SHOCK, Vol. 43, No. 6, pp.

530Y539, 2015

Review Article
BLOOD PRESSURE TARGETS FOR VASOPRESSOR THERAPY:
A SYSTEMATIC REVIEW

Frederick D’Aragon,* Emilie P. Belley-Cote,* Maureen O. Meade,†

François Lauzier,‡ Neill K.J. Adhikari,§ Matthias Briel,|| Manoj Lalu,¶ Salmaan Kanji,**
Pierre Asfar,†† Alexis F. Turgeon,‡‡ Alison Fox-Robichaud,§§ John C. Marshall,||||
and François Lamontagne¶¶ for the Canadian Critical Care Trials Group
*Department of Anaesthesia, Division of Critical Care, McMaster University; and † Departments of Medicine,
and Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Ontario; ‡ Centre Hospitalier
Universitaire de Québec Research Center, Population Health and Optimal Health Practices, Division of
Critical Care Medicine, Department of Anesthesiology, Department of Medicine, Université Laval, Québec,
Québec; and §Department of Critical Care Medicine, University of Toronto, Toronto, Ontario, Canada;
||
Basel Institute for Clinical Epidemiology and Biostatistics, University Hospital Basel, Basel, Switzerland;
¶
Department of Anesthesiology, University of Ottawa, and **Department of Pharmacy, The Ottawa
Hospital, University of Ottawa, Ottawa, Ontario, Canada; †† Département de Réanimation Médicale et de
Médecine Hyperbare, Centre Hospitalier Universitaire Angers, Angers, France; ‡‡ Centre Hospitalier
Universitaire de Québec Research Center, Population Health and Optimal Health Practices,
Division of Critical Care Medicine, Department of Anesthesiology, Université Laval, Québec,
Québec; §§Department of Medicine, Thrombosis and Atherosclerosis Research Institute, McMaster
University, Hamilton; and ||||Li Ka Shing Knowledge Institute, St. Michaels Hospital, University of
Toronto, Toronto, Ontario; and ¶ ¶ Centre de Recherche du Centre Hospitalier Universitaire de
Sherbrooke, Département de médicine, Faculté de médecine et des sciences de la santé,
Université de Sherbrooke, Sherbrooke, Québec, Canada

Received 7 Feb 2014; first review completed 20 Feb 2014; accepted in final form 23 Jan 2015

ABSTRACT—Physicians often prescribe vasopressors to correct pathological vasodilation and improve tissue perfusion in
patients with septic shock, but the evidence to inform practice on vasopressor dosing is weak. We undertook a systematic
review of clinical studies evaluating different blood pressure targets for the dosing of vasopressors in septic shock. We
searched MEDLINE, EMBASE, CENTRAL (to November 2013), reference lists from included articles, and trial registries for
randomized controlled trials (RCTs) and observational and crossover intervention studies comparing different blood pres-
sure targets for vasopressor therapy in septic shock. Two reviewers independently selected eligible studies and extracted
data on standardized forms. We identified 2 RCTs and 10 crossover trials but no observational studies meeting our criteria.
Only one RCT measured clinical outcomes after comparing mean arterial pressure targets of 80 to 85 mmHg versus 65 to
70 mmHg. There was no effect on 28-day mortality, but confidence intervals were wide (hazard ratio, 95% confidence
interval [95% CI] 0.84 Y 1.38). In contrast, this intervention was associated with a greater risk of atrial fibrillation (relative risk,
2.36; 95% CI, 1.18 Y 4.72) and a lower risk of renal replacement therapy in hypertensive patients (relative risk, 0.75; 95% CI,
0.57 Y 1.0). Crossover trials suggest that achieving higher blood pressure targets by increasing vasopressor doses increases
heart rate and cardiac index with no effect on serum lactate. Our findings underscore the paucity of clinical evidence to guide
the administration of vasopressors in critically ill patients with septic shock. Further rigorous research is needed to establish an
evidence base for vasopressor administration in this population.
KEYWORDS—Vasopressors, critical care, systematic review, septic shock, randomized controlled trials

INTRODUCTION vasopressors under the assumption that increasing arterial blood

pressure improves tissue perfusion (2, 3). In septic shock, hy-
Because the severity of hypotension is associated with ad-
potension is largely attributable to pathological vasodilation (4),
verse outcomes in critically ill patients (1), physicians prescribe
and vasopressors constitute a mainstay of therapy (2). Animal
studies suggest that coronary, renal, and cerebral blood flow
Address reprint request to François Lamontagne, MD, Service de médecine
interne, Département de médecine, Faculté de médecine et des sciences de la santé,
autoregulation is lost with mean arterial pressures (MAPs)
Université de Sherbrooke, 3001 12e Avenue Nord, Sherbrooke, Québec, Canada below 60 mmHg (5, 6). Although ensuring that MAP remains
J1H 5N4. E-mail: Francois.Lamontagne@USherbrooke.ca. above this threshold seems to prevent intestinal injury in pre-
This work was performed at the Centre de Recherche du Centre Hospitalier
Universitaire de Sherbrooke, Université de Sherbrooke, Sherbrooke, Québec, Canada.
viously healthy children undergoing surgery (7), experts have
Supplemental digital content is available for this article. Direct URL citation highlighted the somewhat arbitrary nature of these targets
appears in the printed text and is provided in the HTML and PDF versions of this and point out that systolic blood pressure values or MAPs
article on the journal_s Web site (www.shockjournal.com).
DOI: 10.1097/SHK.0000000000000348
between 50 and 60 mmHg are also used to guide vasopressor
Copyright ! 2015 by the Shock Society use (8). Guidelines recommend titrating vasopressors to a target
530
Copyright © 2015 by the Shock Society. Unauthorized reproduction of this article is prohibited.
SHOCK JUNE 2015 OPTIMAL VASOPRESSOR TITRATION 531

MAP of 65 mmHg or more (9, 10), acknowledging that the studies, this is the most conservative statistical assumption. We also did not
quality of the evidence supporting this recommendation is weak consider possible interactions between treatment effect and treatment order.
We assessed the risk of bias in RCTs using the Cochrane Collaboration
(GRADE 1C) (10). toolkit (18). Specifically, we determined whether the investigators randomized
The risks of vasopressor-associated adverse effects, such as and concealed treatment allocation, blinded study personnel, obtained complete
visceral ischemia (11, 12), must be weighed carefully against the outcome data, and avoided selective reporting. For risk of bias assessment in
crossover trials, we derived explicit criteria from the recommendations of the
danger of under-resuscitating hypotensive patients. Correcting EPOC (Effective Practice and Organisation of Care) Cochrane Group (19).
hypotension with vasopressors while tissue dysoxia persists may Specifically, we assessed randomization and concealment of the treatment allo-
worsen outcomes (13). A recent survey of Canadian inten- cation sequence, blinding of study personnel, likelihood that the medical con-
dition would remain stable during the experiment, incomplete outcome data,
sivists suggests that Busual care[ regarding vasopressor use in selective reporting, prespecification of the primary outcome, and expected
sepsis is highly variable, especially when treating patients with effect. We assessed the risk of bias across studies following the previously
comorbidities (14). Faced with similar clinical scenarios, inten- cited GRADE approach (20) for 28-day mortality, atrial fibrillation, initia-
tion of renal replacement therapy, cardiac index, heart rate, and blood lactate
sivists adopt different strategies regarding vasopressor dosing. levels. We used funnel plots of standard error of treatment effect vs. treatment
Systematic reviews integrate existing information for ratio- effect to assess the risk of publication bias.
nal decision making (15). Even when available data do not Summary measures
provide clear guidance, explicit methods improve the reli- Mortality and other clinical outcomes were reported in only one RCT. For
ability of their conclusions. Although expert opinions may also secondary physiological outcomes reported in crossover trials, we pooled re-
sults for outcomes selected based on their clinical relevance under standard
guide clinical decisions, they may be discordant with best evi- care conditions. Thus, we summarized mean differences (MDs) in cardiac in-
dence (16). We undertook a systematic review of clinical stud- dex, heart rate, and blood lactate levels between patients treated with higher
ies evaluating different blood pressure targets for the dosing versus lower blood pressure targets.
Statistical pooling using a random-effects model (inverse variance method)
of vasopressors in septic shock. The research question was BDo was conducted with Review Manager (version 5.2). We describe statistical
higher blood pressure targets for vasopressor therapy impact heterogeneity with the I2 statistic. We did not plan subgroup analyses a priori.
mortality in patients with septic shock?[ Post hoc, we explored whether certain study characteristics could explain
statistical heterogeneity observed in the meta-analysis for cardiac index.
Available candidate explanatory variables were admission APACHE II (Acute
METHODS Physiology And Chronic Health Evaluation II) scores of more than 25, dif-
The design and reporting of this systematic review followed the PRISMA guide- ference in target mean arterial blood pressure of more than 20 mmHg, total
lines (see Supplemental Digital Content 1 at http://links.lww.com/SHK/A281). duration of the experiment of more than 120 min, propofol-based sedation, and
full publication of study versus abstract only.
Search strategy and study selection
With the assistance of a medical librarian, we developed a search strategy
for the electronic databases MEDLINE, EMBASE, and CENTRAL from each
database_s inception to November 2013 without any language restriction.
Database-specific subject headings and corresponding key words for sepsis and
septic shock were entered and combined with subject headings and key words
for vasopressors (see Supplemental Digital Content 2 at http://links.lww.com/
SHK/A282). Adequate sensitivity of the search strategy was demonstrated by
verifying that it captured studies previously identified as relevant from a broad
scoping review. In addition to clinical trials, we sought observational studies
reporting associations between blood pressure targets (as opposed to blood
pressure values) and clinical outcomes but excluded other noninterventional
studies (care reports, case series). Two reviewers independently screened ab-
stracts for potentially relevant studies; the full text publication of any citation
considered relevant by either reviewer was retrieved. The two reviewers read
all potentially relevant publications and selected randomized controlled trials
(RCTs), crossover intervention studies, and observational studies evaluating
the impact of blood pressure targets on patients with septic shock. Blood
pressure targets, as opposed to observed blood pressure values or vasopressor
doses, reflect intentional therapeutic strategies. Accordingly, we excluded ob-
servational studies reporting associations between clinical outcomes and either
observed vasopressor doses or blood pressure because these associations are
prone to residual confounding. We also excluded comparisons between dif-
ferent agents titrated to the same resuscitation end points, comparisons among
different baseline preresuscitation clinical states, and comparisons between
fixed doses of the same drug. We considered studies enrolling patients of any
age and used the authors_ definitions of septic shock. In addition, reviewers
scanned the references of included studies for additional eligible studies and
searched www.clinicaltrials.gov for unpublished studies. Disagreements on
article selection were resolved in consultation with a third reviewer.

Data extraction and analysis

Using pretested standardized forms, two reviewers independently abstract-
ed data from each included study. Forms captured information related to study
design, risk of bias, and results. We contacted the investigators of one RCT
because the original study report did not include parallel arm comparisons
(17). For the purpose of this systematic review, this study was treated as a
crossover trial. When studies examined more than two blood pressure targets,
we used the lowest and highest blood pressure targets for the meta-analytical
comparison. When pooling crossover trials, we did not account for the pairing
of data (correlation coefficient between comparisons equals zero). In the
absence of studies reporting the SD of mean differences in relevant crossovers FIG. 1. Flow diagram.

Copyright © 2015 by the Shock Society. Unauthorized reproduction of this article is prohibited.
532 SHOCK VOL. 43, NO. 6 D_ARAGON ET AL.

RESULTS points unrelated to blood pressure (41) and, another, to vari-

Literature search
The initial literature search yielded 4,418 citations. After
title and abstract review, we screened the full text of 67 pub-
lications and ultimately included two RCTs (21, 22) and 10
crossover trials (Fig. 1). Among the studies that we excluded
because they did not compare resuscitation end points were
observational and crossover studies that measured associations
between dose of vasopressor therapy or severity of hypoten-
sion with clinical outcomes (23Y25) or that compared different
agents titrated to the same resuscitation end points (26Y29).
We also excluded studies that compared the effects of fixed
doses of vasopressors (30Y37) or of vasopressor therapy versus
baseline status before resuscitation was initiated (38, 39). One
observational study described prescribed blood pressure tar-
gets, but clinical outcomes were not reported (40); one study
measured the immediate effects of vasopressors titrated to end
able pre-illness blood pressure values (42). We identified only
one additional ongoing trial comparing different blood pres-
sure targets in shock (OVATION [Optimal Vasopressor Ti-
tration], NCT01800877) the study is not reported.
Study characteristics
Important differences separate the two RCTs (Table 1). Suk
et al. (21) enrolled 16 patients with septic shock within the
first 6 h of diagnosis. The investigators adjusted norepineph-
rine infusions to achieve a MAP target of 65 mmHg in one
group and 85 mmHg in the other for 8 h. However, the inter-
pretation of this trial is challenged by the fact that neither
achieved blood pressure nor norepinephrine dose differed be-
tween the two arms, suggesting poor protocol adherence or
between-arm differences in illness severity. The other larger
more recent trial enrolled 776 patients with septic shock within

TABLE 1. Randomized controlled trials evaluating vasopressor titration to different blood pressure targets in humans with sepsis
Suk et al. (21) Asfar et al. (22)
n total 16 776
Eligibility Within 6 h of septic shock Within 6 h of vasopressor initiation
Primary outcome Unclear 28-day mortality
Vasopressor used Norepinephrine Not protocolized, but 95.5% were receiving norepinephrine at baseline
n per group 8 8 388 388
Mean arterial pressure target, mmHg 65 85 65 Y 70 80 Y 85
Age, mean (SD) 64 (8) 50 (16) 65 (15) 65 (13)
SOFA score, mean (SD) 13.0 (2.6) 13.0 (2.0) 10.8 (3.1) 10.7 (3.1)
Source of infection, n (%)
Lung 1 (12.5) 2 (25) 200 (51.5) 202 (52.1)
Intra-abdominal 3 (37.5) 4 (50) 67 (17.3) 65 (16.8)
Urinary tract infection 0 (0) 1 (12.5) 44 (11.3) 44 (11.3)
Bacteremia 1 (12.5) 0 (0) NR NR
Others 3 (37.5) 1 (12.5) 73 (18.8) 72 (18.6)
Vasopressor treatment duration, Protocolized duration of 8 h 3.7 (3.2) 4.7 (3.7)
mean (SD), days
Blood pressures achieved, mmHg Unclear Day 1 MAP
73.1 (72.7 to 73.5) 82.4 (82.0 to 82.8)
28-Day mortality, n (%) NR NR 132 (34.0) 142 (36.6)
90-Day mortality, n (%) NR NR 164 (42.3) 170 (43.8)
Adverse events, n (%) Unclear Unclear
Myocardial infarction 2 (0.5) 7 (1.8)
Atrial fibrillation 11 (2.8) 26 (6.7)
Ventricular fibrillation/tachycardia 15 (3.9) 22 (5.7)
Measures of macrocirculation Cardiac output: NS NR
Creatinine clearance: NS
Measures of microcirculation Mixed venous oxygen saturation: NS NR
Lactate production and elimination: NS
Gastric tonometry: NS
Clearance of ICG: NS
ICG indicates indocyanine green; NR, not reported; NS, not significant (P value not reported); BP, blood pressure; MAP, mean arterial pressure.

Copyright © 2015 by the Shock Society. Unauthorized reproduction of this article is prohibited.
SHOCK JUNE 2015 OPTIMAL VASOPRESSOR TITRATION 533

6 h of vasopressor initiation (22). Patients received vasopres-
sors to achieve a MAP of 65 to 70 mmHg or 80 to 85 mmHg.
Although the choice of vasopressor was left to the treating
team, norepinephrine was the most common agent. The pri-
mary outcome was 28-day mortality, but patients were fol-
lowed until day 90. In this study, measured blood pressure and
vasopressor doses were higher in the higher MAP.
One study enrolled 28 patients with septic shock and dys-
function of two or more organs and randomized them to re-
ceive norepinephrine to attain a MAP of 65 mmHg for 8 h
(n = 14) or 65 mmHg for 4 h followed by 85 mmHg for 4 h
(n = 14) (17). The publication of this parallel-arm clinical trial
reports only before-after observations of physiological variables
in the experimental arm. After contacting the investigators, we
conducted parallel-arm comparisons and found no statistically
significant differences at the end of this 8-h experiment. We
therefore classified this study as a crossover trial.
All crossover trials were small. Sample sizes ranged from
8 to 32. In every experiment, adjustments in vasopressor doses
hinged on MAP targets (Table 2). The median initial target
was 65 mmHg, and the median maximal target was 85 mmHg.
The duration of the study ranged between 10 and 405 min, and
norepinephrine was uniformly used across all studies. Heart
rate, cardiac index, and serum lactate were the most consis-
tently reported outcomes. A number of studies also measured
microcirculatory blood flow.

TABLE 2. Crossover studies evaluating vasopressor titration to different blood pressure targets in humans with sepsisVdesign
Duration of
intervention at Total
APACHE II Sequential steady targeted duration
Description Score, mean MAP targets, MAP before of the Primary
n of population or median (SD) mmHg measurements, min study, min VP used outcome
Ledoux 2000 10 Septic shock requiring 29 (2.7) 65, 75, 85 60 + (equilibration 315 NE NR
VP to maintain MAP period = 45)
Q60 mmHg with
PAOP Q12 mmHg
Bourgoin et al. (17)* 14 Severe septic shock 28 (NR) 65, 85 240 + (equilibration 480 and 525 NE NR
with hemodynamic period = 45)
instability and 2
or more organ
dysfunctions
Deruddre 2007 11 Within 12 h of ICU NR 65, 75, 85 120 (equilibration 360 NE Renal resistances
admission and period = NR)
requiring VP after
fluid resuscitation
Kozieras 2007 24 Septic shock 26 (8) Baseline†, 90 5 10 NE NR
mechanically
ventilated
Ibsen 2007 8 Septic shock with NR 65, 75, 85 120 360 NE Luminal concentrations
NE 90.1 2g kgj1 of L-lactate in
minj1 to reach the rectum
MAP ~70 mmHg
with stable NE
dose for 2 h and
ScvO2 970%
Dubin 2009 20 Mechanically ventilated 24.4 (5.4) 65, 75, 85 30 (equilibration 180 NE NR
and requiring VP period = NR)
to maintain MAP
Q65 mmHg despite
fluid resuscitation
Jhanji 2009 16 Septic shock requiring 23 (10) 60, 70, 80, 90 45 (equilibration 240 NE Sublingual
VP despite fluid period = NR) microvascular
resuscitation flow index
Tishkov 2009 20 Septic shock requiring NR 65, 75, 85 NR NR NE NR
VP to reach MAP
Q65 mmHg after
fluid resuscitation to
a PAOP Q12 mmHg
Georger 2010 32 Septic shock without NR 65 Y 70, 70 Y 80, 980 NR NR NE Tissue hemoglobin
fluid responsiveness saturation
and MAP 980 mmHg
on NE
Thooft 2011 13 Septic shock for 22.7 (6.9) 65, 75, 85 30 (equilibration 150 NE NR
G48 h requiring NE period = 15)

*Although this study randomized patients to two arms, the publication only described the before and after changes in the group where vasopressors were sequentially titrated
to a MAP of 65 and 85 mmHg.
†
Baseline MAP targets were not reported, but mean baseline blood pressure was 66 mmHg.
ICU indicates intensive care unit; MAP, mean arterial pressure; NE, norepinephrine; NR, not reported; NS, not significant; PAOP, pulmonary artery occlusion pressure; VP,
vasopressor.

Copyright © 2015 by the Shock Society. Unauthorized reproduction of this article is prohibited.
534 SHOCK VOL. 43, NO. 6 D_ARAGON ET AL.

TABLE 3. Risk of bias assessmentVRCTs

Suk et al. (21) Asfar et al. (22)
Allocation sequence generation Unclear Low
Concealment of allocation Unclear Low
Blinding of outcome assessors Unclear Low for mortality (primary outcome);
high for other outcomes*
Loss to follow-up; incomplete outcome data Unclear Low
Selective outcome reporting Unclear Low
Other problems Not stopped early; no protocol for cointerventions; Not stopped early
unclear for outcome assessors interventions
Stopped early? Protocol for relevant cointerventions?
Could the outcomes have been influenced by unblinded
outcome assessors?
*Lack of blinding affects outcomes that involve subjective assessments.

Methodologic quality
Risk of bias was high in the smallest RCT (21) because of the
lack of allocation concealment and blinding of outcome asses-
sors (Table 1). Although the larger RCT (22) was also un-
blinded, the primary outcome could not be biased by subjective
assessments. For outcome assessments potentially influenced by
knowledge of treatment allocation, like the need for renal re-
placement therapy, risk of bias is high (Table 3).
Overall, risk of bias was high in crossover trials because
of lack of randomization of the treatment sequence and un-
blinded outcome assessments (Table 4). Few studies pre-
specified a primary outcome and the expected effect. Also,
confidence in the stability of the illness during the entire
course of the experiment was low and observations could be
attributable to changes in patients_ status across time rather
than the intervention.
Data synthesis
Randomized controlled trial results do not suggest that one
blood pressure target is superior (Table 1). Suk et al. (21) re-
ported no difference in cardiac index, diuresis, mixed venous
oxygen saturation, lactate production, creatinine clearance, pH/
base excess, hepatic blood flow, and gastric PCO2 gap. Asfar
et al. (22) found no difference in mortality (hazard ratio 1.07
[0.89 Y 1.29] at 28 days and hazard ratio 1.03 [0.88 Y 1.22]
at 90 days). However, the investigators reported a greater risk

TABLE 4. Risk of bias assessmentVcrossover studies

Incomplete data
Random Stable reporting (loss Selective Prespecified Prespecification
sequence Concealed Blinding across time to follow-up) reporting primary outcome of effect
LeDoux 2000 High High (High) High (Low) (Low) High High
Bourgoin et al. (17) High High (High) High (Low) (Low) High High
Deruddre 2007 High High (High) High (Low) (Low) Low High
Kozieras High High (High) Low (Low) (Low) High High
Ibsen 2007 Low Low (High) High High (Low) Low High
Dubin 2009 High High (High) High (Low) (Low) High Low
Jhanji 2009 High High Low for all sublingual High High (Low) Low High
microcirculation
measurements
(High) for cutaneous
microcirculation
measurements
Tishkov 2009 High High (High) (High) (Low) (High) High High
Georger 2010 High High (High) (High) (Low) High Low High
Thooft 2011 High High (High) High High1 (Low) High Low
(High): not clearly reported but judged to be high.
(Low): not clearly reported but judged to be low.
Stable across time: Given the rapidly changing nature of septic shock, we considered that a patient’s condition was stable if all measurements occurred
within 30 min (for all BP targets).
Selective reporting: When protocols not published, we deemed studies at (low) risk of bias if they reported at least some hemodynamic outcomes in
addition to either macrocirculation or microcirculation data.
Incomplete data reporting (loss to follow-up): When not clearly stated, we deemed that studies were at (low) risk of bias unless some patients were
excluded from some analyses.
*Some outcomes were reported only for some patients.

Copyright © 2015 by the Shock Society. Unauthorized reproduction of this article is prohibited.
SHOCK JUNE 2015 OPTIMAL VASOPRESSOR TITRATION 535

of atrial fibrillation in the higher MAP arm (6.7% vs. 2.8%,
P = 0.02). Although the proportion of patients requiring renal
replacement therapy was not different overall (33.5% vs. 35.8%,
P = 0.5), the subgroup of patients with a history of hyperten-
sion required less renal replacement therapy during the first
week of the trial (32% vs. 42%, P = 0.046).
The results of the crossover trials are summarized in Table 5.
Pooled results suggest that achieving higher MAPs is associated
with higher cardiac indices (MD, 0.65 L minj1 mj2; 95% CI,
0.2 Y 1.1) and heart rates (MD, 5 beats minj1; 95% CI, 2 Y 8)
but not with significant changes in lactate levels, which were in
the normal range or only mildly elevated (Fig. 2). The I2 sta-
tistic for cardiac index was 80%. Post hoc subgroup analyses
based on experiment duration, propofol-based sedation, and pub-
lication status reduced statistical heterogeneity (see Supplemental
Digital Contents 3, 4, and 5 at http://links.lww.com/SHK/A283,
http://links.lww.com/SHK/A284, and http://links.lww.com/SHK/A285).
Funnel plots suggest that the positive association between higher
MAP targets and heart rate but not cardiac indices could be the
result of publication bias (Fig. 3).
A summary of the available evidence for risk of death, atrial
fibrillation, and requirement for renal replacement therapy

TABLE 5. Results of crossover studies evaluating vasopressor titration to different blood pressure targets in sepsis
VP dose Hazard ratio CI Lactate Indicators of microcirculatory flow
LeDoux 2000 j j NS NS Gastric tonometry: NS
Arterial to intramucosal (gastric) PCO2 gradient: NS
Cutaneous capillary blood flow: NS
Cutaneous red cell velocity: NS
Bourgoin et al. (17) j NS j NS NR
Deruddre 2007 j* NS j* NS NR
Kozieras 2007 j NS j NR NR
Ibsen 2007 j* NR NS NS L-lactate in gastric lumen: NS
L-lactate in rectal lumen: NS
Dubin 2009 j NS j NS Gastric tonometry: NS
Sublingual vascular density: NS
Sublingual MFI: NS
Sublingual proportion of perfused vessels: NS
Sublingual heterogeneity flow index: NS
Jhanji 2009 j NS j† NS Cutaneous PtO2: j
Cutaneous microvascular red blood cell flux: j
Sublingual MFI: NS
Sublingual vessel density: NS
Sublingual proportion of perfused vessels: NS
Sublingual perfused vascular density: NS
Sublingual heterogeneity flow index: NS
||
Tishkov 2009 NR NR j NS Gastric arterial-intramucosal PCO2 gradient: NS
Cutaneous capillary blood flow: NS
Cutaneous microvascular red blood cell velocity: NS
Georger 2010 NR NR NR NR Tissue hemoglobin saturation: NS
Recovery slope after vascular occlusion test: NS
Thooft 2011 j§ NS j|| ,§ Thenar eminence StO2: NS
Thenar eminence descending slope of StO2: NS
Thenar eminence ascending slope of StO2: j§
Sublingual MFI: j||
Sublingual proportion of perfused vessels: NS
Sublingual vessel density: NS
Sublingual perfused vessel density: j||
*Only significant for 65 mmHg vs. 75 mmHg.
†
Only significant for MAP 65 mmHg vs. 85 mmHg.
‡
Only significant for 80 mmHg and 90 mmHg vs. 60 mmHg.
§
Only significant for 75 mmHg and 85 mmHg vs. 65 mmHg.
||
Only significant for 85 mmHg vs. 65 mmHg.
NR indicates not reported; NS, not significant; MFI, microvascular flow index; StO2, tissue hemoglobin oxygen saturation.

Copyright © 2015 by the Shock Society. Unauthorized reproduction of this article is prohibited.
536 SHOCK VOL. 43, NO. 6
FIG. 2. Pooled data from crossover trials. A, Cardiac index. B, Heart rate. C, Lactate.
appears in Table 6. Only one trial provides data for these
clinical end points (22). In absolute terms, this trial shows that
a MAP target of 80 to 85 mmHg could be associated with a
large survival benefit (number needed to treat of 26) or sig-
nificant harm (number needed to harm of 10). A summary of
effects table for heart rate, cardiac index, and blood lactate
level presents mean changes with higher blood pressure targets
in crossover trials (Table 7).
DISCUSSION
The findings from this systematic review of clinical studies
comparing the effects of different blood pressure targets for
patients in septic shock neither support nor reject the current
recommendation of the Surviving Sepsis Campaign (10). Out
of three published clinical trials, only one addressed the
effect of different blood pressure targets sustained during a
course of many days and reported on clinical outcomes includ-
ing mortality (22). However, this trial was underpowered to
rule out a clinically important improvement in survival. Statis-
tically significant differences in secondary outcomes suggest
that benefit (lower risk of renal replacement therapy in chroni-
cally hypertensive patients) and harm (higher risk of atrial fi-
brillation) may result from setting higher blood pressure targets
for vasopressors. Crossover experiments suggest that higher
blood pressure targets are associated with faster heart rates and
higher cardiac indices without short-term differences in lactate
levels. These results apply to norepinephrine.
This review underscores the paucity of published clinical
data to inform practice regarding vasopressor dosing. The
Copyright © 2015 by the Shock Society. Unauthorized reproduction of this article is prohibited.
D_ARAGON ET AL.
strength of this work lies in the application of rigorous sys-
tematic review methodology to a clinically important but
overlooked research question. Systematic reviews finding in-
sufficient data to guide clinical practice underscore the un-
certain effects of variation in usual practices. Nevertheless, our
study has limitations. First, we focused on comparisons be-
tween blood pressure targets. Although some clinicians may
titrate vasopressors to other physiologic end points, these al-
ternative practices are less common and not incorporated in
clinical practice guidelines. Second, the association between
higher blood pressure targets and cardiac index varies con-
siderably across crossover trials, suggesting that it may depend
on specific study characteristics. Unfortunately, many poten-
tial explanatory variables were not measured in included
studies. Post hoc subgroup analyses should be considered hy-
pothesis generating. At this time, we cannot identify with
certainty under what conditions this association exists. Third,
although our search strategy for published studies was com-
prehensive, we cannot rule out publication bias. Funnel plots
analysis suggests that the risk of publication bias is more likely
to affect the positive association between blood pressure tar-
gets and heart rate. Fourth, we found no observational study
directly addressing our research question, which highlights the
paucity of the evidence informing vasopressor use. Cohort
studies reporting associations between vasopressor doses or
observed blood pressure values and clinical outcomes are sub-
ject to residual confounding and are not suitable to provide
guidance for titration of vasopressors. Finally, given the limited
experimental evidence in humans, a summary of preclinical data
SHOCK JUNE 2015 OPTIMAL VASOPRESSOR TITRATION 537

Increasing research activity in this field reflects a growing

interest in the question of optimal vasopressor dosing and
provides context to our work. Our findings point to a paradox
of contemporary critical care: fewer than 850 patients have
been enrolled in clinical trials of blood pressure targets for
septic shock, yet administration of vasopressors to the sickest
patients in our hospitals occurs daily. Because strong recom-
mendations for clinical care require concordant results from a
number of clinical trials (10, 20), further research comparing
the target recommended by the Surviving Sepsis Campaign
with alternative dosing regimens is needed.
Future directions
In addition to the data incorporated in this systematic re-
view, our work highlights important questions that remain
unaddressed. First, certain patient subgroups may require dif-
ferent blood pressure targets. Although the SEPSISPAM Trial
addressed the potential effect modifier role of chronic hyper-
tension (22), other preliminary data suggest that race could
influence response to vasopressors (45). Second, as mentioned,
the studies included herein used norepinephrine. Whether dif-
ferent blood pressure targets achieved with other vasopressors
would have different effects may prove important. The dif-
ferent effects of dopamine and norepinephrine on mortality
(46) could be specific to the dose that was necessary to achieve
the blood pressure that was targeted. Using different blood
pressure targets, the splanchnic preservation of dopamine
(47, 48) may be achieved without the adverse cardiac effects.
The effects of vasopressin analogs may also hinge on the blood
pressure targets used in previous clinical trials. If these agents
were used to target a higher blood pressure without increasing
catecholamines, improvements in renal function may occur
without an increased risk of cardiac arrhythmias. Terlipressin
has been used safely in both adults and children (49Y51).
Third, specific practical methodological aspects of existing
trials may influence study results. For example, radial arterial
lines frequently underestimate central arterial pressure, which
FIG. 3. Funnel plots. A, Cardiac index. B, Heart rate. C, Lactate. could lead to unnecessary upward titration of vasopressor
doses (52). Fourth, what are the long-term effects (e.g., cog-
may be of interest. We opted to focus on clinical studies be- nitive and physical function) of higher blood pressure targets
cause the limited clinical relevance of animal experiments pre- and what outcomes (e.g., renal function versus atrial fibrilla-
cludes any inference to clinical practice even when the results tion) are most important to patients? And finally, considering
are compelling (43, 44). that physicians frequently prescribe blood pressure targets but
TABLE 6. Summary of the evidence on risk of death, atrial fibrillation, and requirement for renal replacement therapy
Illustrative comparison
Assumed risk Corresponding risk
MAP target MAP target
Outcomes 65 Y 70 mmHg 80 Y 85 mmHg RR (95% CI) n Quality of evidence Comments
28-Day mortality 34 per 100 36.6 per 100 (30 Y 44) 1.08 (0.89 Y 1.30) 776 Moderate due to Although 3 RCTs have
imprecision compared the effects of
different blood pressure
targets, only 1 measured
clinical outcomes
Atrial fibrillation 2.8 per 100 6.7 per 100 (3.3 Y 13.2) 2.36 (1.18 Y 4.72) 776 High
Renal replacement therapy 35.8 per 100 33.5 per 100 (27.6 Y 40.5) 0.94 (0.77 Y 1.13) 776 Moderate because
of lack of blinding
MAP indicates mean arterial pressure; RR, relative risk; 95% CI, 95% confidence interval; RCT, randomized controlled trial.

Copyright © 2015 by the Shock Society. Unauthorized reproduction of this article is prohibited.
538 SHOCK VOL. 43, NO. 6
TABLE 7. Summary of effects for heart rate, cardiac index, and blood lactate levels
Outcomes MD (95% CI) n Quality of evidence
Cardiac index, L minj1mj2 0.65 (0.2 Y 1.1) 128 Low because of high risk of bias 1) These effects represent short-term effects of higher
Heart rate, beats minj1 5 (2 Y 8) 108 Low because of high risk of bias 2) Average blood lactate levels when patients were treated
Blood lactate, mEq Lj1 0.08 (j0.6 to 0.4) 70 Low because of high risk of bias
MD indicates mean difference (mean in higher MAP target Y mean with lower MAP target); 95% CI, 95% confidence interval.
adjust these end points to other physiological variables, clini-
cal trials comparing vasopressor dosing with different end
points are lacking. Many studies report the prognostic value of
lactate clearance and the feasibility of using real-time varia-
tions as markers of therapeutic success, even within normal
range (9, 53, 54). Echocardiography (55) and near-infrared
spectroscopy (56) may provide acceptable resuscitation end
points. However, despite the limitations of gross resuscitation
end points such as MAP (57), the superiority of these alter-
natives should be proven in rigorous clinical trials powered
to show meaningful differences in important outcomes.
CONCLUSIONS
This systematic review underscores the paucity of clinical
evidence to guide the common administration of vasopressors
in critically ill patients with septic shock. Further rigorous re-
search is needed to establish an evidence base for vasopressor
administration in this population.
ACKNOWLEDGMENTS
The authors thank Neera Bhatnagar, Claudio Martin, Lauralyn McIntyre, and
Robert Green for their critical review of the manuscript and Nicolay Ferrari for his
administrative support. This work was conducted without designated funding.
REFERENCES
1. Jones AE, Aborn LS, Kline JA: Severity of emergency department hypotension
predicts adverse hospital outcome. Shock 22(5):410Y414, 2004.
2. Hollenberg SM: Vasopressor support in septic shock. Chest 132(5):1678Y1687, 2007.
3. Hollenberg SM: Vasoactive drugs in circulatory shock. Am J Respir Crit Care
Med 183(7):847Y855, 2011.
4. Buckberg G, Cohn J, Darling C: Escherichia coli bacteremic shock in con-
scious baboons. Ann Surg 173(1):122Y130, 1971.
5. Hollenberg SM, Ahrens TS, Annane D, Astiz ME, Chalfin DB, Dasta JF, Heard
SO, Martin C, Napolitano LM, Susla GM, et al.: Practice parameters for he-
modynamic support of sepsis in adult patients: 2004 update. Crit Care Med
32(9):1928Y1948, 2004.
6. Kirchheim HR, Ehmke H, Hackenthal E, Löwe W, Persson P: Autoregulation
of renal blood flow, glomerular filtration rate and renin release in conscious
dogs. Pflugers Arch 410(4Y5):441Y449, 1987.
7. Thuijls G, Derikx JP, de Kruijf M, van Waardenburg DA, van Bijnen AA,
Ambergen T, van Rhijn LW, Willigers HM, Buurman WA: Preventing enterocyte
damage by maintenance of mean arterial pressure during major nonabdominal
surgery in children. Shock 37(1):22Y27, 2012.
8. Dries DJ: Vasoactive drug support in septic shock. Shock 26(5):529Y530, 2006.
9. da Silva Ramos FJ, Azevedo LC: Hemodynamic and perfusion end points for
volemic resuscitation in sepsis. Shock 34(Suppl 1):34Y39, 2010.
10. Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, Opal SM,
Sevransky JE, Sprung CL, Douglas IS, Jaeschke R, et al.: Surviving sepsis
campaign: international guidelines for management of severe sepsis and septic
shock: 2012. Crit Care Med 41(2):580Y637, 2013.
11. Bulkley GB, Womack WA, Downey JM, Kvietys PR, Granger DN: Collateral
blood flow in segmental intestinal ischemia: effects of vasoactive agents. Sur-
gery 100(2):157Y166, 1986.
Copyright © 2015 by the Shock Society. Unauthorized reproduction of this article is prohibited.
D_ARAGON ET AL.
Comments
blood pressure targets.
with lower blood pressure targets were in the normal
range or mildly elevated.
12. Pawlik W, Shepherd AP, Jacobson ED: Effect of vasoactive agents on intestinal
oxygen consumption and blood flow in dogs. J Clin Invest 56(2):484Y490, 1975.
13. Sterling SA, Puskarich MA, Shapiro NI, Trzeciak S, Kline JA, Summers RL, Jones
AE; Emergency Medicine Shock Research Network (EMSHOCKNET): Charac-
teristics and outcomes of patients with vasoplegic versus tissue dysoxic septic
shock. Shock 40(1):11Y14, 2013.
14. Lamontagne F, Cook DJ, Adhikari NK, Briel M, Duffett M, Kho ME, Burns
KE, Guyatt G, Turgeon AF, Zhou Q, et al.: Vasopressor administration and
sepsis: a survey of Canadian intensivists. J Crit Care 26(5):532 e1Ye7, 2011.
15. Mulrow CD: Rationale for systematic reviews. BMJ 309(6954):597Y599, 1994.
16. Antman EM, Lau J, Kupelnick B, Mosteller F, Chalmers TC: A comparison
of results of meta-analyses of randomized control trials and recommenda-
tions of clinical experts. Treatments for myocardial infarction. JAMA 268(2):
240Y248, 1992.
17. Bourgoin A, Leone M, Delmas A, Garnier F, Albanèse J, Martin C: Increasing
mean arterial pressure in patients with septic shock: effects on oxygen variables
and renal function. Crit Care Med 33(4):780Y786, 2005.
18. Higgins J, Green S, editors. Cochrane Handbook for Systematic Reviews of
Interventions Version 510 (updated March 2011).
19. EPOC EPaOoC: EPOC Resources for review authors Oslo: Norwegian
Knowledge Centre for the Health Services. 2013. Available from: http://
epocoslo.cochrane.org/epoc-specific-resources-review-authors. Accessed Jan 2, 2015
20. Guyatt GH, Oxman AD, Vist GE, Kunz R, Falck-Ytter Y, Alonso-Coello P,
Schünemann HJ; GRADE Working Group: GRADE: an emerging consensus on
rating quality of evidence and strength of recommendations. BMJ 336(7650):
924Y926, 2008.
21. Suk P, Hruda J, Leverve X, Sramek V: Early resuscitation of septic shock
to different levels of arterial blood pressure [In Czech]. Anesteziologie a
Intenzivni Medicina 18(3):150Y156, 2007.
22. Asfar P, Meziani F, Hamel JF, Grelon F, Megarbane B, Anguel N, Mira JP,
Dequin PF, Gergaud S, Weiss N, et al.: High versus low blood-pressure target
in patients with septic shock. N Engl J Med 370(17):1583Y1593, 2014.
23. Dünser MW, Ruokonen E, Pettilä V, Ulmer H, Torgersen C, Schmittinger CA,
Jakob S, Takala J: Association of arterial blood pressure and vasopressor load
with septic shock mortality: a post hoc analysis of a multicenter trial. Crit Care
13(6):R181, 2009.
24. Janssen van Doorn K, Verbrugghe W, Wouters K, Jansens H, Jorens PG: The
duration of hypotension determines the evolution of bacteremia-induced acute
kidney injury in the intensive care unit. PLoS One 9(12):e114312, 2014.
25. Subramanian S, Yilmaz M, Rehman A, Hubmayr RD, Afessa B, Gajic O:
Liberal vs. conservative vasopressor use to maintain mean arterial blood pres-
sure during resuscitation of septic shock: an observational study. Intensive Care
Med 34(1):157Y162, 2008.
26. Loeb HS, Winslow EB, Rahimtoola SH, Rosen KM, Gunnar RM: Acute
hemodynamic effects of dopamine in patients with shock. Circulation
44(2):163Y173, 1971.
27. Nouira S, Dhainaut JF, Brunet F, Armaganidis A, Giraud T, Garrauste MT, Corsia
G, Schremmer B, Monsallier JF: [Septic shock: hemodynamic effects of nor-
adrenaline and a noradrenaline-dopexamine combination]. Annales francaises
d_anesthesie et de reanimation 8(Suppl):R234, 1989.
28. Redl-Wenzl EM, Armbruster C, Edelmann G, Fischl E, Kolacny M, Wechsler-
Fordos A, Sporn P: The effects of norepinephrine on hemodynamics and renal
function in severe septic shock states. Intensive Care Med 19(3):151Y154, 1993.
29. Winslow EJ, Loeb HS, Rahimtoola SH, Kamath S, Gunnar RM: Hemodynamic
studies and results of therapy in 50 patients with bacteremic shock. Am J Med
54(4):421Y432, 1973.
30. Day NP, Phu NH, Bethell DP, Mai NT, Chau TT, Hien TT, White NJ: The
effects of dopamine and adrenaline infusions on acidYbase balance and sys-
temic haemodynamics in severe infection. Lancet 348(9022):219Y223, 1996.
31. Hall LG, Oyen LJ, Taner CB, Cullinane DC, Baird TK, Cha SS, Sawyer MD, et al.:
Fixed-dose vasopressin compared with titrated dopamine and norepinephrine as
SHOCK JUNE 2015
initial vasopressor therapy for septic shock. Pharmacotherapy 24(8):1002Y1012,
2004.
32. Mayeur N, Vallée F, De Soyres O, Mebazaa A, Salem R, Fourcade O, Minville V,
Genestal M, Dopexamine Test in septic shock with hyperlactatemia. Annales
francaises d_anesthesie et de reanimation 29(11):759Y764, 2010.
33. Meier-Hellmann A, Bredle DL, Specht M, Hannemann L, Reinhart K:
Dopexamine increases splanchnic blood flow but decreases gastric mucosal pH
in severe septic patients treated with dobutamine. Crit Care Med 27(10):
2166Y2171, 1999.
34. Meier-Hellmann A, Bredle DL, Specht M, Spies C, Hannemann L, Reinhart K:
The effects of low-dose dopamine on splanchnic blood flow and oxygen uptake
in patients with septic shock. Intensive Care Med 23(1):31Y37, 1997.
35. Moran JL, O_Fathartaigh MS, Peisach AR, Chapman MJ, Leppard P: Epi-
nephrine as an inotropic agent in septic shock: a dose-profile analysis. Crit Care
Med 21(1):70Y77, 1993.
36. Obritsch MD, Jung R, Fish DN, MacLaren R: Effects of continuous vasopressin
infusion in patients with septic shock. Ann Pharmacother 38(7Y8):1117Y1122, 2004.
37. Persichini R, Silva S, Teboul JL, Jozwiak M, Chemla D, Richard C, Monnet X:
Effects of norepinephrine on mean systemic pressure and venous return in hu-
man septic shock. Crit Care Med 40(12):3146Y3153, 2012.
38. Fukuoka T, Nishimura M, Imanaka H, Taenaka N, Yoshiya I, Takezawa J:
Effects of norepinephrine on renal function in septic patients with normal and
elevated serum lactate levels. Crit Care Med 17(11):1104Y1107, 1989.
39. Loeb HS, Cruz A, Teng CY, Boswell J, Tobin JR Jr. Gunmar RM:
Haemodynamic studies in shock associated with infection. Br Heart J 29(6):
883Y894, 1967.
40. St-Arnaud C, Éthier J, Leclair M, Lamontagne F: Prescribed targets for vaso-
pressor titration in septic shock: a retrospective cohort study. Intensive Care
Med 37(Suppl 1):231, 2011.
41. Naumann CP, Ruetsch YA, Fleckenstein W, Fennema M, Erdmann W, Zäch
GA: pO2-profiles in human muscle tissue as indicator of therapeutical effect in
septic shock patients. Adv Exp Med Biol 317:869Y877, 1992.
42. Xu YJ, Yang Y, Qiu BH: Titrating mean arterial pressure in consideration of
interindividual effect improves microcirculation in patients with septic shock.
Intensive Care Med 38(Suppl 1):S87, 2012.
43. Lamontagne F, Briel M, Duffett M, Fox-Robichaud A, Cook DJ, Guyatt G, Lesur
O, Meade MO: Systematic review of reviews including animal studies addressing
therapeutic interventions for sepsis. Crit Care Med 38(12):2401Y2408, 2010.
44. Lamontagne F, Meade M, Ondiveeran HK, Lesur O, Fox-Robichaud AE: Nitric
oxide donors in sepsis: a systematic review of clinical and in vivo preclinical
data. Shock 30(6):653Y659, 2008.
Copyright © 2015 by the Shock Society. Unauthorized reproduction of this article is prohibited.
OPTIMAL VASOPRESSOR TITRATION 539
45. Bauman ZM, Killu KF, Rech MA, Bernabei-Combs JL, Gassner MY, Coba VE,
Tovbin A, Kunkel PL, Mlynarek ME: Racial differences in vasopressor re-
quirements for septic shock. Shock 41(3):188Y192, 2014.
46. De Backer D, Aldecoa C, Njimi H, Vincent JL: Dopamine versus norepi-
nephrine in the treatment of septic shock: a meta-analysis. Crit Care Med
40(3):725Y730, 2012.
47. Guérin JP, Levraut J, Samat-Long C, Leverve X, Grimaud D, Ichai C: Effects
of dopamine and norepinephrine on systemic and hepatosplanchnic hemody-
namics, oxygen exchange, and energy balance in vasoplegic septic patients.
Shock 23(1):18Y24, 2005.
48. Patel GP, Grahe JS, Sperry M, Singla S, Elpern E, Lateef O, Balk RA: Efficacy
and safety of dopamine versus norepinephrine in the management of septic
shock. Shock 33(4):375Y380, 2010.
49. Leone M, Albanèse J, Delmas A, Chaabane W, Garnier F, Martin C:
Terlipressin in catecholamine-resistant septic shock patients. Shock 22(4):
314Y319, 2004.
50. Matok I, Vard A, Efrati O, Rubinshtein M, Vishne T, Leibovitch L, Adam M,
Barzilay Z, Paret G: Terlipressin as rescue therapy for intractable hypotension
due to septic shock in children. Shock 23(4):305Y310, 2005.
51. Russell JA, Walley KR, Singer J, Gordon AC, Hébert PC, Cooper DJ, Holmes
CL, Mehta S, Granton JT, Storms MM, et al.: Vasopressin versus norepineph-
rine infusion in patients with septic shock. N Engl J Med 358(9):877Y887, 2008.
52. Kim WY, Jun JH, Huh JW, Hong SB, Lim CM, Koh Y: Radial to femoral
arterial blood pressure differences in septic shock patients receiving high-dose
norepinephrine therapy. Shock 40(6):527Y531, 2013.
53. Dettmer M, Holthaus CV, Fuller BM: The impact of serial lactate monitoring
on emergency department resuscitation interventions and clinical outcomes in
severe sepsis and septic shock: an observational cohort study. Shock 43(1):
55Y61, 2015.
54. Wacharasint P, Nakada TA, Boyd JH, Russell JA, Walley KR: Normal-range
blood lactate concentration in septic shock is prognostic and predictive. Shock
38(1):4Y10, 2012.
55. Noritomi DT, Vieira ML, Mohovic T, Bastos JF, Cordioli RL, Akamine N,
Fischer CH: Echocardiography for hemodynamic evaluation in the intensive
care unit. Shock 34(Suppl 1):59Y62, 2010.
56. Skarda DE, Mulier KE, Myers DE, Taylor JH, Beilman GJ: Dynamic near-
infrared spectroscopy measurements in patients with severe sepsis. Shock
27(4):348Y353, 2007.
57. Dahlqvist M, Hasibeder WR, Dünser MW: Hemodynamic and perfusion end
points for volemic resuscitation in sepsis. Shock 35(4):434, 2011.