotic strategies, Meta-analysis, Systematic review, Intensive care unit

Background prescribed in hospital settings are either unnecessary or

Timely diagnosis and appropriate antimicrobial treat- inappropriate [3]. Nowadays, long-term antimicrobial
ment of infection remain a major challenge in critical regimens applied to critically ill patients are common
care settings. Delay in diagnosis due to lack of specific and often based on empiric rules [4, 5]. This may result
clinical signs in the early stage of infection may with- in increased medical costs, emergence of resistant patho-
hold or delay antibiotic therapy. On the other hand, con- gens, prolonged length of stay (LOS), and risk of mortal-
cern of not treating potentially life-threatening infection ity [6, 7].
and the risk of recurrence frequently leads clinicians to Recently, procalcitonin (PCT) has shown to be a prom-
antimicrobial overuse in intensive care unit (ICU) [1, 2]. ising biomarker for identification of bacterial infections
Studies have demonstrated that up to 50% of antibiotics and is correlated with the severity of infection [8–11].
The 2016 Surviving Sepsis Campaign (SSC) guidelines
offered a weak recommendation (low quality of evi-
*Correspondence: [email protected]
1
Medical ICU, Peking Union Medical College Hospital, Peking Union
dence), favouring that measurement of procalcitonin
Medical College and Chinese Academy of Medical Sciences, 1 Shuai Fu levels can be used to support shortening the duration of
Yuan, Beijing 100730, People’s Republic of China antimicrobial therapy in sepsis patients [12]. However,
Full list of author information is available at the end of the article

© The Author(s) 2017. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License
(http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium,
provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license,
and indicate if changes were made.
Huang et al. Ann. Intensive Care (2017) 7:114 Page 2 of 10

one recent large study failed to show any benefit of daily without mentioning of PCT assay methods. Articles
PCT measurement with regard to time to appropriate available only in abstract form or meeting reports were
therapy or survival, but resulted in a longer antibiotic also excluded.
course and ICU stay [13]. To date, several meta-analyses
have assessed the value of PCT to guide antibiotic stew- Data extraction and outcomes
ardship in ICU patients [14–22]. Findings of these reports Two reviewers independently extracted data from
showed that utilizing PCT to guide antibiotic decisions included studies on the first author, year of publication,
could significantly reduce antibiotics use, but did not country, sample size, study design, ICU type, compared
improve patient outcomes, such as mortality, hospital, or protocols, methods of PCT assay, methodological quality,
ICU LOS. However, one major limitation of these meta- as well as all outcomes of interest.
analyses was unexplainable significant heterogeneity We stratified different PCT-guided strategies according
among included trials, possibly due to the fact that the to medical decision with regards to antimicrobial therapy.
different PCT guidance strategies (including antibiotic In brief, strategy of antibiotic initiation referred to the
initiation, discontinuation, or combination of antibiotic decision to or not to start antibiotics, and decision of the
initiation and discontinuation strategies) had been evalu- intensified monitoring, diagnostic efforts, and interven-
ated in these trials. Since the considerable differences in tions to explore uncontrolled sources of infection based
methodologies and research purposes associated with on a predefined threshold of baseline PCT concentration,
the different PCT-guided strategies, the previous meta- while strategy of antibiotic discontinuation meant mak-
analyses might not accurately evaluate the effects of PCT- ing the decision to stop antibiotics according to a prede-
based algorithms (Additional file  1: Table S1). Recently, fined threshold of PCT concentration, or the PCT level
two large-scale randomized controlled trials (RCTs) of dropped by a certain proportion predefined compared
PCT-guided antibiotic strategy in ICU patients have been with the previous value. The primary outcomes were the
published, with inconsistent results [23, 24]. Of note, the duration of antibiotic use and the short-term mortality,
study by de Jong et al., the largest PCT trial to date, dem- while the latter was defined as ICU or hospital or 28-day
onstrated an unexpected and significant survival benefit, mortality [26, 27]. Secondary outcomes included ICU
in addition to less antibiotic exposure. and hospital LOS.
Therefore, with the aid of increased power of meta-
analytic techniques, we sought to expand the previous Quality assessment
analyses by including studies published recently, stratify- Two independent reviewers evaluated the quality of
ing different strategies to have a more accurate analysis of studies using the risk of bias tool recommended by the
the influence of different PCT algorithms to guide anti- Cochrane Collaboration [28]. We assigned a value of
microbial decisions. high, unclear, or low to the following items: sequence
generation; allocation concealment; blinding; incom-
Methods plete outcome data; selective outcome reporting; and
Search strategy and selection criteria other sources of bias. Discrepancies were identified and
This systematic review and meta-analysis were con- resolved through discussion.
ducted in accordance with the PRISMA guidance [25].
We searched RCTs in PubMed, Embase, Web of Science, Statistical analysis
and Cochrane Central Register of Controlled Trials from The results from all relevant studies were combined to
inception through 25 February 2017 to identify poten- estimate the pooled risk ratio (RR) and associated 95%
tially relevant studies. Search included the following key confidence intervals (CIs) for dichotomous outcomes. As
words: (“Procalcitonin” OR “PCT”) AND (“intensive to the continuous outcomes, mean differences (MD) and
care” OR “critically ill” OR “critical care”). No language 95% CI were estimated as the effect results. Heterogene-
restriction was imposed. Reference lists of relative arti- ity was tested by using the I2 statistic. An I2  <  50% was
cles were also reviewed. considered to indicate insignificant heterogeneity, and
Studies were included if they are enrolling adult ICU a fixed-effect model was used, whereas a random-effect
patients, with confirmed or suspected infection, assigned model was used in cases of significant heterogeneity
to either a PCT-guided therapeutic strategy group or a (I2 > 50%). Before data analysis, we estimated mean from
standard care group. Standard care referred to antimicro- median and standard deviations (SD) from IQR using the
bial regimens based on clinical signs, laboratory results, methods described in previous studies [29]. Sensitivity
and empiric rules or guidelines, without consideration analyses were performed by excluding trials that poten-
of PCT level. We excluded studies enrolling children or tially biased the results of primary outcomes. Publication
patients without any evidence of infection and studies bias was evaluated by visually inspecting funnel plots. All
Huang et al. Ann. Intensive Care (2017) 7:114 Page 3 of 10

analyses were performed using Review Manager version the included RCTs (Additional file 4: Table S4). Of these
5.3. included studies, seven were multicenter studies. A total
of 5136 patients comprised 2588 in the PCT-guided
Results group and 2548 in the standard care group. These studies
Study selection evaluated the effects of PCT-guided strategies in antibi-
The literature search yielded 881 records through data- otic discontinuation (n = 8) [23, 24, 31–34, 37, 39], anti-
base searching, and 13 RCTs fulfilled inclusion criteria biotic initiation (n = 3) [13, 30, 36], or the combination of
were eligible for the final analysis [13, 23, 24, 30–39]. The the antibiotic initiation and discontinuation (n = 2) [35,
overview of the study selection process is presented in 38]. Study population included surgical patients (n  =  3)
Fig.  1. In the study by Jensen et  al., some patients with- [30, 31, 33] and mixed medical–surgical patients (n = 10)
out infection were also included [13]; therefore, we only [13, 23, 24, 32, 34–39]. PCT assays adopted varied across
included patients with severe sepsis or septic shock the included studies.
that fulfilled our inclusion criteria from this study. The
Cochrane risk of bias score for each citation varied across Data synthesis
the studies (Additional file 2: Table S2). Procalcitonin‑guided discontinuation of antibiotics
The use of a PCT algorithm compared with standard care
Study characteristics to guide antibiotic discontinuation in critically ill patients
The main characteristics and predefined outcomes of the was evaluated in eight RCTs [23, 24, 31–34, 37, 39]. All
13 included studies are shown in Table 1 and Additional eight studies reported outcomes including total days
file 3: Table S3. The degree of non-compliance with PCT with antibiotics or antibiotic-free days. The aggregated
algorithm recommendations for antibiotics varied among data suggested that the duration of antibiotic treatment
was 1.67  days shorter in PCT-guided group (n  =  3404;
MD −  1.66  days; 95% CI −  2.36 to −  0.96; I2  =  71%;
P  <  0.00001) [23, 24, 31–34, 37, 39] (Fig.  2), while anti-
biotic-free days were 2.26  days longer (n  =  2120; MD
2.26  days; 95% CI 1.40–3.12; I2  =  0%; P  <  0.00001) [23,
24, 32, 34] when compared with that of standard care
group. Results showed patients in PCT-guided group
had lower short-term mortality than standard care group
(n = 3414; RR 0.86; 95% CI 0.76–0.98; I2 = 0%; P = 0.02)
[23, 24, 31–34, 37, 39] (Fig. 3), while no differences were
found in ICU LOS (n  =  3326; MD −  0.00  days; 95% CI
−  0.58 to 0.58; I2  =  0%; P  =  0.99) [23, 24, 31–33, 37,
39] and hospital LOS (n = 3290; MD 0.43 days; 95% CI
− 0.83 to 1.70; I2 = 30.4%; P = 0.50) [23, 24, 32, 34, 37,
39]. There was significant heterogeneity in the outcome
of duration of antibiotic treatment between the pooled
studies. Therefore, we conducted sensitivity analyses to
explore potential sources of heterogeneity. Exclusion of
the trial by Bloos and colleagues resolved the heterogene-
ity without alerting the result (n = 2338; MD − 1.97 days;
95% CI − 2.27 to − 1.68; I2 = 0%; P < 0.00001) [24, 31–
34, 37, 39].

Procalcitonin‑guided initiation of antibiotics

Three studies examined the efficacy of PCT-guided ini-
tiation of antibiotics [13, 30, 36]. Only one study exam-
ined the efficacy of PCT-guided initiation of antibiotics.
In this case, antibiotic consumption was comparable
between groups with the treatment days represented
62.6% and 57.7% of ICU stays in the PCT and standard
care groups, respectively (P  =  0.11) [36]. There was no
Fig. 1  Selection process for RCTs included in the meta-analysis statistically significant difference between groups in the
Table 1  Characteristics of included studies
Study/year Trial design Population Type of ICU N PCT/Ctrl PCT-guided group protocol Control group protocol PCT assay

Svoboda et al. [30] SC, P, R, OL Postoperative severe Surgical 38/34 AI: prompted change of ABT and catheter Standard evaluation by consultant PCT-Q
sepsis (≥ 2 ng/ml), prompted to repeated radio‑ surgeon
graphic and/or surgical evaluation (< 2 ng/
ml)
Schroeder et al. [31] SC, P, R, OL Postoperative severe Surgical 14/13 AD: if clinic signs and symptoms improved and According to clinical signs and LIA
sepsis PCT < 1 ng/ml or 25–35% of baseline empiric rules
Nobre et al. [32] SC, P, R, OL Sepsis Mixed 39/40 AD: if baseline PCT > 1 μg/L, re-evaluation at Regimens according to guidelines Kryptor
day 5. ABT discontinuation if PCT < 0.25 μg/L
or PCT dropped by > 90% from the baseline
peak level. If baseline PCT < 1 μg/L, re-
evaluation at day 3. ABT discontinuation if
Huang et al. Ann. Intensive Care (2017) 7:114

PCT < 0.1 μg/L and careful clinical evaluation

Hochreiter et al. [33] SC, P, R, OL Infection Surgical 53/57 AD: if clinic signs and symptoms improved and Standard regimen over 8 days LIA
PCT < 1 ng/ml or 25–35% of initial value over
3 days
Stolz et al. [34] MC, P, R, OL Ventilator-associated Mixed 51/50 AD: strongly encouraged (< 0.25 μg/L), According to clinical signs and Kryptor
pneumonia encouraged (0.25–0.5 μg/L or a empiric rules
decrease ≥ 80%), discouraged (0.5–1.0 μg/L
or a decrease < 80%) or strongly discouraged
(> 1.0 μg/L)
Bouadma et al. [35] MC, P, R, OL Bacterial infection or Mixed 311/319 AI: ABT was strongly discouraged (< 0.25 μg/L), Regimens according to interna‑ Kryptor
sepsis discouraged (0.25–0.49 μg/L), encouraged tional and local guidelines
(0.5–0.99 μg/L) or strongly encouraged
(≥ 1 μg/L)
AD: strongly encouraged (< 0.25 μg/L), encour‑
aged (0.25–0.49 μg/L). continuing of ABT was
encouraged (0.25–0.5 μg/L or > 80% peak)
and change of ABT (> peak concentration
and > 0.5 μg/L)
Jensen et al. [13] MC, P, R, OL Severe sepsis/septic Mixed 212/247 AI: if PCT ≥ 1 μg/L that was not decreasing by According to current guidelines Kryptor
shock at least 10% from previous day: increasing
the antimicrobial spectrum and intensify‑
ing diagnostic efforts to find uncontrolled
sources of infection
Layios et al. [36] SC, P, R, OL Infection Mixed 258/251 AI: ABT was strongly discouraged (< 0.25 μg/L), No reports Kryptor
discouraged (0.25–0.5 μg/L), encouraged
(0.5–1.0 μg/L) or strongly encouraged
(> 1.0 μg/L)
Annane et al. [38] MC, P, R, OL Septic shock Mixed 31/31 AI/AD: ABT was not to be started or was to ABT at the discretion of the Kryptor
be discontinuation (< 0.25 μg/L); strongly patient’s physician
discouraged (≥ 0.25 to < 0.5 μg/L); was
recommended (≥ 0.5 to < 5 μg/L) and
was strongly recommended (≥ 5 μg/L).
For patients enrolled ≤ 48 h after surgery,
the respective PCT cut-offs were < 4 μg/L,
4–9 μg/L and ≥ 9 μg/L
Page 4 of 10
 Huang et al. Ann. Intensive Care (2017) 7:114

Table 1  continued
Study/year Trial design Population Type of ICU N PCT/Ctrl PCT-guided group protocol Control group protocol PCT assay

Deliberato et al. [39] SC, P, R, OL Sepsis Mixed 42/39 AD: if PCT dropped > 90% from the peak level The possible source of infection Vidas
or the absolute value < 0.5 ng/ml and local susceptibility profile
Shehabi et al. [23] MC, P, R, SB Bacterial infection or Mixed 200/200 AD: cease ABT when PCT < 0.1 ng/ml or PCT According to the ABT guidelines Automated immu‑
sepsis was 0.1–0.25 ng/ml and infection is highly noassay analysers
unlikely or PCT level decreased > 90% from
baseline
De Jong et al. [24] MC, P, R, OL Infection Mixed 776/799 AD: if PCT value decreased over 80% or PCT Guidelines and the discretion of Vidas, Roche or
value lower than 0.5 μg/L attending physicians Kryptor machine
Bloos et al. [37] MC, P, R, OL Severe sepsis/septic Mixed 587/593 AD: stopping ABT if PCT level on day 7 or According to the local sepsis Kryptor
shock later < 1 ng/ml r or dropped > 50% from the guidelines
previous value

ABT antibiotics, AD antibiotic discontinuation, AI antibiotic initiation, Ctrl control, ICU intensive care unit, LIA immunoluminometric assay, MC multi-centre, Mixed surgical and medical intensive care unit, OL open label, P
prospective, PCT procalcitonin, PCT-Q procalcitonin immunochromatographic technology, R RCT, SC single centre
Page 5 of 10
Huang et al. Ann. Intensive Care (2017) 7:114 Page 6 of 10

Fig. 2  Effects of PCT-guided antimicrobial strategies on total days of antibiotics

Fig. 3  Effects PCT-guided antimicrobial strategies on short-term mortality

Huang et al. Ann. Intensive Care (2017) 7:114 Page 7 of 10

risk of short-term mortality (n = 1040; RR 1.01; 95% CI would result in a shorter duration of antibiotic treatment
0.84–1.23; I2 = 0%; P = 0.90) [13, 30, 36] (Fig. 3) or ICU of about 1.67  days, as well as lower short-term mortal-
LOS (n = 581; MD − 1.22 days; 95% CI − 4.34 to 1.90; ity compared with standard care. Due to the insufficient
I2 = 60%; P = 0.44) [30, 36]. evidence, a baseline PCT value should not be used as a
marker to guide antibiotic initiation.
Procalcitonin‑guided antibiotic initiation Our study had several strengths. The current meta-
and discontinuation analysis provided robust evidence to support and expand
Two studies employed a PCT-guided strategy of antibi- the weak suggestion in the 2016 SSC guidelines, i.e. use
otic initiation and discontinuation [35, 38]. No differences of low PCT level to assist the clinician in the discontinu-
were observed between the PCT and standard care group ation of empiric antibiotic [12]. In addition to those RCTs
in total days with antibiotics (n = 679; MD − 1.90 days, in previous meta-analyses, we had included three addi-
95% CI − 5.62 to 1.83; I2 = 96%; P = 0.32) (Fig. 2), antibi- tional RCTs recently published, and this added to the
otic-free days (n = 679; MD 1.31 days; 95% CI − 1.34 to statistical power by having 3414 cases to evaluate the pri-
3.95; I2 = 90%; P = 0.33), short-term mortality (n = 682; mary outcome. Moreover, we had stratified enrolled RCTs
RR 1.10; 95% CI 0.86–1.39; I2 = 30%; P = 0.46) (Fig. 3), according to different PCT-guided strategies, in order
the ICU LOS (n = 682; MD − 1.45 days; 95% CI − 0.91 to to eliminate the potential confounding factors caused
3.80; I2 = 0%; P = 0.23), and hospital LOS (n = 679; MD by different strategies. Though significant heterogeneity
−  0.43  days; 95% CI −  3.36 to 2.49; I2  =  0%; P  =  0.77) was observed among these studies, our sensitivity analy-
[35, 38]. ses demonstrated that the heterogeneity was resulted
Summary of findings for the effect of PCT-guided strat- from the trial by Bloss et al. [23]. This study was different
egy antibiotic on predefined outcomes in ICU patients is from the other trials in some aspects. On the one hand,
described in Table  2. We did not assess the publication it was designed as a 2 × 2 factorial trial and the interac-
bias because of the limited number of studies included in tion between the two treatment factors was unclear; on
each analysis. the other hand, in the trial [23], the clinicians used a 50%
decrease from previous value as a stopping rule, which
Discussion was lower than that of other studies. As a result, we also
PCT-guided strategies had been examined in multiple demonstrated a significant improvement in short-term
studies to optimize antibiotic treatment, with conflict- mortality associated with PCT-guided antibiotic discon-
ing results. The current meta-analysis justified a low tinuation. This had added robustness to findings of reduc-
PCT level to discontinue antibiotic treatment, which tion in antibiotics usage, since studies have demonstrated

Table 2  Summary of findings for the effect of procalcitonin-guided strategy on predefined outcomes in intensive care
unit patients
PCT-guided strategy Predefined outcome Number of trials N Estimated benefit with antibiotic I2 (%) P value

ABT discontinuation Duration of antibiotic use 8 3404 − 1.66 days (− 2.36, − 0.96) 71 P < 0.0001
Antibiotic-free days 4 2120 2.26 days (1.40, 3.12) 0 P < 0.0001
Short-term mortality 8 3414 0.86 (0.76, 0.98) 0 0.02
Length of stay in ICU 7 3326 − 0.00 days (− 0.58, 0.58) 0 0.99
Length of stay in hospital 6 3290 0.43 days (− 0.83, 1.70) 30 0.50
ABT initiation Duration of antibiotic use – – – – –
Antibiotic-free days – – – – –
Short-term mortality 3 1040 1.01 (0.84, 1.23) 0 0.90
Length of stay in ICU 2 581 − 1.22 days (− 4.34, 1.90) 60 0.44
Length of stay in hospital – – – – –
ABT initiation and discontinuation Duration of antibiotic use 2 679 − 1.90 days (− 5.62, 1.83) 96 0.32
Antibiotic-free days 2 679 1.31 days (− 1.34, 3.95) 90 0.33
Short-term mortality 2 682 1.10 (0.86, 1.39) 30 0.46
Length of stay in ICU 2 682 − 1.45 days (− 0.91, 3.80) 0 0.23
Length of stay in hospital 2 750 − 0.43 days (− 3.36, 2.49) 0 0.77
ABT antibiotics, PCT procalcitonin, ICU intensive care unit
Huang et al. Ann. Intensive Care (2017) 7:114 Page 8 of 10

that strategies aiming at restricting antibiotic overuse but not all RCTs, showed significant variability, ranging
could help improve survival [40, 41]. Our findings that from 0 to 59%. In most case of non-compliance, physi-
short-term mortality was significantly reduced in the cians were reluctant to stop antibiotics, even with a very
PCT-guided discontinuation group contrasted those of low PCT level (Additional file 4: Table S4), possibly due
previous meta-analyses. Despite the fact that no heteroge- to the concerns about the accuracy of single PCT value
neity was detected, the beneficial effect was clearly driven as a biomarker of infection [42, 43]. This might lead to
by the study results of de Jong et al. [24]. However, as the unnecessarily prolonged exposure of antibiotics, which
authors acknowledged, this study was a non-inferiority supported the robustness of our findings that imple-
study; therefore, the beneficial effect of mortality in PCT- mentation of PCT algorithm was associated with shorter
guided group was unexpected, which merited cautious duration of antibiotic treatment and longer antibiotic-
interpretation and further validation. free days.
As for a PCT strategy that combined initiation and dis- Our meta-analysis has some limitations. First, studies
continuation of antibiotics, we found no beneficial effect examining the PCT-guided strategies other than discon-
with regard to any predefined outcomes. The reason for tinuation of antibiotic treatment were scarce, with lim-
this failure may be that only two trials (one positive [35] ited number of studies available as well as small number
and one negative [38]) investigated the combined strat- of enrolled patients. As such, data on these strategies
egies varying in objective and the methodology. On the were insufficient to draw solid conclusions. Second, the
other hand, reported non-compliance rate was high. For high exclusion rate of screened patients in the included
example, in a prospective, multicenter, open-label rand- studies (such as immunosuppressed patients and those
omized trial (PRORATA trial) involving 630 non-surgical requiring long-term antibiotic therapy) precluded gen-
patients with suspected bacterial infections [35], patients eralization of the study results. Third, antibiotic strat-
in the PCT group had significantly more antibiotic-free egy in the control group (indications to initiate and
days. However, the algorithm-guided treatment recom- discontinue antibiotics) was not specified in most stud-
mendation was not strictly followed in 53% of patients ies. Whether the variation in antibiotic strategy, if any,
in the PCT group. Moreover, this trial [35] reported a in the control group might have affected the results of
higher standard deviation with regard to duration of anti- our meta-analysis is unclear. Thus, a more uniform
biotic treatment as well as antibiotic-free days (possibly approach to evaluating and reporting standard care
due to reported higher non-compliance rate), compared related to antibiotic use would be needed in future stud-
to that in the negative trial [38], which weaken its statis- ies. Fourth, the uneven distribution of different underly-
tic weight in the meta-analysis. Interestingly, it was note- ing diseases among included studies might also exert a
worthy that initial antibiotic prescription rate was similar prognostic value. Of note, in the two recently published
in PCT and standard care groups, suggesting that the systematic reviews and individual patient data meta-
improvement in antibiotic-free days was more likely the analysis [44, 45], Schuetz et  al. demonstrated with suf-
result of antibiotic discontinuation, while it was less likely ficient evidence that PCT-guided antibiotic treatment
due to exclusion of potential infection. in patients with acute respiratory infections reduced
In our study, we could not verify the efficacy of PCT- antibiotic exposure and side effects and improved sur-
based antibiotic initiation strategy because we found vival. Finally, different cut-off values of PCT and differ-
only three RCT through the literature search [13, 30, ent PCT measurements were reported across included
36]. Of these trials, only the trial by Layios and cowork- studies, which might also lead to bias in our results.
ers reported the antibiotics exposure and concluded that We had originally tried to perform subgroup analyses
PCT measuring for the initiation of antibiotics failed to exploring studies according to all the diversities. How-
decrease the antibiotic consummation [36]. The reasons ever, there were insufficient data.
for this failure may be that almost half of PCT serum
samples were >  1  μg/L, thus encouraging the antibiotic Conclusions
treatment, and the relative low proportion of patient- In summary, based on the results of our meta-analysis,
days with antibiotic treatment in the control group (57%). we recommend use of PCT to guide antibiotic discon-
Another reason could be related to the high non-com- tinuation, which was associated with a reduction in anti-
pliance rate with the PCT-guided antibiotic initiation biotic exposure and lower short-term mortality. Further
strategy described in the study. The authors reported that studies are needed to define the optimal cut-off value of
nearly 64% of patients in the PCT group received antibi- PCT for antibiotic discontinuation and to generalize our
otics regardless of a normal PCT level (< 0.5 μg/L). findings in other patient population including immu-
The incidence of non-compliance with the recommen- nocompromised patients and those received long-term
dations based on PCT algorithm, as reported in some, antibiotic therapy in ICU.
Huang et al. Ann. Intensive Care (2017) 7:114 Page 9 of 10

Additional files 3. Dellit TH, Owens RC, McGowan JE Jr, et al. Infectious Diseases Society of
America and the Society for Healthcare Epidemiology of America guide‑
lines for developing an institutional program to enhance antimicrobial
Additional file 1: Table S1. Summary of the RCTs included by previous stewardship. Clin Infect Dis. 2007;44(4):263–4.
and current meta-analysis. 4. Dimopoulos G, Poulakou G, Pneumatikos IA, Armaganidis A, Kollef MH,
Matthaiou DK. Short- vs long-duration antibiotic regimens for ventilator-
Additional file 2: Table S2. Predefined outcomes of the included RCTs.
associated pneumonia: a systematic review and meta-analysis. Chest.
Additional file 3: Table S3. Risk of bias table for included randomized 2013;144(6):1759–67.
control trials. 5. Najafi A, Khodadadian A, Sanatkar M, et al. The comparison of procalci‑
Additional file 4: Table S4. Non-compliance rate reported in the tonin guidance administer antibiotics with empiric antibiotic therapy in
included RCTs. critically ill patients admitted in intensive care unit. Acta Medica Iranica.
2015;53(9):562–7.
6. Evans HL, Lefrak SN, Lyman J, et al. Cost of gram-negative resistance. Crit
Care Med. 2007;35(1):89–95.
Abbreviations 7. Vincent JL, Rello J, Marshall J, et al. International study of the prevalence
CI: confidence interval; ICU: intensive care unit; IQR: interquartile range; LOS: and outcomes of infection in intensive care units. J Am Med Assoc.
length of stay; MD: mean difference; PCT: procalcitonin; RR: risk ratio; RCTs: ran‑ 2009;302(21):2323–9.
domized controlled trials; SD: standard deviations; VAP: ventilator-associated 8. Luyt CE, Combes A, Reynaud C, et al. Usefulness of procalcitonin for
pneumonia. the diagnosis of ventilator-associated pneumonia. Intensive Care Med.
2008;34(8):1434–40.
Authors’ contributions 9. Simon L, Lacroix J. Serum procalcitonin and C-reactive protein levels as
H-BH contributed to the conception of the study, data collection, analysis, markers of bacterial infection: a systematic review and meta-analysis. Clin
and drafting of the article. WJ and C-YW contributed to data collection and Infect Dis. 2004;39:206–17.
analysis. LW and J-MP contributed to design and revisions of this manuscript. 10. Wacker C, Prkno A, Brunkhorst FM, Schlattmann P. Procalcitonin as a diag‑
BD was responsible for the integrity of the work as a whole, from inception to nostic marker for sepsis: a systematic review and meta-analysis. Lancet
publication of the article. All authors read and approved the final manuscript. Infect Dis. 2013;13(5):426–35.
11. Schuetz P, Muller B, Christ-Crain M, et al. Procalcitonin to initiate or dis‑
Author details continue antibiotics in acute respiratory tract infections. Evid Based Child
1
 Medical ICU, Peking Union Medical College Hospital, Peking Union Medi‑ Health. 2013;8:1297–371.
cal College and Chinese Academy of Medical Sciences, 1 Shuai Fu Yuan, 12. Rhodes A, Evans LE, Alhazzani W, et al. Surviving sepsis campaign: inter‑
Beijing 100730, People’s Republic of China. 2 Department of Critical Care Medi‑ national guidelines for management of sepsis and septic shock. Crit Care
cine, The First Affiliated Hospital of Fujian Medical University, Fuzhou, China. Med. 2017;45(3):486–552.
13. Jensen JU, Hein L, Lundgren B, et al. Procalcitonin-guided interventions
Acknowledgements against infections to increase early appropriate antibiotics and improve
We thank Jing-Chao Luo, MD, for his assistance in searching the literature. survival in the intensive care unit: a randomized trial. Crit Care Med.
2011;39(9):2048–58.
Competing interests 14. Kopterides P, Siempos II, Tsangaris I, Tsantes A, Armaganidis A. Procalci‑
The authors declare that they have no competing interests. tonin-guided algorithms of antibiotic therapy in the intensive care unit: a
systematic review and meta-analysis of randomized controlled trials. Crit
Availability of data and materials Care Med. 2010;38(11):2229–41.
All data generated or analysed during this study are included in this published 15. Heyland DK, Johnson AP, Reynolds SC, Muscedere J. Procalcitonin for
article. reduced antibiotic exposure in the critical care setting: a systematic
review and an economic evaluation. Crit Care Med. 2011;39(7):1792–9.
Consent for publication 16. Agarwal R, Schwartz DN. Procalcitonin to guide duration of antimicro‑
Not applicable. bial therapy in intensive care units: a systematic review. Clin Infect Dis.
2011;53(4):379–87.
Ethics approval and consent to participate 17. Schuetz P, Chiappa V, Briel M, Greenwald JL. Procalcitonin algorithms
Not applicable. for antibiotic therapy decisions: a systematic review of randomized
controlled trials and recommendations for clinical algorithms. Arch Intern
Funding Med. 2011;171(15):1322–31.
Cams Innovation Fund for Medical Sciences (2016-12M-1-014). 18. Soni NJ, Pitrak DL, Aronson N, Samson DJ, Galaydick JL, Vats V. Procalci‑
tonin-guided antibiotic therapy. Agency for Healthcare Research and
Quality. 2012.
Publisher’s Note 19. Matthaiou DK, Ntani G, Kontogiorgi M, Poulakou G, Armaganidis A, Dimo‑
Springer Nature remains neutral with regard to jurisdictional claims in pub‑ poulos G. An ESICM systematic review and meta-analysis of procalci‑
lished maps and institutional affiliations. tonin-guided antibiotic therapy algorithms in adult critically ill patients.
Intensive Care Med. 2012;38(6):940–9.
Received: 7 July 2017 Accepted: 12 November 2017 20. Tang H, Huang T, Jing J, Shen H, Cui W. Effect of procalcitonin-guided
treatment in patients with infections: a systematic review and meta-
analysis. Infection. 2009;37(6):497–507.
21. Prkno A, Wacker C, Brunkhorst FM, Schlattmann P. Procalcitonin-guided
therapy in intensive care unit patients with severe sepsis and septic
shock—a systematic review and meta-analysis. Crit Care. 2013;17(6):R291.
References 22. Westwood M, Ramaekers B, Whiting P, et al. Procalcitonin testing to guide
1. Braykov NP, Morgan DJ, Schweizer ML, et al. Assessment of empirical antibiotic therapy for the treatment of sepsis in intensive care settings
antibiotic therapy optimisation in six hospitals: an observational cohort and for suspected bacterial infection in emergency department settings:
study. Lancet Infect Dis. 2014;14(14):1220–7. a systematic review and cost-effectiveness analysis. Health Technol
2. Llor C, Bjerrum L. Antimicrobial resistance: risk associated with antibiotic Assess. 2015;19(96):1–236.
overuse and initiatives to reduce the problem. Ther Adv Drug Saf. 23. Bloos F, Trips E, Nierhaus A, et al. Effect of sodium selenite administration
2014;5(6):229–41. and procalcitonin-guided therapy on mortality in patients with severe
Huang et al. Ann. Intensive Care (2017) 7:114 Page 10 of 10

sepsis or septic shock: a randomized clinical trial. JAMA Intern Med. 35. Bouadma L, Luyt CE, et al. Use of procalcitonin to reduce patients’ expo‑
2016;176(9):1266–76. sure to antibiotics in intensive care units (PRORATA trial): a multicentre
24. de Jong E, van Oers JA, Beishuizen A, et al. Efficacy and safety of proc‑ randomised controlled trial. Lancet. 2010;375(9713):463–74.
alcitonin guidance in reducing the duration of antibiotic treatment in 36. Layios N, Lambermont B, Canivet JL, et al. Procalcitonin usefulness for the
critically ill patients: a randomised, controlled, open-label trial. Lancet initiation of antibiotic treatment in intensive care unit patients. Crit Care
Infect Dis. 2016;16(7):819–27. Med. 2012;40(8):2304–9.
25. Moher D, Liberati A, Tetzlaff J, Altman DG, Group TP. Preferred reporting 37. Shehabi Y, Sterba M, Garrett PM, et al. Procalcitonin algorithm in critically
items for systematic reviews and meta-analyses: the PRISMA statement. J ill adults with undifferentiated infection or suspected sepsis. A rand‑
Chin Integr Med. 2010;8(5):336–41. omized controlled trial. Am J Respir Crit Care Med. 2014;190(10):1102–10.
26. Huang H, Li Y, Ariani F, Chen X, Lin J. Timing of tracheostomy in critically ill 38. Annane D, Maxime V, Faller JP, et al. Procalcitonin levels to guide antibi‑
patients: a meta-analysis. PLoS ONE. 2014;9(3):e92981. otic therapy in adults with non-microbiologically proven apparent severe
27. Wang F, Wu Y, Bo L, et al. The timing of tracheotomy in critically ill patients sepsis: a randomised controlled trial. BMJ Open. 2013;3(2):e002186.
undergoing mechanical ventilation: a systematic review and meta-analy‑ 39. Deliberato RO, Marra AR, Sanches PR, et al. Clinical and economic impact
sis of randomized controlled trials. Chest. 2011;140(140):1456–65. of procalcitonin to shorten antimicrobial therapy in septic patients with
28. Higgins JP, Altman DG, Gotzsche PC, et al. The Cochrane Collaboration’s proven bacterial infection in an intensive care setting. Diagn Microbiol
tool for assessing risk of bias in randomised trials. BMJ. 2011;343(oct18 Infect Dis. 2013;76(3):266–71.
2):d5928. 40. Levy SB, Marshall B. Antibacterial resistance worldwide: causes, chal‑
29. Wan X, Wang W, Liu J, Tong T. Estimating the sample mean and standard lenges and responses. Nat Med. 2004;10(12 suppl):S122–9.
deviation from the sample size, median, range and/or interquartile range. 41. Cohen ML. Epidemiology of drug resistance: implications for a post-
BMC Med Res Methodol. 2014;14(8):735–46. antimicrobial era. Science. 1992;257(257):1050–5.
30. Svoboda P, Kantorová I, Scheer P, Radvanova J, Radvan M. Can procalci‑ 42. Schuetz P, Affolter B, Hunziker S, et al. Serum procalcitonin, C-reactive
tonin help us in timing of re-intervention in septic patients after multiple protein and white blood cell levels following hypothermia after cardiac
trauma or major surgery? Hepatogastroenterology. 2007;54(74):359–63. arrest: a retrospective cohort study. Eur J Clin Invest. 2010;40(4):376–81.
31. Schroeder S, Hochreiter M, Koehler T, et al. Procalcitonin (PCT)-guided 43. Jung B, Molinari N, Nasri M, et al. Procalcitonin biomarker kinetics fails to
algorithm reduces length of antibiotic treatment in surgical intensive predict treatment response in perioperative abdominal infection with
care patients with severe sepsis: results of a prospective randomized septic shock. Crit Care. 2013;17(5):R25.
study. Langenbecks Arch Surg. 2009;394(2):221–6. 44. Schuetz P, Wirz Y, Sager R, et al. Procalcitonin to initiate or discontinue
32. Nobre V, Harbarth S, Graf JD, Rohner P, Pugin J. Use of procalcitonin to antibiotics in acute respiratory tract infections. Cochrane Database Syst
shorten antibiotic treatment duration in septic patients: a randomized Rev. 2017;10:CD007498.
trial. Am J Respir Crit Care Med. 2008;177(5):498–505. 45. Schuetz P, Wirz Y, Sager R et al. Effect of procalcitonin-guided antibi‑
33. Hochreiter M, Kohler T, Schweiger AM, et al. Procalcitonin to guide otic treatment on mortality in acute respiratory infections: a patient
duration of antibiotic therapy in intensive care patients: a randomized level meta-analysis. Lancet Infect Dis. 2017. https://doi.org/10.1016/
prospective controlled trial. Crit Care. 2009;13(3):R83. S1473-3099(17)30592-3.
34. 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(6):1364–75.