Studying Placebo Effects

ResearchBlogging.orgMeasuring placebo effects (often misleadingly referred to as the placebo effect – singular) is a part of standard clinical trial design, because they need to be distinguished from the physiological effects of the treatment under study. Rarely, however, are placebo effects the actual target being measured, but such is the case with a new study published in the most recent edition of the British Medical Journal (BMJ) – Components of placebo effect: randomised controlled trial in patients with irritable bowel syndrome. (Here is a summary if you cannot access the article directly.)

Dr. Ted Kaptchuk studied the response to various placebo treatments in 262 adults with irritable bowel syndrome (IBS). The three groups were designed to address three major categories of placebo effects: 1) response to the process of being assessed and observed, 2) response to being given a placebo treatment, and 3) response to the patient-practitioner relationship. These types of placebo effects were represented by three treatment arms: 1) observation alone, 2) placebo acupuncture, 3) placebo acupuncture plus an “augmented” practitioner-patient relationship – with added “warmth, attention, and confidence.”

IBS was chosen as the target symptom because prior studies have shown it is highly amenable to placebo effects – with most studies showing a 40% response. This is consistent with the observation that IBS (while its ultimate causes are still unclear) is exacerbated by psychological stress, and therefore can response symptomatically to stress reduction.

Patients were assessed on several scales: a 7 point improvement scale (with 4.0 being no change), an IBS symptom severity scale, a quality of life scale, and a yes/no assessment of “have you had adequate relief.” After three weeks of treatment “scores on the global improvement scale were 3.8 (SD 1.0) v 4.3 (SD 1.4) v 5.0 (SD 1.3) for waiting list versus “limited” versus “augmented,” respectively (P<0.001 for trend).” What this means is that the observation and placebo treatment groups had essentially no change while the augmented therapy had “slight improvement.” On the symptom severity scale, at three weeks, the waiting list group had about a 30 point drop in symptom severity with almost 30% of patients reporting adequate relief; for the treatment group it was 42 points and 44% and for the augmented group over 80 points and over 60% respectively.

What does all this mean? Well, it does confirm what was already known – there are significant placebo effects from the process of being treated, especially with symptoms that are amenable to psychological factors. This is the first study, however, that tried to separate out different types of placebo effect. This new data suggests that different sources of placebo effect add together to give a greater overall placebo effect.

The authors conclude that “the patient-practitioner relationship is the most robust component.” While I agree with this assessment, it should also be pointed out that the wait list group produced about 30 points improvement on the severity scale and almost 30% of subjects reported adequate relief of symptoms – even though on average there was no significant change.  This is important for understanding placebo effects because it means doing nothing but entering a study will create the appearance of benefit for about 1/3 of subjects.  I also agree that this study shows that these different types of placebo effects add together.

I think this is a very useful study in many ways (possibly a first for a study funded by the National Center for Complimentary and Alternative Medicine – NCCAM). It reinforces the important concept that “the” placebo effect is actually a complex combination of many effects – many of which do not involve a “mind-over-matter” physical response to belief in the treatment but are essentially artifacts of the process of observation and reporting.

The study does also reinforce (as the authors emphasized) that having a positive practitioner-patient interaction does provide incremental placebo benefit, at least in symptoms that are responsive to stress reduction. This is an important factor to keep in mind as managed care and the overall economic stresses of modern medicine are causing, if anything, a trend toward decreased quality time between practitioners and patients.

The study further reinforces an important principle of evidence-based and science-based medicine – that studies need to very carefully and thoughtfully controlled for all sources of placebo effects before they can be used to conclude that the treatment in question has any physiological effect. In this study placebo acupuncture plus a supportive therapist caused “adequate relief” in 62% and 61% of subjects at 3 and 6 weeks respectively. This strongly implies that unless acupuncture studies employ properly blinded sham acupuncture, positive results cannot be used to conclude that acupuncture has any physiological effect.

Understanding the nature and role of placebo effects is critical to clinical trial technology, and therefore to the practice of science-based medicine. I hope to see more studies like this that focus directly on placebo effects.


Kaptchuk, T.J., Kelly, J.M., Conboy, L.A., Davis, R.B., Kerr, C.E., Jacobsen, E.E., Kirsch, I., Schyner, R.N., Nam, B.H., et al, . (2008). Components of placebo effect: randomised controlled trial in patients with irritable bowel syndrome. BMJ, online first(April 7), 1-8.

Posted in: Clinical Trials, Science and Medicine

Leave a Comment (27) ↓

27 thoughts on “Studying Placebo Effects

  1. qetzal says:

    Fascinating study and post, Dr. Novella. Thanks very much for highlighting it.

    I hope you will forgive some minor disagreement with your interpretations, however.

    From my reading of the paper, the 3.8 score on the global improvement scale does not mean that the waiting list (i.e. observation) group improved 3.8 points, versus 4.3 and 5.0 points for the other two groups. From the paper:

    Following validated procedures in research on irritable bowel syndrome, our a priori primary outcome was a change from baseline at three weeks in the global improvement scale, which asks participants, “Compared to the way you felt before you entered the study, have your IBS symptoms over the past 7 days been: (1)=substantially worse, (2)=moderately worse, (3)=slightly worse, (4)=no change, (5)=slightly improved, (6)=moderately improved, or (7)=substantially improved.”21 22 Our other main outcome was adequate relief, which is a single dichotomous categorisation that asks participants “Over the past week have you had adequate relief of your IBS symptoms?”23 24 Neither of these primary outcomes were measured at baseline.

    So a 3.8 score essentially means that on average, patients in the observation group reported very slight worsening of symptoms. In contrast, the limited and augmented treatment groups reported very slight or slight improvement, respectively. More importantly, it’s not correct to state that the observation group “had a 3.8 point improvement” and that the additional effects contributed only an additional 0.5 and 0.7 points.

    The authors did include a symptom severity questionnaire as a secondary endpoint, which I think can be used to estimate the relative change in each group, as you were attempting to do. Those scores do seem in line with the authors’ claims that the patient-practitioner relationship “is the most robust component.” However, I agree with you that there is still room to dispute the relative contributions of sham treatment versus relationship.

    What I found most interesting was this passage:

    The magnitude of non-specific effects in the augmented arm is not only statistically significant but also clearly clinically significant in the management of irritable bowel syndrome. A decrease in the symptom severity score of 50 reliably indicates improvement in symptoms,9 and our study indicates that 61% and 59% of patients in the augmented arm achieved this level of improvement at three and six weeks, respectively. Likewise, the changes we observed in quality of life indicate at least moderate clinical improvement in symptoms.25 Finally, the percentage of patients reporting adequate relief (62% and 61% at three and six weeks, respectively) is comparable with the responder rate in clinical trials of drugs currently used in the treatment of irritable bowel syndrome.26 27

    In other words, with proper application of ‘placebo’ effects, one can achieve improvements in IBS that are in line with current efficacy standards for IBS drugs.

    That doesn’t necessarily mean that IBS drugs don’t really work, except as placebos. However, it does reinforce the importance of well designed, double-blind, placebo controlled trials, so that any pharmacological effects of a drug can be distinguished from ‘placebo’ effects.

    It also highlights how CAM practioners can have substantial effects on patients’ perceptions of some illnesses, simply by being caring and confident and performing some action that the patient perceives as a treatment.

  2. daedalus2u says:

    This is a very interesting study. You are completely correct, it does show the severe problem in unblinded trials, or even in trials where the placebo effects of the chosen placebo are not as good as the placebo effects of the treatment arm.

    Since they used actual practitioners of acupuncture to do the placebo acupuncture, I would anticipate that when those practitioners are actually doing “real” acupuncture, the patients would be able to pick up on their confidence (however misplaced) that they are doing something positive and the acupuncture part of this trial may achieve better results than the placebo legs.

    Maybe we need to coin a new term, “placeboey”, (along the lines of “truthy” (something that feels true regardless of whether it is or not)) to denote the proper bedside manner that invokes placebo effects in patients.

  3. Harriet Hall says:

    The obvious next step is to look at the effect of an “augmented relationship” alone. I wonder why they didn’t include that in the study.

  4. qetzal – I read the study again and you are correct. I confused the improvement scale with the severity scale. I am used to using improvement scales that use “0” as no change (which is more intuitive). I will make the appropriate changes to the post.

  5. David Gorski says:

    The obvious next step is to look at the effect of an “augmented relationship” alone. I wonder why they didn’t include that in the study.

    I agree, but they may have something like that in there. Clearly they plan to test “real” acupuncture, as evidenced by these passages in the methods section:

    Although the trial was prospectively designed to investigate non-specific effects in irritable bowel syndrome, its design included a nested acupuncture substudy that allowed potential participants in the “treatment” arms to be told, truthfully, that they had a 50% chance of receiving genuine acupuncture during the trial. When the study ended, a letter was sent to all participants explaining the exact purpose of the study and offering them the opportunity to withdraw their original consent to use their data. All study personnel, except the practitioners, were blinded to participant assignment. Blinded registered nurses who were otherwise unconnected to the study conducted assessments.


    At the end of the first three week period, participants in groups 2 (limited) and 3 (augmented) were, without their knowledge, randomised a second time in equal numbers either to continue with sham acupuncture or to receive genuine acupuncture. Patient-practitioner relationships for these groups, however, remained the same. (Results of this nested secondary study, comparing acupuncture and sham acupuncture, in the second three week period will be reported elsewhere.) Data from patients in groups 2 and 3 who remained on placebo for the second period, however, again as planned prospectively, are included in this report. Participants in group 1 (waiting list) remained on the list for the second three week period. Results at three weeks provided data for the primary end point; those who remained on placebo for the additional three weeks served to provide observations on non-specific effects over time.

  6. David Gorski says:

    I read the study again and you are correct. I confused the improvement scale with the severity scale.

    That’s what I thought, but qetzal beat me to commenting on it.

  7. pmoran says:

    Most interesting, although I think the results (but not necessarily their magnitude) could be predicted from everything else that we have learnt about placebo and the effects of lack of blinding on the results of RCTs (randomised controled trials). There is, of course, no possible way of blinding patients as to whether they are in the waiting list, normal placebo and enhanced placebo arms and it has been consistently observed that non-blinded studies of subjective responses to inert treatments tend to give positive results even when properly blinded ones do not.

    So, as the authors admit, it is difficult to say how much of the apparent patient benefit is due to distorted patient reporting (patients trying to give the nice doctors the “right” answer) or due to “true” placebo phenomena prompted by the belief that they are receving an effective treatment.

    I wish we knew for sure, as it would certainly colour my attitude towards the placebo medicines of CAM. Last time I raised this question Harriet sensibly mentioned studies suggesting endorphin release by placebos, which, if confirmed as a predictable effect of receiving a placebo, would suggest at least some genuine temporary benefits for patients. Brain scans showing certain effects form placebos might also be relevant if we knew for sure what they meant.

    The most interesting finding is the magnitude of the “effects” shown here. They are rather larger than those detected in other studies where placebo treatments have been compared with no treatment, perhaps because the usual RCT hopes to minimise placebo effects, not enhance them.

  8. BlazingDragon says:

    Is there any evidence (other than anecdotal) that a bad doctor-patient relationship negatively affects diseases such as IBS? From my point of view, they take an already bad situation and make it a lot worse. I’d be curious to see if my experience (and that of several other people I know) has ever been studied in a systematic way.

    Doctors should (as a group) do a hell of a lot more to use peer pressure and even sanctions (like some kind of scoring system on their license when they go to get it renewed) to get doctors in line when it comes to bed-side manner. Managed care and cost-containment are a reason why doctors are stressed out and may end up taking it out on their patients, but it’s not an excuse for bad behavior. Too many doctors just don’t give a damn.

  9. daedalus2u says:

    PM, I have my own hypothesis of the physiology behind the placebo effect (which I distinguish from non-physiological effects such as reporting bias by patients to help the “nice” doctor). I see it as the normal allocation of resources between different tasks, primarily ATP, but no doubt there is allocation of all cellular resources.

    Under “stress”, resources are mobilized and kept at the ready to be able to be used at a moments notice, for things such as running from a bear. The only high capacity sources of ATP are mitochondria, and it takes a while for them to “rev up” to produce ATP at maximum rates. That delay is intolerable under some circumstances (as when a bear is eying you and you may need to reach maximum running speed as fast as possible).

    The optimum ATP allocation strategy in those circumstances would be to rev the mitochondria up to produce maximum ATP, and then dissipate that ATP as heat until the precise moment that it is needed. Essentially all of the degenerative diseases do show increased basal metabolic rate (including obesity, ALS, dilative cardiomyopathy, cirrhosis, kidney failure). That increased basal metabolic rate declines when the degenerative condition is improved. I would not expect that actual ATP demand to be increased, rather ATP generation is less efficient because mitochondria are operating at a higher potential where there is more slip and it takes more substrates to produce the same ATP.

    ATP used for healing is not available for “running from a bear”. It takes some time to efficiently shut down those processes. I see this as very much akin to the state of ischemic preconditioning. Where brief periods of ischemia induce a physiological state where ATP demand is lower and so longer periods of ischemia can be survived with less damage.

    Some amount of stress would be expected to turn off long term ATP consuming pathways. Stress reduction would allow those pathways to be turned back on.

    Pathways for healing are extremely complicated and so are necessarily under complicated physiologic control. No doubt there are many ways to disrupt them. The only way to improve them is to allow them to function properly. No amount of external control can possibly be as effective as the normal regulatory pathways that evolved.

    With this view of the placebo effect, all placebos are essentially “the same”, they are “something” that causes physiology to “stand down” from the state induced by high stress so that the normal allocation of resources to healing during low stress can resume. Once that maximal allocation of resources to healing has occurred, there is nothing more that any placebo can do. The healing produced by the placebo effect then reaches the physiological limit.

  10. pmoran says:

    daedalus2u, my knowledge of subcellular physiology and its regulation is now some decades old, so I find it a little hard to comment on the specifics of what you claim.

    One difficulty I have: — I don’t believe that placebos have any significant *objective* effect on most kinds of human illness or “healing”. So it is not clear to me what end points might demonstrate the kind of placebo activity you are postulating.

    The available studies suggest that any “true” placebo benefits are at the level of patient perceptions or focus, or symptom mitigation through other psychological mechanisms or the release of neurotransmitters, or possibly the switching off of illness to the extent that can be a kind of behaviour. It is all bound to be very complex.

    Somewhere down in the middle of an onion of possible influences may be mechanisms such as you suggest but I assume they would have to be mediated by stress acting on the pituitary-adrenal axis. Other hormones e.g. the thyroid could affect the processes you describe, but I know of no evidence that placebos have any effect on them.

    So might your idea only be relevant to some kinds of illness, such as those with a large component of stress? And what physiological or pathological manifestations would you expect from it?.

  11. daedalus2u says:

    PM, There are some studies and reviews that have claimed the only true placebo effects are patient expectations and bias. I think that view is incorrect. There have been demonstrable positive (as well as negative) effects of placebos. The effects may be small, but if they are non-zero, there must be physiological mechanism(s) to produce them.

    There have been instrumental measures of physiological changes. In the blog post I linked to earlier, I discuss studies where they looked at the effect of a placebo (an inert material said to make nausea better) a nocebo (the same inert material said to make the nausea worse), and the same inert material said to be inert. They found that the nocebo reduced nausea both subjectively as reported by the subjects and objectively via gastroelectromyogram. The placebo made it worse compared to the material said to be inert.

    The authors of the above study did suggest that expectations played a role in nausea and that nausea anticipated is nausea lessened. My explanation is that because much of the enteric nervous system is nitrergic, being given a nocebo causes “stress”, which lowers NO levels in anticipation, and that lowered NO reduces the activity of the enteric nervous system reducing both nausea and gastric motility. Their suggestion that patients be told how unpleasant their procedure will be so as to induce negative expectation is (I think) mistaken. I think that negative expectations will only produce beneficial effects for the relatively few conditions that are exacerbated by high NO (such as nausea).

    Whether a placebo is going to improve things or make them worse depends on the physiology behind what is “wrong” in the first place. I think that the placebo effect that I am talking about is more “fundamental” than effects mediated by hormones. Hormones may certainly be involved, but I think they are part of downstream pathways that mediate the specific effects depending on the idiosyncratic physiology involved in specific tissue compartments. Steroids are synthesized (and also metabolized) by the cytochrome P450 enzymes which are regulated by NO (low NO makes them more active). NO is the normal regulatory pathway, so presumably changes in NO levels due to stress that affect those pathways are part of their normal regulation under stress. Steroids are carried around kind of slowly. Acute stress responses need to be regulated by something that signals faster than steroids do. I think that something is NO.

    We know that ischemic preconditioning does reduce ATP consumption acutely and that the ischemic preconditioned state persists for some period of time during which there is reduced ATP consumption. We know that one some of the triggers for ischemic preconditioning require the production of ROS (which causes an acute drop in the NO level). We also know that if cells could stay in an ischemic preconditioned state indefinitely, they would have evolved to do so because there would be more ATP for reproduction. Cells and organisms haven’t evolved to stay in the ischemic preconditioned state, so there must be something incompatible with being in the ischemic preconditioned state and survival and/or reproduction.

    We know that in the ischemic preconditioned state many ATP consuming pathways are reduced. I presume that those are pathways that can be put off for some period of time, but cannot be put off indefinitely.

    During an “ATP crisis”, any physiological pathway that takes longer to produces its product than the duration of the “ATP crisis” can safely be shut down to conserve ATP. Perfect and “ideal” ATP regulation would do exactly that. How close to that “ideal” and perfect ATP regulation are actual organisms? We don’t know the answer to that. We know that some things are shut down during ischemic preconditioning. We don’t know the details, but we know it happens.

    We know that under severe stress healing occurs more slowly. Healing is a very complex process under very complex control. That control is very complex no matter at what rate healing occurs. Slower healing while under stress is also a controlled process. Healing is very important, but it is less important than dealing with things that can kill you faster than not healing. Presumably being ready to respond to “stress” (depending on the severity of the stress) can be more important than healing and is the reason stress slows healing. Presumably something like that is responsible for the slow healing under stress, perhaps the same thing is responsible for the slow healing observed in all the various degenerative diseases. I think that slow healing is just the normal regulation of healing rate, to “optimize” healing rate as a function of substrate availability, and the signal that the body uses to signal that between tissue compartments is NO.

    Something like cancer is more complicated because many growth factors have effects mediated through NO. Low NO might slow tumor growth, but then sometimes differentiation is triggered by high NO, as is apoptosis sometimes. Low NO causes more inflammation and reduces angiogenesis. High NO might increase the growth rate of tumors, but might prevent them from becoming malignant. It might even do different things to different cells in the same tumor. It isn’t something that can be ethically tried in humans, so we likely will never know (unless you want to use all the CAM victims as retrospective “placebo” test subjects).

    In the NEJM paper, the continuous conditions listed in table 2 are all things I would expect to be exacerbated by low NO, so it is not a surprise that placebos had small positive effects. I would expect nausea to be made worse, and don’t know what effect there would be on smoking. Depression should get better, but that may be quite idiosyncratic, and depression certainly isn’t well represented as a binary condition. A trial that would attempt to do so isn’t very sophisticated.

  12. daedalus2u says:

    I would like to back up and use the hypothesis of the placebo effect that I just presented as an example of what I consider an “ordinary” hypothesis, a hypothesis that is completely consistent with everything that is well known. To my way of thinking, a hypothesis that is completely consistent with everything that is well known is a completely acceptable “default” hypothesis; it is something that does not require “extraordinary” evidence to adopt by default. To my way of thinking, it is as good as any other hypothesis that is consistent with all the data (or with as much of the data).

    People do have a default hypothesis of the placebo effect. Many consider that it is only patient expectations and clinician bias. Certainly there is a component of that, but I think the data is inconsistent with that being the only thing that is going on.

    I appreciate that many people are unable to accept this hypothesis of the placebo effect as their default hypothesis. Why that is, and how that happens is not something I understand. I know it is happening; the mechanism by which it is happening is not something I understand. I think it has something to do with how people think using “folk logic”, that is the normal “common sense” that people use which is based on everyday experience and feelings rather than on formal analysis.

    This comment is more specifically directed at the thread where Dr Atwood had the discussion of prior probabilities. To my mind, a hypothesis that is consistent with everything in the literature has a high prior probability. Depending on how much data there is in the literature that prior probability could be very high. According to my analysis I find the prior probability of the above NO hypothesis of the placebo effect to be very high. I appreciate that I am (probably) alone on this thread in coming up with that very high prior probability. I would appreciate a brief description of the thought processes that others are using to come up with a low prior probability.

  13. pmoran says:

    “I think that the placebo effect that I am talking about is more “fundamental” than effects mediated by hormones.”

    This may be part of the problem, at least for me. You seem to be suggesting that in some unspecified but fairly direct way the psychological strimulus of the placebo can affect low-level biochemical processes such as ATP and nitric oxide production.

    Is that so?

  14. daedalus2u says:

    Yes, but I wouldn’t characterize ATP and NO production as “low level biochemical processes”. I would characterize them as more fundamental biochemical processes. Processes that are much more difficult to study because their regulation, control and utilization is much more complex and much more coupled to other parts of physiology than are things such as steroids. There are no techniques to measure either NO or ATP concentrations in vivo at the length and time scales that we know are important.

    I think the characterization of NO and ATP as “low level processes” is unfortunate. I think the idea that they are “low level processes” is a result of the difficulty in measuring them precisely, not because there is any data or compelling theoretical arguments that changes in their levels are not important.

    Direct production of NO by nerves is well known. That is the mechanism for most smooth muscle relaxation. Steroid synthesis is known to be regulated by NO. For regulated effects to occur via steroid synthesis, that synthesis needs to be regulated which requires regulation of NO upstream of that steroid synthesis.

    Steroid physiology uses P450 enzymes, which are highly uncoupled and so make lots of superoxide (some P450 enzymes are nearly 50% uncoupled, that is ~50% of O2 consumed ends up as superoxide). That superoxide is confined to the inside of the microsome the P450 enzyme is in, NO diffuses through the lipid membrane and is destroyed at near diffusion limited kinetics by that superoxide. NO binds to the heme of the P450 enzyme and inhibits it. For estrogen synthesis, the product estrogen activates the estrogen receptor, activates nitric oxide synthase and produces NO which feed-back inhibits the enzyme making estrogen. There are many products and substrates in steroid physiology, many P450 enzymes, so the effects of NO level are not simple (products of one enzyme are substrates of another). Nitric oxide synthase itself is a P450 enzyme.

    NO inhibits cytochrome c oxidase. For high aerobic ATP production, the NO level has to be made very low. One way to do that is via generation of superoxide which destroys NO at near diffusion limited kinetics, even when that superoxide is confined inside vesicles (which it virtually always is). The superoxide generated by mitochondria at the high membrane potential required for a high flux ATP production destroys NO and disinhibits cytochrome c oxidase as required for high O2 consumption at the low O2 concentration needed for high flux (purely passive) O2 diffusion from the blood vessel to the mitochondria down the concentration gradient. For the O2 flux to increase 10x, the concentration gradient has to increase also about 10x. The O2 level at the blood vessel doesn’t change, so the level at the mitochondria has to drop.

    If NO has to be lowered under “stress” to allow for a high aerobic ATP production rate, it would be plausible that over evolutionary time other pathways requiring control during “stress” would evolve to use low NO as a signaling pathway too. Actually it would be implausible that they would not. We know that NO is involved in the regulation of many hundreds of pathways. We don’t know the details of any of them.

  15. daedalus2u says:

    PM, I don’t want you to have the impression that the involvement of low NO in many stress pathways is hypothetical. There are gigantic amounts of data implementing low NO in many of the individual pathways known to be activated during stress. Ischemic preconditioning is triggered in large part by ROS which lowers NO levels. Immune cell activation causes the “respiratory burst” which generates ROS and lowers NO levels. Mast cell degranulation causes the release of proteases and histamine which activate xanthine oxidoreductase to produce superoxide which lowers NO levels and increases mast cell sensitivity. Ischemia or hypoxia cause production of superoxide which lowers NO levels.

    A state of “stress” has to be communicated between cells so that entire tissue compartments can respond to that stress. The “signal” that communicates that state of stress has to diffuse rapidly and through cell membranes. NO is such a molecule, and NO is coupled to the ATP level through sGC (the sensitivity of sGC to NO depends on the ATP concentration).

    There have been a number of reports where topical NO donors/sources have been used to speed healing. The details of how that works are not understood. Healing is a very complex process under very complex physiological regulation. It is not possible for a simple treatment to substitute for that complex feedback control. What a simple treatment can do is shift the “setpoint” of that complex control system. That is what I think higher NO levels are doing, shifting the “setpoint” of the extremely complex control system that is allocating resources to the myriads of tasks each cell is doing in concert with the other cells in that tissue compartment. Shifting the setpoint from less healing to more healing.

    Many of the diseases that are characterized by slow healing are also characterized by low NO including diabetes, hypertension, kidney failure, all the degenerative diseases, obesity, and so on. I think that slow healing is a completely natural side effect of low NO. It is physiology acting completely normally, but controlling physiology to a bad setpoint.

    In a disease state, when a number of pathways are known to be regulated by nitric oxide, and are all observed to deviate in the direction characterized by low nitric oxide, and it is known that nitric oxide is perturbed in the low direction in that disease state; I find it plausible that the deviation is related to NO levels. I find it implausible that NO is unrelated. For NO to be unrelated would require that the normal NO regulation in myriad pathways is replaced by something else in each and every pathway that is otherwise known to be regulated by NO. I find that alternative preposterously implausible.

  16. pmoran says:

    I am trying to follow your argument without having much intimacy with the fields you are talking about.

    By “low level” I was thinking of processes that are completely unconscious and not known to be under neural or psychological control, such as the many metabolic process of the liver. It would be highly inefficient if every biochemical process or system of signallling had to have direct psychological input to function correctly and I think the evidence suggests that most of the functions of the body are NOT under any immediate or tight control by psychological inputs. So, firstly, you need some way of showing how this would work — where does the NO come from and how is its release regulated, especially if it is not a humoral mechanism?

    Your use of the word “healing” is confusing as that word should only be used in technical dicussions in its true medical meaning — the repair of tissue injury. has coopted the word to into meaninglessness, in order to give themselves an aura of niceness and of being empowering to the patient, but mainly to avoid having to explain what their placebo medicines actually do. Who would be churlish enough to want to analyse this wonderful thing?

    When used in this correct sense, I don’t think there is any evidence that placebos accelerate “healing” of any disease or physical proclem. Yet in order to show that placebo activity is not “all in the mind” or at the level of neural processes, you would need some such objective measure of placebo activity, and then some evidence that it is correlated with NO release.

    So I think you are a long way off having a hypothesis that would have general appeal.

  17. daedalus2u says:

    I was using “healing” in the medical sense of natural physiologic repair of damaged tissue, but also to cover the normal “housekeeping” that cells do on an ongoing basis, some of which can be put off for different lengths of time. I was not using it in any alt-med sense but I should have been more specific. I will use wound healing as a specific example because it is an externally measurable quantity. There are similar parameters of cell “health” that are real but not easily measured.

    We know that if an injured individual with a wound is subjected to high stress, the wound will not heal as quickly as if they are subjected to low stress. That slowing of wound healing under conditions of high stress is mediated by the normal physiological control of wound healing. The mechanisms of how stress causes that slow down are not well understood. Presumably that slow down of wound healing under conditions of high stress is due to the normal and (presumably) efficient allocation of resources toward wound healing and toward other things.

    Wound healing requires many (virtually all) of “the same” physiological processes to occur whether wound healing occurs under low stress or under high stress. The only difference is the state of “stress” of the individual. “Stress” is a physiological state brought about by psychological insults, such as noise, threats, fear, worry, perceptions of danger, fear of violence. There must be neurogenic mechanisms that communicate that state of “stress” to the tissues that are undergoing wound healing so as to regulate that wound healing.

    My default assumption is that the process of wound healing under low stress and under high stress is fundamentally “the same” wound healing process. My default assumption is that the “wound healing process” simply uses the state of stress as a control parameter to adjust the rate at which it proceeds. In other words I see “wound healing under high stress” and “wound healing under low stress” as simply two different instances of the same “wound healing process”. Under different physiological states it is controlled differently and so it proceeds at different rates (and perhaps with different out comes). In no way am I attempting to minimize the complexity of wound healing. It involves many thousands or tens of thousands of non-linear coupled pathways working together.

    If we consider the dynamic range of “stress” that it is possible for an individual to experience; from feeling totally safe and protected as in being cradled in their mother’s arms, to being tortured and in great pain and fear of death. Presumably the “wound healing process” active under those conditions is actively regulating the rate of healing of wounds according to whatever optimization scheme the “wound healing process” has evolved to implement.

    The low stress individual has rapid wound healing and the high stress individual has slow wound healing. In normal everyday circumstances, an individual will be somewhere in that spectrum of “stress” between those two extremes. There will be some variation, a bad day at work will increase the stress, a good weekend will lower it. In terms of how that “stress” affects wound healing, it is obviously multifactorial and more complicated than a simple linear progression. We don’t know the conditions that will lead to the “most optimum” wound healing; no doubt they are idiosyncratic and not measurable except by experiments which would not be ethical.

    My conceptualization of a placebo is is something that causes an individual to move from a “higher stress state” toward a “lower stress state”. The magnitude of that shift depends on the efficacy of the placebo. The archetypal placebo is a mother’s “kiss it and make it better”. The mother is communicating to the child that the need to be in the “fight or flight” state has passed, and that the child can now reallocate resources according to being in a different state of “stress”. Healing will then proceed more quickly due to the psychological resolution of the “stress”. I think that is the essence of all placebos.

    I have used the example of “wound healing”, but in a broader sense I mean all the “housekeeping” that cells do which maintains the cell in a viable state and performing the normal metabolic functions of that cell in that tissue compartment (most of which remains unknown). That includes damaged protein disposal, protein synthesis, repair of DNA, transcription, mitochondrial biogenesis, all the things that cells must do to survive and to carry out the functions required for survival of the cell, the tissue and the organism. There are many tens of thousands or hundreds of thousands of things that each cell must do, at many different time scales depending on what needs to be done. Those myriads of things are done based on a priority that the cell has which in part is determined by the state of “stress” of the organism. Some of those can be put off, but there is a cost in terms of decreased function and even of damage. But damage to any particular cell, tissue compartment or organ is tolerable if it leads to survival of the organism. My conceptualization of “stress”, is as danger to the survival of the organism where injury and damage to cells, tissue compartments or organs becomes more tolerable.

    Virtually all organs are highly connected to the nervous system. Presumably those neural connections are for feedback and control of those organs. Organs are totally de-innervated during transplant and continue to function, so the default operation of those organs doesn’t require neural input. We don’t know precisely what those neural connections are doing, but presumably it is something important (or neural connections to organs wouldn’t be so conserved). If they are not important in the “normal” default operation, presumably they are important at other times, perhaps during non-default operation, perhaps under extreme operating conditions where external feedback neural control is needed to coordinate activity with the rest of the organism. When operating margins are much smaller (as during extremes of stress), the open-loop control of a de-innervated organ isn’t good enough.

    Muscle cells in the heart, will continue beating until ATP levels are so depleted that the cells necrose. Presumably organisms evolved that way because a beating heart with dead spots is better than a non-beating heart that remains undamaged. Preventing damage to the heart is useless if the rest of the body receives lethal damage. Presumably all organs have similar operating paradigms. Presumably there is something that communicates to each organ, the state of the rest of the organism, so the organ can “decide” what level of damage is tolerable to inflict upon itself to maintain the organ outputs that the organism needs for survival. Pumping blood is a pretty simple function, the liver has the much more complex biochemical task of gluconeogenesis, degradation and synthesis of many compounds, removal of ammonia.

    Stress is a low NO state. NO is used in many thousands of pathways, and so is already in the active range for each of those pathways. That is, the local concentration of NO is actively regulating each of those specific pathways in physiology. A change in the NO level will then lead to a change in the output of each of those pathways being regulated by NO. All NO sensors only sense the sum of NO from all sources and subtracted by all sinks. The background NO level is necessarily a component of the regulation of every pathway that uses NO as a signaling molecule. If the NO background is low, then all NO mediated pathways are skewed in a low NO direction.

    NO is a freely diffusible molecule. There are no tissue compartments that block the diffusion of NO, except perhaps crystalline bone. Everything else is permeable to NO. Presumably the normal regulation of pathways using NO accommodates as part of its regulatory paradigm the change in basal NO levels as occurs during normal and extreme conditions in physiology. Presumably low NO during stress exerts the appropriate skewing of NO mediated pathways to provide the proper response to stress of those NO mediated pathways.

  18. pmoran says:

    “There must be neurogenic mechanisms that communicate that state of “stress” to the tissues that are undergoing wound healing so as to regulate that wound healing. ”

    But there is not a scrap of evidence that I can think of showing that placebos hasten wound healing. I am a surgeon, This is my field.

    Re stress: Moreover, most placebo responsive states are not the acute, high stress ones that might be associated with slow tissue healing. Should that occur it can be adequately explained by other factiors such as cortisol release.

    Re neural anatomy and phsyiology: It is also not true that all organs are “highly connected” to the nervous sytem or that we don’t have any handle on what the flimsy nervous connections of many tissues and organs do (mostly only sympathetic nervous supply to blood vessels).

    You have a nice theory, but I think if you study human physiology at the macroscopic level more you will start to see the weaknesses in it as any major component of placebo effects.

    There is no question that expectation is the key to whether placebo effects(/reactions/phenomena?) occur, and the shortest distance between that and perceptions regarding levels of pain or other symptoms is unquestionably the neural connections in the brain.

  19. David Gorski says:

    But there is not a scrap of evidence that I can think of showing that placebos hasten wound healing. I am a surgeon, This is my field.

    Nor is there a scrap of evidence that placebos slow down or halt the growth of cancer. Of course, with few exceptions, it’s generally unethical to do placebo-controlled cancer trials. Usually most trials involve comparing new agents with the standard of care or adding new agents to standard chemotherapy; so it’s not as big an issue in cancer research.

  20. daedalus2u says:

    I think some of the difficulty I am having trying to get my point across is that we are using different conceptualizations of what the term “placebo effect” means. I am not trying to skirt the issue. There are “treatments” which I would call “placebos” which are good medical practice. A good bedside manner, using the right tone of voice, emotional connection to the patient, these are not “biologically active treatments”, but the do have effects on outcomes.

    In his earlier post on placebo effects, Dr. Novella said: “The operational definition of a placebo effect is any health effect measured after an intervention that is something other than a physiological response to a biologically active treatment.”

    That definition is problematic because any treatment that does produce a physiological response can (at least in that sense) be said to be “biologically active”. That limits any “placebo effect” (by definition) to only non-physiological responses. The paper I cited earlier where pharmacologically inert pills caused measured increases or decreases in gastric motility could not be about “placebos” (by definition) because the inert pills produced physiological responses, hence they must have been a “biologically active treatment”. If a homeopathic treatment produces a physiological response, then is it a “placebo”? If you restrict the “physiological response” to “greater than placebo”, then the definition of placebo becomes circular and you need to have a known “placebo” which you can measure other treatments against to see if they have greater or lesser physiological responses.

    In Dorland’s medical dictionary, they have: a medical preparation having no specific pharmacological activity against the patient’s illness or complaint given solely for the psychophysiological effects of the treatment.

    Limiting the definition of “placebos” to inactive treatments given only in a therapeutic context inappropriately limit its scope. To try and understand what the “placebo effect” actually is, we need to expand the definition beyond treatments given in a therapeutic context. The physiology behind the placebo effect didn’t evolve in a therapeutic context. We are only going to understand the placebo effect in the broader context where it did evolve.

    Defining “placebo effect” for non-pharmacological interventions is more difficult. Sham surgery can have as large a therapeutic effect as the real surgery. An earlier comment by David Gorski

    mentions a surgical procedure, pericardial poudrage which worked as well as placebo and makes the statement that ” One thing that this study demonstrates is that surgery itself is among the most powerful of placebos.”

    Another surgical technique for ischemic heart that of internal mammary artery ligation worked quite well, with something like a 90% improvement rate. However placebo treatment gave as good or better results.

    Sham laser myocardial revascularization and also sham FGF2 infusion also produce a good therapeutic effect as measured by exercise time (p less than 0.001), angina class (p less than 0.001), quality of life (p less than 0.001). These improvements persisted for 2 years.

    It is pretty clear that the holes blasted in the heart by the laser don’t cause revascularization. The holes quickly fill up with cells and scar tissue. Sham treatments produce pretty much the same improvements.

    My definition of the placebo effect is: positive (placebo) or negative (nocebo) health effects from medically inactive treatments. This simply moves the definition difficulty to what is “medically inactive”. Pharmacologically inactive is relatively straightforward to understand; treatments without an effective surgical component are harder to distinguish.

    For example, dilative cardiomyopathy is a disease that is associated with low NO. It is reliably produced in animal models with simple metabolic overload either through pacing or outlet constriction. It is reliably made worse by stress, and reliably improved by stress reduction. Is “stress reduction” in the context of dilative cardiomyopathy a “placebo”? Or is it improvement due to a medically active treatment? Is psychotherapy a placebo or a medically active treatment?

    Regarding innervation of organs, the usual mechanism by which blood flow is regulated is via release of NO which activates sGC producing cGMP and relaxing smooth muscle. Some of that NO is carried downstream both as NO, and as NO metabolites, nitrite and no doubt other things (S-nitrosothiols for example). Regulating the NO level may be the major mechanism by which there is neurogenic control of organs. In the brain, neural activity is highly correlated with blood flow, so much so that in BOLD fMRI what is measured are the relative levels of O2Hb and Hb. The vasodilation that causes that increased blood flow is due to NO release. In a very real sense, all that BOLD fMRI is measuring are the effects of neurogenic NO release on blood flow. Which is the “cause” of the neural activity? The increased blood flow, or the NO release that preceded the increased blood flow?

  21. pmoran says:

    Daedalus2 >”In his earlier post on placebo effects, Dr. Novella said: “The operational definition of a placebo effect is any health effect measured after an intervention that is something other than a physiological response to a biologically active treatment.”

    I agree with you that this definition has limited application when it comes to analysing what is actually going on in the placebo arm of a controlled trial, or when someone claims spectacular results from a quack remedy that we know has no intrinsic medical activity. Even the “effects” is an unfortunate use of the word, as patient outcomes with many conditions are partly or even largely due to the fact that many conditions fluctuate naturally in severity or get better by themsleves. It is not an “effect” in any real sense at all.

    It also unjustifiably focuses attention upon what the placebo or sham treatment is “doing” whereas the Kaptchuk study supports what we always suspected , that it is everything else about the therapeutic environment — the physician’s “schtick”, if you like — that is the main influence upon patient responses (or reporting).

    I am happy to include any physiological effects of taking a placebo, such as endorphin release, as a component of “placebo-related phenomena”, so we can work on that basis.

    Daedalus2> “For example, dilative cardiomyopathy is a disease that is associated with low NO. It is reliably produced in animal models with simple metabolic overload either through pacing or outlet constriction. It is reliably made worse by stress, and reliably improved by stress reduction. Is “stress reduction” in the context of dilative cardiomyopathy a “placebo”? Or is it improvement due to a medically active treatment? Is psychotherapy a placebo or a medically active treatment? ”

    “Stress” is another word with many meanings and it is not clear what kind of stress you are talking about here — work stress, as in having to pump harder, which will certainly make things worse, or psychological strress. But it doesn’t matter. Even psychological stress will speed up the heart rate and that alone may exascerbate severe dilative cardiomyopathy.

    I don’t think I have anything to add to what I have already been said. The principal problem you have is that placebos don’t seem to do anything very much at all at the physical level, leaving you with a hypothesis that is looking around for something to explain.

  22. daedalus2u says:

    I completely agree that in many cases placebos don’t do much. So long as they don’t do zero in all cases, then there are physiological effect(s) that need to be understood; particularly if we are going to understand the effects of individual placebos, the additive (or subtractive) effects of multiple placebos and how placebos interact with “real” treatments.

    Any understanding of the placebo effect has to explain cases where placebos do something, and cases where placebos do nothing. I think including good bedside manner into “placebo-related phenomena” is technically correct and is completely appropriate. A bad bedside manner is the opposite of a placebo and is an example of a nocebo.

    I completely agree that “stress” has many meanings and there are many physiological responses to the many different types of “stress”. All “stress” responses must have two components, something to turn on the stress response , and something to turn off the stress response . If there is nothing to turn off the stress response then the response becomes permanent and isn’t a “stress” response any more. I see placebo effects as just one of a number of mechanisms that turn off the stress response. That idea is not something that I originated.

    My contribution to this idea is incorporating ATP regulation as an additional fundamental component. ATP and NO are directly coupled because they both affect the sensitivity of sGC to the other. At high ATP it takes a higher NO level to activate sGC. This coupling of NO and ATP then “tunes” NO responses to the ATP status of the cell, and as ATP levels fall, sGC becomes more sensitive and is activated even if NO levels remain constant.

    This hypothesis predicts that once you are fully out of the “fight or flight” state that is diverting resources away from normal cell maintenance and healing (and holding them in reserve for responding to what ever “stress” is anticipated), then placebos will do nothing. The physiological details of that “fight or flight” state are quite complex because it is a very complex coordinated change in the physiology of essentially every organ and tissue compartment. Those changes are different depending on the specifics of the “stress” that invokes the “fight or flight” state. The control system to coordinate those changes must be extremely complex because it has the capacity to invoke every stress-mediated physiological state and “choose” which one to invoke, in what time frame, and when to de-invoke it and return to the basal “at rest” state.

    A difficulty people have with this hypothesis is that it is not consistent with the idea of homeostasis (which happens to be wrong). There are many observations that are inconsistent with homeostasis. The very concept of a physiological stress response is incompatible with physiology keeping things static. Stasis and change are incompatible control paradigms. Cells didn’t evolve physiological pathways to keep things static, they evolved those pathways to survive, and adjustment of any and/or all parameters is within the scope of what can evolve.

    ATP level isn’t kept static, it is used as a control parameter to regulate pathways that generate and consume ATP. When the ATP level is low, only the highest priority pathways are allowed to consume ATP. In the heart, muscle cells will work themselves to death. Presumably “healing” has a lower priority and is turned off well before cells start to die from low ATP. Ischemic preconditioning due to an acute lowering of ATP levels changes the control paradigm of that cell to reduce ATP consumption for a while. That reduction in ATP consumption is what results in survival from ischemic insults later. Obviously, what ever is being turned off during ischemic preconditioning isn’t acutely needed during that ischemic preconditioned period.

    Virtually all of the (extremely many) details of what goes on during the “fight or flight” state remain unknown. We know that many thousands of things happen, in concert, in multiple tissue compartments and that these are primarily “stress responses”. While we don’t know the details of what is happening, we do know how these stress responses arose, they evolved. Stress responses are among the most conserved pathways, and many of them date to deep evolutionary time. The more fundamental the stress response, the deeper in evolutionary time its regulation dates to.

    One of the most fundamental stresses is an ATP crisis, an acute depletion of ATP. Single cells had to learn how to regulate their ATP levels by regulating production and consumption before they could do much else with that ATP. ATP crisis responses fall into two categories, make more ATP, or use less ATP. The first organisms that evolved to deal with an ATP crisis date to pre-eukaryotic time. Every organism today is a descendent of those organisms that first learned to deal with an ATP crisis. Because ATP is such a fundamental need for cells, essentially every other need is subservient to the ATP need. When there isn’t enough ATP, the highest priority pathways are those for immediate survival and pathways to make more ATP. If a pathway isn’t needed for immediate survival, and isn’t involved in supplying more ATP, it can supply ATP by shutting itself off. Any pathway that takes longer to produce its product than the length of the “ATP crisis” can’t contribute anything to survival during that ATP crisis and so can contribute ATP by being shut down. The usual “ATP crisis” is short because compensatory pathways usually restore ATP production and/or reduce consumption. I think the chronic effects of stress occur because pathways are shut down for longer than is “usual”, and not restored before the adverse effects of chronic stress occur.

    The usual mechanism by which physiological pathways evolve is by elaboration of existing pathways (rather than generation of new pathways de novo). Presumably stress responses utilized this mechanism too. Simple and fundamental stress response pathways likely became more elaborate and complicated as they evolved to mitigate additional stresses. Presumably many of those pathways would then retain the same turn on and turn off mechanisms, there being no evolutionary advantage to new mechanisms in that part of the pathway.

    I suspect that the evolution of pathways by elaboration of existing pathways is part of why NO is involved in so many stress pathways. Oxidative stress often occurs due to Fenton active metals (which generate hydroxyl radical from H2O2) and heme containing enzymes, both of which can be neutralized by NO.

    Nitric oxide synthase (NOS) makes both NO and superoxide but under different conditions (not yet fully understood). When NOS is subjected to oxidation by peroxynitrite (formed by the near diffusion limited reaction between NO and superoxide), the Zn thiolate cluster becomes oxidized and NOS then becomes uncoupled and makes only superoxide. Neuronally activated NOS (usually nNOS) can then produce either NO, or superoxide or a mix of the two (peroxynitrite).

    In the “long run”, essentially all pathways in organisms are needed. Pathways that are not needed are not conserved or are shut off permanently. Pathways turned off during ischemic preconditioning need to be turned back on when the ischemic preconditioning is over. If they are not turned back on, then what ever product of those pathways that the organism needs is not supplied and (I think), the organism goes down a slow degenerative death spiral. I think that all disorders that are exacerbated by “stress” are fundamentally of this type, and any form of stress reduction will have positive effects via turning off of the stress response.

  23. garrick says:

    What about placebo on a dog? I’ve seen acupuncture work on my own dog (me being the skeptic, lost that fight with the woman). I have an old arthritic dog that had a rather profound improvement with acupuncture. The effect only lasted a day or so, but she actually walked down the stairs that she couldn’t do for months prior.

    The acupuncture was repeated once a week for 4 weeks and the outcome was repeated all 4 times. Though the last time was harder to see because she was also benefiting from physical therapy (strengthening exercises) at the same time.

    Could a dog have improved for a day by keying into our expectation of therapy? Can a dog have a placebo effect?

Comments are closed.