The gold standard in medicine is the double-blinded randomized placebo-controlled trial (RCT).
This means that the researchers don’t know which group is getting the treatment, the patients themselves don’t know if they are getting treated or not (the placebo group), and the patient cohorts are randomized to be as equal as possible without potential confounders (age, socioeconomic status, gender, race, etc).
Some confounders could be too many sick people in one group, too many older people in one group, too much of one condition represented in one group, and the list goes on.
However there are problems with RCTs.
You will almost never find a study (unless the specific intervention in question) that controls for the diet and lifestyle between groups, and indeed still fewer who control for microbiota content between groups. How about their exposure to sunlight? What amount of sleep they are getting? Hormone levels?
If the study is not specifically looking for this data, then it won’t be controlling for any of these factors (+ 100 more). In order to control for all the confounders that can’t be controlled easily (or wouldn’t be measured anyway), researchers also aim to have huge groups to compare, ideally thousands of patients in the treatment group and the comparison group.
This way, all these “other” confounders magically balance themselves out. But do they? Really? The answer is maybe sometimes but definitely not all the time.
Now we turn to HBOT clinical trials
It is VERY difficult to do an RCT in hyperbaric medicine because of the chambers that patients have to be in for treatment. If one person goes into a chamber and one doesn’t, there will be a different experience between the two groups and will not fit the RCT criteria.
If the HBOT group gets better and the group that didn’t get into the chamber doesn’t get better, was it the treatment itself that helped? Or was being in the chamber somehow psychologically therapeutic?
What most HBOT trials have done, as a result, is put all treatment and placebo groups in the chamber but just do the “active” treatment (or the treatment being studied) in the treatment arm of the study.
Placebo and Sham
For definition's sake, instead of a placebo, we use the word “sham” to describe a placebo treatment when there is a procedure involved. This applies to the “sham” group in an HBOT trial just like it would apply to a patient who had a sham knee surgery.
In this latter example, both patients are brought to the OR. The treatment group gets the surgery. The sham group gets an incision on the skin but no surgery and both then get sutured back up looking like they had surgery. Then the two groups are compared (outcomes, etc.).
This, in theory, is what these HBOT trials are doing. “Treating” one group of participants and just putting the other group in the chamber and pretending to treat them. But it’s just not that simple!
When you’re in an HBOT environment, the experience is one of simulating the pressure you feel under a certain amount of seawater and when you are simulating this pressure, you’ll feel that pressure sensation in your ears.
So to make it a real “sham” treatment, the “sham” group will need to “feel that pressure change” as well so they have the same experience in the chamber.
Otherwise, the sham participant will leave the chamber and, as people do, inevitably talk to the people in the treatment group who are getting the pressure changes for therapeutic investigation. When they find out that their ears aren’t pressurizing, they’ll know right away that they are in the placebo group, proceeding to screw up the whole study.
In order to prevent this from happening, many of the HBOT studies will compare a patient pressurized with 100% oxygen to a treatment depth of 1.5 ATA to 3.0 ATA (or greater) depending on the indication to a “sham treatment” of 1.3 ATA using pressurized 21% (sea level) oxygen.
Sometimes they’ll use 1.1 ATA, 1.2 ATA, or more recently, some studies are now drifting patients back down to 1.1 ATA after quickly at 1.3 ATA to try to make it as shamtastic as possible.
In other studies, they’ll actually use the same pressure as the treatment group but just change the mix of nitrogen and oxygen so the amount of oxygen infused at the same pressure for the sham group equals a milder pressure, usually around 1.3 ATA.
Pressure and Physiology
And what about the direct effects of pressure on physiology? Even small changes in pressure have a direct effect on blood vessels and cell walls, causing shear stress and directly contributing to energy production at the cellular level. The pressure is also likely creating more blood and lymphatic flow and helping with detox as well.
Thus, comparing 1.3 ATA with 21% oxygen to 2.0 ATA with 100% oxygen is not a sham vs. treatment trial. It is instead better described as comparing two doses of hyperbaric therapy because both change physiology.
For studies where they use the same pressure (say, 2.0 ATA) but change the nitrogen/oxygen mixture to make them treatment/sham, the same is true. They are forgetting the massive effects of pressure on circulation! There’s also the increased nitrogen in these kinds of mixtures that can cause more changes in mental status (nitrogen narcosis) and may lead to increased reactive nitrogen species which can also contribute therapeutically.
Two ways to overcome RCT issues in HBOT trials
This issue with RCT in HBOT trials has long been described by many of my colleagues in the field. There are only really two ways to do a real RCT in an HBOT trial. The first is to have patients fully sedated and on ventilators so they have no idea if they are getting treated or not. These patients would also have tubes placed in their ears to equalize pressure since they won’t be able to do it on their own.
This very trial design was implemented in a study of patients with severe TBIs who were admitted to a neuro ICU after the trauma. All patients had standard-of-care treatment including burr holes in their brains to relieve pressure. The patients who received HBOT were sedated, on ventilators, and had ear tubes placed. But of course, they were oblivious to this while it was ongoing.
I would argue that there is probably one other way to do an RCT that doesn’t require intubation and sedation but it would definitely require ear tubes. Both the treatment and the sham group would get the tubes and each of them would no longer feel the treatment pressure, making it equal as long as the sham group also had a “sound” in the chamber that represented the pressure changes happening.
But this study design has never been implemented and the reason is that every study that is approved has to be approved by what is called an Institutional Review Board (IRB). This is a good thing as it prevents from people doing crazy things like starving them for weeks or preventing people from sleeping until they become psychotic (yes, both of these studies were done in the mid-20th century!).
The problem with ear tubes is that there is a small chance that those getting them will have decreased hearing after they are placed. The procedure itself when done on adults can be done without sedation in the office but it’s the possible resulting mild hearing loss that has prevented this as being a viable study design strategy except in a few extenuating circumstances like a severe TBI.
Unfortunately, many of the studies set up as RCTs, especially those interested in whether mild or low-pressure HBOT could be helpful in neurologic injury, used the sham framework above and came out negative…i.e. “everyone got better” and thus. were interpreted as being studies that showed that HBOT did not work instead of seeing for what they really were: HBOT dosing trials!
This was true for several studies on TBI that the military spent millions of dollars running. In all the trials, all groups in the studies got better. The sham groups got better (all receiving pressure and/or getting more oxygen into circulation). The treatment groups also got better. However the difference in improvements was not enough to be “statistically significant” which led the military to completely disregard the use of HBOT for TBI. There are still some major efforts to get Veterans HBOT but these efforts have been hamstrung by the misleading results in this study.
Some HBOT RCT's work
The RCT is not dead when it comes to non-neurologic HBOT indications. There are many RCTs, like for example Radiation Injury or Diabetic Foot ulcers, that compare a sham treatment depth and 21% oxygen to deep pressure HBOT at 2.0 or 2.4 ATA with 100% oxygen and there are massive benefits to the latter compared to the former.
Because of the issues with RCTs, many researchers, with the Israelis at the forefront, first introduced the concept of the crossover designed RCT to HBOT trials. In this type of study, everyone is treated with the intervention in question but just at different points in time. The groups are randomized and matched demographically and for the treatment in question. For example, all research participants have had a stroke between 6 months and 3 years prior to enrollment of the study. The patients are then matched as above.
Once they are matched, they get the same interventions but at different points in time while they go through the same testing in time as well. A good example is the Israeli study on Stroke.
At time 0, everyone gets a SPECT scan of their brain (see chapter blah on monitoring treatment outcomes). Then, the first treatment group (call them A) get 40 HBOT sessions over 2 months. Treatment group A (who also had the SPECT), gets no HBOT.
After two months, both group A and group B get SPECT scans again and then the groups cross over. Group B then gets 40 HBOT treatments.
After group B gets 40 treatments, both group A and group B get SPECT scans again to assess progress. Basically, group B crosses over from being the placebo group to being the treatment group.
In these crossover trials, the Israelis have been very successful showing the effectiveness of HBOT in TBI, Stroke, Fibromyalgia, Alzheimer’s, and for a number of other conditions.
More recently, some studies are now drifting patients back down to 1.1 ATA after quickly at 1.3 ATA to try to make it as shamtastic as possible…i.e. as inert as possible and have shown some great promise, even with some indications that include the central nervous system.
So there you have it. It is quite difficult to do a real randomized placebo-controlled trial using HBOT and the “sham treatments” used in most studies are actually just a lower dose of HBOT and not inert as a sham is supposed to be in clinical trials.
This study design has led to a significant challenge in neurologic-related indications most of all due to the brain and central nervous system being more sensitive to oxygen and pressure. As a result, seeing a difference between the treatment group at 1.5 ATA w/100% oxygen vs. the “sham” group at 1.3 ATA w/21% oxygen is very difficult, especially in very small studies.
And that’s the last limitation to HBOT studies right there: The money. It’s expensive to run HBOT trials and there isn’t a huge amount of incentive to do them. The reason is simple: you can’t patent oxygen or HBOT treatment so there is no drug or intervention at the end of the clinical trial that can go on to be marketed and make billions. The perverse incentives of our medical system in action ladies and gentlemen!