THE PRACTICE OF EMERGENCY MEDICINE/SYSTEMATIC REVIEW SNAPSHOT

Does the Addition of Vasopressin to Catecholamine

Vasopressors Affect Outcomes in Patients With
Distributive Shock?

TAKE-HOME MESSAGE
In patients with distributive (eg, septic) shock, the addition of vasopressin or vasopressin analogues to
catecholamine therapy may decrease the rate of atrial fibrillation and need for renal replacement
therapy but may also increase the risk of digital ischemia.

EBEM Commentators Editor’s Note: This is a clinical

METHODS Steven K. Roumpf, MD synopsis, a regular feature of the
Benton R. Hunter, MD Annals’ Systematic Review Snapshots
Department of Emergency Medicine (SRS) series. The source for this
Indiana University School of Medicine systematic review snapshot is:
DATA SOURCES Indianapolis, IN McIntyre WF, Um KF, Alhazzani
The authors searched MEDLINE, W, et al. Association of
EMBASE, and the Cochrane Jestin N. Carlson, MD, MS, and Alan vasopressin plus catecholamine
Central Register of Controlled Jones, MD, serve as editors of the SRS vasopressors vs catecholamines
Trials from inception through series. alone with atrial fibrillation in
February 25, 2018. Trial registries patients with distributive shock: a
were searched through http:// systematic review and meta-
www.isrctn.com for unpublished or analysis. JAMA. 2018;319:1889-
1900.
ongoing trials. References of eligible
articles were reviewed and experts
were consulted to identify Results
additional trials. Additionally, the
authors hand searched the last 2 Comparison of catecholamines with and without vasopressin on outcomes in distributive
years’ worth of conference shock.
proceedings for the European No. of Studies Relative Risk With
Society of Intensive Care Medicine, Outcome (No. of Participants) Vasopressin (95% CI) I2, %
the Society of Critical Care
AF 13 (1,462) 0.77 (0.67–0.88) 1
Medicine, and the American
1-mo mortality, all studies 17 (2,904) 0.89 (0.82–0.97) 0
Thoracic Society.
1-mo mortality, low risk of bias 2 (1,049) 0.96 (0.84–1.11) 0
RRT 6 (805) 0.74 (0.51–1.08) 70
STUDY SELECTION
Digital ischemia 9 (1,963) 2.38 (1.37–4.12) 0
Studies included randomized trials
comparing vasopressin (or
analogues), with or without CI, Confidence interval; AF, atrial fibrillation; RRT, renal replacement therapy.
catecholamine vasopressors, with
catecholamine vasopressors alone The search identified 1,210 unique types. There were no studies of
in adults with distributive shock. citations, with 23 randomized patients with anaphylaxis or
Distributive shock was defined as trials, totaling 3,088 patients, neurogenic shock. Vasopressin
septic shock, post–cardiovascular included in the review. Twenty- was studied in 13 trials, and terli-
surgery vasoplegia, neurogenic one studies enrolled patients with pressin (8 trials) was the most
shock, or anaphylaxis. Any dose, septic shock, one study enrolled commonly studied analogue. Most
duration, and other cointerventions post–cardiac surgery patients, and trial protocols did not require that
one study included both patient vasopressin be used with other

Volume -, no. - : - 2018 Annals of Emergency Medicine 1

Systematic Review Snapshot

also associated with reduced patients driving the atrial fibrilla-

were allowed. Two reviewers 1-month mortality in the 17 tion results, however, were not
independently screened studies for
studies in which mortality data septic. Hajjar et al4 studied pa-
eligibility and exclusion, with
disagreements resolved through
were available. However, in the tients with vasoplegia after cardiac
discussion. sensitivity analysis limited to trials surgery and found that 82% of
with low risk of bias, there was controls developed atrial fibrilla-
no effect on mortality (Table). tion and carried 75% of the weight
DATA EXTRACTION AND
Overall, rates of renal replacement of the atrial fibrillation analysis.
SYNTHESIS
therapy were similar between Limiting the analysis to only septic
Two reviewers independently
groups. However, the sensitivity patients resulted in a similar point
abstracted data on the intervention
analysis including only trials with estimate in effect size, but the
and outcomes. The primary
outcome was the incidence of atrial low risk of bias found lower rates result was no longer statistically
fibrillation. Planned secondary of renal replacement therapy significant (relative risk 0.76; 95%
outcomes included mortality, need with vasopressin use (relative risk confidence interval 0.55 to 1.05).
for renal replacement therapy, 0.70; 95% confidence interval 0.53
myocardial injury, stroke, to 0.92). Digital ischemia was More important, the apparent
ventricular arrhythmias, and length- more common with vasopressin. effect on mortality found in this
of-stay metrics. Digital ischemia and There were no differences in meta-analysis was driven by small
acute kidney injury were post hoc other secondary outcomes, studies at high risk of bias, including
secondary outcomes. Primary including myocardial injury, some that were published in ab-
study authors were contacted for stroke, ventricular arrhythmias, stract form only. When the mortality
clarifications and outcome data not analysis was restricted to studies
and length-of-stay metrics.
listed in primary reports. Quality
at low risk of bias, there was no
was assessed for individual studies Commentary evidence of benefit.
with the Cochrane Risk of Bias
Tool, and Grading of Despite many comparative studies,
Recommendations Assessment, the ideal vasopressor for patients Outside of one study with high
Development and Evaluation was with shock, including septic shock, risk of bias and markedly outlier
used to define the overall quality of is unknown.1 Vasopressin is an results,5 vasopressin consistently
evidence. Meta-analyses were endogenous peptide that is depleted appeared to decrease the need for
conducted with random effects in in patients with septic shock and renal replacement therapy. This
instances of low clinical likely benefit was countered by a
that can be infused to increase
heterogeneity. The primary
vascular tone and increase blood consistent finding of increased
analyses are reported as relative digital ischemia. Because the pri-
risk with 95% confidence intervals. pressure.2 It has been hypothesized
that vasopressin may decrease mary studies did not provide in-
Sensitivity analyses were
myocardial demand and arrhythmo- formation on the severity of the
performed including only studies at
low risk of bias. genesis in patients with shock by ischemia, which could range from
allowing lower doses of cardiac- transient discoloration to necrosis
stimulating catecholamines.3 This and amputation, the effect of this
vasopressors but allowed the addi- systematic review seeks to define finding on patients is unclear.
tion of catecholamines if needed. whether the addition of vasopressin
The comparator agent was norepi- to catecholamines is superior to Interpretation of the results of
nephrine in 19 of 23 studies. catecholamines alone in distributive this review is complicated by
shock. the clinical heterogeneity of the
Results from 13 studies suggested included studies. Some trials studied
that atrial fibrillation was less Distributive shock may be caused vasopressin; others, terlipressin or
common in patients randomized by anaphylaxis, vasoplegia after other analogues. The control groups
to vasopressin, with the results cardiac surgery, or neurogenic mostly received norepinephrine,
unchanged when only trials with causes but in the emergency but in a few trials vasopressin
low risk of bias were included. department (ED) setting is usually was compared with placebo or
Vasopressin administration was caused by septicemia. Many of the other catecholamines. Titration of

2 Annals of Emergency Medicine Volume -, no. - : - 2018

 Systematic Review Snapshot

vasopressin and comparator agents This review suggests that the addi- of septic shock. Circulation. 1998;98:
187.
was highly variable, as was the tion of vasopressin or analogues to 3. Dunser MW, Hasibeder WR. Sympathetic
addition of other catecholamines to catecholamines may decrease atrial overstimulation during critical illness:
the vasopressin-treated patients. fibrillation and renal replacement adverse effects of adrenergic stress.
Additionally, inclusion and exclu- therapy while increasing digital J Intensive Care Med. 2009;24:293-316.
4. Hajjar LA, Vincent JL, Barbosa Gomes
sion criteria were variable. For ischemia in patients with septic Galas FR, et al. Vasopressin versus
example, cirrhosis or cancer was an shock.6 The effect on mortality is norepinephrine in patients with vasoplegic
inclusion criterion in some trials but unclear. Further work is needed to shock after cardiac surgery: the VANCS
randomized controlled trial. Anesthesiology.
an exclusion criterion in others. It is examine these medications in ED 2017;126:85-93.
thus unclear whether the results patients with distributive shock to 5. Dunser MW, Mayr AH, Ulmer H, et al. Arginine
should be considered widely determine the ideal treatment reg- vasopressin in advanced vasodilatory shock:
a prospective, randomized, controlled study.
generalizable or should be applied imens. Circulation. 2003;107:2313-2319.
differently to different populations. 6. McIntyre WF, Um KF, Alhazzani W, et al.
No studies in patients with anaphy- 1. Gamper G, Havel C, Arrich J, et al. Vasopressors Association of vasopressin plus catecholamine
for hypotensive shock. Cochrane Database vasopressors vs catecholamines alone with
laxis or neurogenic shock were
Syst Rev. 2016;2:CD003709. atrial fibrillation in patients with distributive
identified, and results may not be 2. Bujik SE, Bruining HA. Vasopressin shock: a systematic review and meta-analysis.
generalizable to these populations. deficiency contributes to the vasodilation JAMA. 2018;319:1889-1900.

Volume -, no. - : - 2018 Annals of Emergency Medicine 3