Methylene blue versus vasopressin in sepsis-induced vasoplegia

El Adawi. Ain-Shams J Anaesthesiol 2016;9:319-24. doi:10.4103/1687-7934.189091

Clinical Question

• In adult patients with severe sepsis, does methylene blue compared to vasopressin reduce noradrenaline requirement and increase blood pressure?

Design

• Single centre
• Randomised controlled trial
• “Single blinded” but doesn’t state who or how
• Screening and recruitment method not specified
• Random number tables used for allocation therefore poor concealment
• Powered at 90% with significance defined at 0.05 if 40 patients recruited
  • Primary outcome not define
  • Source of pilot or background date not provided

Setting

• Single academic tertiary Intensive Care Unit in Egypt
• 2010 to 2014

Population

• Inclusion: Within 72 hours of severe sepsis or within 24 hours of septic shock
  • As defined by ACCP/SCCM consensus criteria on sepsis
  • Requiring Noradrenaline 0.2 µg/kg/min to maintain MAP 70-90 mmHg
• Exclusion: Pregnancy, sensitivity to study drugs, G6PD deficiency, under 18 years old, vasospastic diathesis (e.g. Raynaud’s), coronary disease, MAOI administration
• 40 patients enrolled; none lost to follow-up
• Baseline characteristics (Methylene Blue group vs Vasopressin group)
  • Age: 55 vs 59 years
  • SOFA score: 10 vs 10

Intervention

• Methylene Blue
  • Bolus 1 mg/kg
  • Infusion 2 hours later of 0.5 mg/kg/h for 4 hours

Control

• Vasopressin
  • Infusion of 0.02 U/kg/h for 6 hours

Management common to both groups

• Adrenaline and Noradrenaline titrated to achieve MAP 60-80 mmHg
  • Adjusted in steps of 0.02 µg/kg/min every 15 minutes if MAP > 80 mmHg
  • Noradrenaline weaned before Adrenaline
• Fluid resuscitation titrated to cardiac index and pulmonary artery occlusion pressure
• Piperacillin+Tazobactam was first line antibiotic
• Lung protective invasive ventilation if PaO₂< 6.6 kPa (50 mmHg)
• Enteric feeding
• DVT and ulcer prophylaxis

Outcome

• Primary outcome: not specified
  • MAP was higher in the Methylene Blue group (numbers appear wrongly reported in manuscript’s results section but figure and discussion section are in agreement)
    • At 6 hours: 75 ± 5.9 mmHg vs 68 ± 6.8 mmHg
      • Mean difference 7.0 mmHg (95% CI 2.9 to 11.1; P = 0.001)
    • At 24 hours: 74 ± 6.9 mmHg vs 68 ± 6.9 mmHg
      • Mean difference 6.0 mmHg (95% CI 1.5 to 10.4; P = 0.009)
  • Dose of noradrenaline became significantly lower in the Methylene Blue group compared to the Vasopressin group at 6 hours
  • Dose of adrenaline became significantly lower in the Methylene Blue group compared to the Vasopressin group at 2 hours
  • Oxygen extraction ratio was significantly lower in the Methylene Blue group compared to the Vasopressin group at 6 hours
  • No significant difference in length of ICU stay

Authors’ Conclusions

• In sepsis-induced refractory vasoplegia methylene blue may be more effective than vasopressin but further studies are required

Strengths

• Given the findings of VANISH, asking research questions like this is clinically important
• Appropriate to examine this clinical question by a randomised, controlled trial design
• Appropriate statistical tests for the data types presented

Weaknesses

• Single centre (and published in that single centre’s own medical journal)
• Selection bias likely given only 40 patients recruited over 4 years
• Without adequate randomisation sequence concealment, the potential impact on selection bias is huge
  • That is, researchers may have selected certain patients for the trial knowing that they will enter one group or the other
• If ‘single blinded’ means the patient didn’t know but the clinician did, then allocation bias is possible (Cochrane Centre call this performance bias)
  • That is, clinicians may have given non standard care to favour one group over the other
• Unclear primary outcome
  • Although they have stated a power calculation, it is meaningless if we are not told what is the primary outcome
• Outcomes are not patient-centred
• Errors in the manuscript

The Bottom Line

• Although an exciting and relevant clinical question, the methodological flaws in this study potentially introduce far too many biases for the conclusion to be accepted
• Further patient-centred trials are required before I will be reaching for the methylene blue

External Links

• [article] Methylene blue versus vasopressin in sepsis-induced vasoplegia
• [further reading] Methylene Blue by LITFL
• [further reading] Use of methylene blue in sepsis: a systematic review

Metadata

Summary author: Duncan Chambler
Summary date: 8 February 2017
Peer-review editor: Adrian Wong