Background:

Sepsis is life-threatening organ dysfunction because of a dysregulated host response to infection. Disturbed microvascular blood flow is associated with excess mortality and is a potential future target for interventions. This review addresses the evidence for pharmacological manipulation of the microcirculation in sepsis assessed by techniques that evaluate the sublingual microvasculature.

Methods:

Systematic review using a published protocol. Eligibility criteria were studies of septic patients published from January 2000 to February 2018. Interventions were drugs aimed at improving perfusion. Outcome was improvement in microvascular flow using orthogonal polarization spectral, sidestream dark field, or incident dark field imaging (Grades of Recommendation, Assessment, Development, and Evaluation criteria used).

Results:

Two thousand six hundred and six articles were screened and 22 included. (6 randomized controlled trials, 12 interventional, 3 observational, and 1 pilot, n = 572 participants). Multiple measurement techniques were described, including: automated analyses, subjective, and composite scoring systems. Norepinephrine was not found to improve microvascular flow (low-grade evidence, n = 6 studies); except in chronic hypertension (low, n = 1 study). Addition of arginine vasopressin or terlipressin to norepinephrine maintained flow while decreasing norepinephrine requirements (high, n = 2 studies). Neither dobutamine nor glyceryl trinitrate consistently improved flow (low, n = 6 studies). A single study (n = 40 participants) demonstrated improved flow with levosimendan (high). In a risk of bias assessment 16/16 interventional, pilot and observational studies were found to be high risk.

Conclusions:

There is no robust evidence to date that any one agent can reproducibly lead to improved microvascular flow. Furthermore, no study demonstrated outcome benefit of one therapeutic agent over another. Updated consensus guidelines could improve comparable reporting of measurements and reduce bias, to enable meaningful comparisons around the effects of individual pharmacological agents.