Methylene blue in ICU

This post came about after a recent discussion amongst the Broome crew. We were reviewing a case of severe sepsis with refractory shock in which the receiving ICU team suggested we try IV methylene blue in addition to all the typical vasopressors and resuscitation tricks that we have in our small hospital. This was a new concept for many of us and inevitably the question was asked… is this something that we should do routinely? We treat more than our share of severe sepsis and have no access to other haemodynamic supports – so maybe this intervention is something that we should be doing sooner rather than later? But what is the evidence? Is methylene blue a magic bullet in the context of septic shock? What could go wrong? Should we add this to our arsenal of medications?

Before we answer those questions – a quick look at the pathophysiology and where methylene blue fits into the management of vasoplegic shock. We are all familiar with the use of traditional vasopressors such as noradrenaline and adrenaline. These are the workhorses of the ICU and remain first line in the care of patients who need vasopressor support.

In recent years we have started using vasopressin as a fixed dose infusion as a second line therapy for the management of vasoplegia. Typically most ICUs will have a threshold of norad above which vasopressin is added… so far so good. We also use a lot of steroids and judicious fluids if we are not winning with modest doses of traditional vasopressor agents.

So where does methylene blue fit? How does it work? Well – it is all about nitric oxide synthase. In the context of severe sepsis and the accompanying “cytokine storm” a subtype of nitric oxide synthase (iNOS) tends to be induced resulting in increased NO levels and this results in vasodilation, the vessels become less responsive to intrinsic vasopressors [and the ones we add in to fix the mix]. So the basic concept is that by blocking the NOS we will be able to use less vasopressors and hence maintain better tissue perfusion to the organs that matter… that the theory.

It is important to know that we also have other subtypes of NOS such as endothelial NOS. Needless to say – this is not a simple up/down, feedback loop – there is a lot going on at the microvascular level and it is likely that there is a well evolved subtle balance of nitric oxide activity happening. We just do not have the physiological finesse to be fine tuning this process in 2024.

Way back in 2004 Lopez et al ran a phase III trial involving 797 ICU sepsis patients of a novel non-selective nitric oxide synthase agent [ called somewhat unimaginatively: 546C88 – there are so many better acronyms out there!] Now this trial was done 20 years ago and our management of septic shock hass changed significantly… sort of… maybe … well we still just do source control, ABs, vasopressors and fluids… but more carefully nowadays! Long story short – this drug resulted in a higher mortality. Mainly due to cardiovascular deaths and pulmonary hypertension – this makes sense when you consider that NO itself is a treatment for pulmonary hypertension…. so we do know that inhibiting NOS indiscriminately can lead to real harms.

Now we get to methylene blue [Editors note: hydroxycobalamin aka Vitamin B12 has also been used in a similar vein (awesome pun!)]. MB is a more selective nitric oxide synthase inhibitor and in theory should target the excess NO production that is induced in the endotoxin / sepsis environment. Early trials showed that measured values such as NO levels, mean arterial pressure and cardiac index were improved / normalised by the use of MB. However, there was still the side efects of increased pulmonary resistance and problems with gas exchange in the lungs. However, what we really care about is the patient, not the numbers… so we need to know does MB do anything to make our patients better? Can it reduce mortality or get them better quicker?

Randomised controlled trials of methylene blue in severe sepsis [there are also a few looking at post cardiac surgery vasoplegia] are few in number and somewhat small in scale. There are a number of small case series and non-controlled trials. However as of this moment we have 6 RCTs which included a total of 292 patients [until last year it was 5 trials and 200 patients… aka only about 20 people in each arm of the trials on average.]. These are very small trials which are prone the effects of random chance, bias and insufficient power to make meaningful conclusions. All of the trials used MOOs (monitor-oriented outcomes) as their primary outcomes – none used a patient-oriented outcome eg. mortality or functional outcomes.

It is probably also worth mentioning that although these were placebo controlled… the fact that methylene blue turns the patient blue and the urine green makes me wonder if a significant unblinding bias occurs in the treating teams? This effect is quite dramatic. It would require an inert placebo that also changes skin and urine to match this one – tricky!

In what may be the most telling moment of pause when considering the research into methylene blue a quick PubMed search yields significantly more meta-analyses and systematic reviews than there are actual trials of this intervention. Not sure if this phenomenon has a name? Let me dub it Morgenstern’s Malady? There are almost as many review articles as there are patients in the trials! Anyway – if you were going to pick one – then I would read this one: (free PDF on link)

Alessandro Pruna, Alessandra Bonaccorso, Alessandro Belletti, Stefano Turi, Ambra Licia Di Prima, Filippo D’amico, Alberto Zangrillo, Yuki Kotani, Giovanni Landoni,
Methylene Blue Reduces Mortality in Critically Ill and Perioperative Patients: A Meta-Analysis of Randomized Trials.    Journal of Cardiothoracic and Vascular Anesthesia, Volume 38, Issue 1, 2024,
Pg 268-274

This is a comprehensive review with an unfortunate title. A case of “the large print giveth, yet the small print taketh away”… to quote the great Tom Waits. Can a metanalysis of small trials, which were not powered to look at mortality… in which no individual trial showed a statistically significant decrease in mortality( as a secondary outcome) claim that there is in fact a decrease in mortality. I am not so sure. If we had granular patient level data, then maybe we could pool them all. However, the actual interventions were different in each individual ICU, so it seems fraught to me.

To summarise the mortality data into a neat Forest plot we can look at figure 2. here:

Fig 2 from Pruna et al

The data from the primary outcomes is more convincing:

Fig 3 from Pruna et al

There is a clear signal to higher blood pressure (and subsequent less vasopressor requirements) in the MB groups. Note this analysis includes studies in post-surgical patients, not just sepsis. So that is useful information – MB looks like it may help with the numbers.. and this may (or may not) translate into better patient outcomes if we look at a bigger population.

However, we also need to think about the harms. Recall the 2004 paper with that non-selective NOS inhibitor… it contained 4 times as many patients as all of the trials in this metanalysis and found clear harms. We know that harms are tough to find in clinical trials. Typically an uncommon harm needs 10 times as many participants to find a signal of such.

So with the current data set on MB for sepsis I do no think we can reliably comment on the safety of this drug in the ICU. For example, methylene blue is a potent monoamine oxidase inhibitor – and as such may precipitate serotonin syndrome in patients on other serotinergic agents [ there is no hard data on this… but that is largely because we just have not looked at enough folk yet in a controlled fashion].

Summary: methylene blue is potentially a useful agent that may improve haemodynamics in our vasoplegic septic patients. It may reduce the need for vasopressors, time on pressor support and time in the ICU. We do not know if there are any patient-oriented benefits such as mortality or functional outcomes for these very sick people. We also do not know if there are uncommon but serious harms from the use of methylene blue in septic shock.

Practice: at this point in time I do not see a role for MB in routine ICU care. I think that with such a sick cohort with a relatively common disease process we should be waiting for a large, multi-centre trial to give us the information we need to guide care.

We know that small, single-centre ICU trials when repeated as large multi-centre trials have about a 5% chance of replicating their mortality benefit under the larger microscope. (Ref: Kotani Et al ). So I would bet against MB in the future if it were to be studied across a large population.

Having said that, if I have a patient who is rapidly dying from sepsis and it seems all is los then the balance of risk / harm seems to favour trying this out – altough I do not anticipate any magic bullets. The data form these trials shows a small improvement in the MAP, not a sudden bolt from the blue!

Add a Comment

Your email address will not be published. Required fields are marked *