Sterile Water Injections for Pain Relief

Before ethical standards changed, doctors used to occasionally fool patients with placebo injections of sterile saline or water. If my obstetrician had tried to give me sterile water instead of an epidural, I probably would have hit him. But apparently women are getting sterile water injections for childbirth and are telling us they work. What’s going on?

A recent study in Sweden compared sterile water injections to acupuncture for relief of labor pain. It found that sterile water produced significantly greater pain relief and relaxation. It concluded, “Women given sterile water injection experience less labor pain compared to women given acupuncture.”

I’m puzzled, because the study also says “there were no significant differences regarding requirements for additional pain relief after treatment between the 2 groups.” 85% and 90% got nitrous oxide, 40% and 47% got epidurals, and other conventional interventions were also used. It seems to me the conclusion could just as well have been “Women given sterile water injections report less labor pain than women given acupuncture, but require just as much additional pain relief.”

Could this be a case of biased researchers selectively interpreting data in the most favorable light? I read the abstract to my husband (who is not a doctor) and he immediately noticed the apparent contradiction. But Dr. George Lundberg didn’t mention it when he reported on the study for the Medscape Medical Minute. In fact, he inexplicably said “Pain relief was sufficient in both groups to prevent the need for other pain therapy.” That’s not what the article says.

I won’t attempt to dissect the study. I’ll assume that it was properly carried out and that its main conclusion (water injections better than acupuncture) is correct. But I do have questions about what it really tells us.

This experiment was not blinded: patients can tell whether they are getting an injection or acupuncture. I wondered why they didn’t include any placebo controls like sham acupuncture or dry needling. It seems they already assumed that both treatments were more effective than placebo; but I don’t think that has been established.

Perhaps all this study shows is that a placebo injection is more effective than a placebo needling procedure. We already knew that placebos are “effective” in reducing pain, and that there is a hierarchy of placebo effects with injections being higher on the list than practically anything besides sham surgery. If we assume that sterile water injections are a placebo, this study just provides one more confirmation that acupuncture is no more effective than placebo.

Do sterile water injections make sense?

Is the method plausible? Anywhere from 0.1 ml (that’s 1/50 of a teaspoonful) to 0.5 ml of sterile water is injected into or just under the skin, forming a small bleb. How could a small superficial bleb affect deep nerves and affect pain transmission up the spinal cord?

Several hypothetical mechanisms have been proposed:

1. A counterirritant effect. Sterile water injections hurt. This hypothesis fits with the fact that sterile saline is less painful and doesn’t work as well as sterile water (sterile water has an osmotic effect in addition to the distention of tissues by the fluid). If it’s only a counterirritant, there might be better, less invasive ways of providing it.
2. Endorphin production. But we know placebos can produce endorphins too.
3. The gate control theory of pain. But other methods that are supposed to work by the gate control theory, like acupuncture or TENS (transcutaneous electrical nerve stimulation), don’t work as well as sterile water.
4. “It may work like acupuncture.” But there is no good evidence that acupuncture is anything more than an elaborate placebo system. And if it works like acupuncture there would be no need for the water; needling could be used alone.
5. “The needles may be hitting acupuncture points.” But no one has convincingly shown that acupuncture points even exist, and acupuncture needles are inserted deeper below the skin.

Sterile water works fast: it sometimes relieves pain for the very next contraction. My colleague Dr. Atwood is an anesthesiologist who has done a lot of epidurals. Before inserting the spinal needle, he numbs the skin by injecting lidocaine intradermally and subcutaneously. It stings. He tells me he has never noticed any effect on the patient’s perception of pain over the next few contractions. Then when the epidural kicks in, there is a dramatic effect.

So sterile water injections don’t look very plausible. Admittedly, implausible things do sometimes turn out to be true. So what does the evidence really show about the effectiveness of this treatment?

The Evidence

I found quite a few studies, but they are problematical. Most are small studies. Some are from non-English-speaking countries that are statistically less likely to publish negative studies (IranThailand, etc.). Some used intradermal injections, others used subcutaneous injections. One study showed that intradermal injections were more painful than subcutaneous ones. Some injected one site, others injected four. Some repeated the injections when the effect wore off after 2-3 hours. Several studies were from Scandinavia, where sterile water injections are popular among midwives. I couldn’t find any US studies.

Most studies used normal saline as a control, and it may not be an adequate control. Several studies reported severe pain with water injections but little or no pain with saline injections. Subjects might be able to guess which group they were in by the amount of pain. If their guesses were better than chance, you wouldn’t have an adequate placebo control. I couldn’t find any evidence of “exit polls” asking patients which group they thought they were in.

A systematic review concluded that it was effective but it was based only on 4 studies they found suitable for analysis:

Ader et al. compared sterile water to saline. Sterile water worked better but there was no difference in the requirement for pethidine (Demerol).
Trolle et al. compared sterile water to saline and found it twice as effective (89% vs. 45%).
Martensson et al. compared 0.1cc of intradermal water, 0.5 cc of subcutaneous water, and 0.1 cc of subcutaneous saline. The two water groups were equally effective and superior to the saline.
Labreque et al. compared sterile water injections to TENS and to standard care (massage, etc.) Water worked better than the other two, but there was no difference in epidural requests, and fewer women said they would choose it again.

Systematic reviews are limited by the quality of the studies they review. Since the controls in these studies are questionable, I don’t know how much we can trust the conclusions.

What about Non-Obstetric Pain?

What really bothers me is that if these injections really work so well, why aren’t they being used for other types of pain like post-op pain and trauma? I found only 3 studies of other types of pain. A study from Iran showed sterile water was more effective than saline for renal colic. A small study in Norway showed no benefit from either saline or sterile water injections for cervicogenic headache. And there was a whiplash study, but it was not really comparable because it targeted trigger points and the pain relief lasted much longer.

Ideological factors

Sterile water injections seem to appeal mainly to midwives and natural childbirth advocates. They argue that it’s inexpensive, readily available, does not require a prescription, has no risks, is easily administered, and “won’t make the drug companies rich.” And it doesn’t require an obstetrician or an anesthesiologist. One commenter on Medscape said, “Sad that the technique isn’t being given a chance here in the U.S. Probably because it doesn’t fit in with the paternalistic, manipulative, technologically dominated, controlled style of obstetric medicine in this country.”

A bias and an agenda shine through such comments. It’s strange that natural childbirth advocates support sterile water injections, because there’s nothing “natural” about them. Some people seem to think there is merit in women suffering pain in labor rather than accepting interventions. I don’t. I think it’s cruel to persuade women to suffer pain unnecessarily when modern obstetrics can relieve pain safely and effectively.

Some support using any placebo that helps the patients get through the ordeal. Placebos relieve pain and are unlikely to harm women or their babies, but they’re UNETHICAL! We need to look for ways to harness the power of the placebo without lying to patients. One way to do that might be to tell them it’s a placebo and get their informed consent. Studies have shown that knowing you’re getting a placebo doesn’t necessarily destroy the placebo effect. I will always remember a patient in my internship who was given a placebo injection instead of the Demerol she usually got in the ER for her severe headaches. It worked just as well as the Demerol! When she was confronted with the fact that a placebo had relieved her pain, she asked if she could get it the next time she had a headache because it really worked!

I wonder about the wisdom of using counterirritants. Hitting your thumb with a hammer would probably work great as a counterirritant, but I couldn’t recommend it. I suspect simple distraction measures and suggestion might be nearly as effective, and they are not invasive and don’t produce pain or stinging.

Mind you, I’m not claiming that sterile water is not better than placebo; I’m just saying it’s premature to conclude that that it is better, based on currently available evidence that is questionable and has not been shown to be properly blinded.

Posted in: Clinical Trials

Leave a Comment (29) ↓

29 thoughts on “Sterile Water Injections for Pain Relief

  1. Erica says:

    My first child was born with an epidural, the second without.

    With the epidural, I noticed a mild immediate effect (followed by total, blissful numbness after a few minutes); however, I am very confident this was more due to psychological relief at knowing the contractions would be gone soon, and so the next few were easier to endure.

    The second time, without the epidural, there was a similar (albeit accidental) placebo-ish experience. During contractions, I was assured it would “be better” when it actually came time to push. Against all logic, I interpreted this as “it won’t hurt as much.” This expectation helped me get through contractions; I believed this was the worst, but it would end. Of course, actual pushing is NOT less painful — the “better” that I was encouraged to look forward to was meant as “pain while you’re doing something productive” instead of “pain while you’re just lying there”. I was furious, but got over it; they didn’t intentionally deceive me and, oddly enough, it helped a little bit. It CERTAINLY helped more than if I’d been hearing, “It’ll hurt a hell of a lot more soon…” :-)

  2. TsuDhoNimh says:

    Anywhere from 0.1 ml (that’s 1/50 of a teaspoonful) to 0.5 ml of sterile water is injected into or just under the skin, forming a small bleb

    Were they told what was in the syringe? Maybe the patients thought they were getting morphine.

  3. Beowulff says:

    Isn’t this just another case of certain placebos working better than others because of different expectations? I seem to recall studies where placebo capsules outperformed placebo tablets, and both were outperformed by placebo injections. If the strength of the placebo effect indeed depends on the apparent impressiveness of the placebo, as seems likely to me, it doesn’t surprise me at all that water injections outperform acupuncture.

    Unfortunately, I couldn’t find any publicly available info on the studies I’m referring to with a couple of Google searches, so if someone knows which ones I’m talking about and can provide a link or two, I’d appreciate it. Of course, I did find sites explaining how we should try to “harness the power of the placebo effect”…

  4. quackdoctor says:

    “And there was a whiplash study, but it was not really comparable because it targeted trigger points and the pain relief lasted much longer.”

    I was not aware that there was any evidence at all that there is such thing as a “trigger point”.

  5. daedalus2u says:

    In the context of labor and delivery, an injection is quick and the person providing the injection can leave and let the woman and who ever is assisting her get back to laboring. In the case of acupuncture, who ever is poking the needles in has to hang around doing the needle manipulation.

    I would think that a quick injection is less intrusive and so is less anxiety provoking than someone hanging around. In the context of labor, anything that leads to more anxiety is going to exacerbate pain and other negative effects.

    I see nothing inconsistent with a pure placebo effect for both of them. The difference being in the details of delivery.

    When they say injections of water, I presume they mean distilled water and not saline. My understanding is that distilled water can be quite harsh and painful to inject because of its osmotic pressure. Saline is isotonic and so has the same osmotic pressure so tissues that come in contact with it don’t gain or lose water by diffusion (the way that distilled water causes). That pain might trigger a better placebo effect than would painless isotonic saline.

  6. gocsick says:

    I know this is anecdotal ans speculative so please read accordingly:

    Many midwives will use sterile water injections along with counterpressure during back labor (When the baby is presenting occipito-posterior i.e. face up ) and a lot of patients claim instant relief.

    Labor and delivery is such a wierd time for the body. Hormone levels are all over the map. Women are in very strange mental states. I wouldn’t be surprised at all to see a much larger placebo effect during labor than for other types of pain. There have been several studies on the effect of professional labor coaches (Birth Doulas) and many showed a remarkable decrease in the need for epidurals (some studies as much as 50%, the better studies typically show a reduction rate of 10-20%) and other interventions. Simple coaching can greatly reduce the perception of pain.

    (see for example A Randomized Controlled Trial of Continuous Labor Support for Middle-Class Couples: Effect on Cesarean Delivery Rates Birth 35:2 2008 further studies can be found in the references)

    I do not think this kind of coaching will work for say kidney stone pain. It is probably important to take the unique circumstances of birthing into account when considering the efficacy of a particular treatment.

    However this does not address the ethical dillema?

    It would be interesting to see if the same effect was seen if the doctor said “I am going to give you an injection of sterile water, we don’t know why it works, it really shouldn’t work, but many women find instant pain relief” Would that address the ethical concerns of a placebo based treatment?

  7. Jules says:

    I would venture to guess that if you’re giving birth, you’re probably producing endorphins up the wazoo anyway…if it’s this “effective” on childbirth pains, shouldn’t it be just as effective on, say, sports injuries?

    That’s what’s got me curious–why haven’t they suggested this for other types of pain? Why only childbirth?

    And my other question: how can a physician who’s trained in accupuncture ethically administer a “false” needling? I’m not being rhetorical–this really puzzles me. It seems to me that sticking needles where they shouldn’t go would hurt. So as long as we’re discussing placebos and controls, would someone mind telling me if this isn’t a contradiction in terms?

  8. quackdoctor says:

    “Mind you, I’m not claiming that sterile water is not better than placebo;”

    I fail to see how someone would attempt to prove something with no plausible mechanism aty all. Would one attempt to prove that the moos is made of green cheese?

    Obviously it is a placebo. Better than some and worse than others.

    Look inject the woman with the water and tell her that it will make her contractions harder and probebly increase her pain and see how much your water removes pain then.

  9. vinny says:

    rj, if you read this comment, I wanted to ask you if you were aware of any side effects from using silver utensils. I understand you developed argyria from allergy drops, but was wondering what danger there is from using a silver spoon?

  10. Mark Crislip says:

    This study, and I have only read Dr. Halls summary, is more evidence for my crank interpretation that there is no real placebo effect.

    Both groups had the same pain medication requirements, which to my way of thinking, says they had the same amount of pain. The objective amount of pain is the same so I would say nothing happened

    Perception is changed, like thinking 70 degrees is cold rather than hot as you just arrived in Portland from Palm Springs.

    Like in the thoracotomy study, one would think that any endogenous pain control system would be maxed out during child birth and it would make no physiologic sense to have any pain control augmented by placebo.

    Also, as best as I can figure out, it makes no evolutionary sense to have developed a pain control system that is only optimized when one is are lead to think something is being done to relieve pain even if nothing is.

    If you are an Ocams kind of guy, then no real placebo effect gives an explanation with the least suppositions.

  11. quackdoctor says:

    The placebo effect may not work through a pain control mechanism. Regardless of the mechanism if the net effect is that the pain is inturpreted as less. Or the pain threshold has changed. Then it is the same thing. The patient percieves less pain. It would however be interesting to see if a patient is under placebo or hypnosis if when they report less pain this diminution in perception is accompanied by a reduction in the other physical changes that may accompany pain.

  12. Harriet Hall says:


    How do you interpret the studies that show that patients report less pain with a placebo but do not report less pain if they are also given an opioid blocker without their knowledge?

  13. Mark Crislip says:

    title, journal, date?

    my brain is too small to remember anything but the lyrics to 70’s pop songs without a direct reference.

  14. Vinny, there is a very large body of med. lit. on silver and argyria and I’ve read almost all of it in English. There was an entire book written by 2 MDs in 1938 called Argyria which is still the definitive work. I have never heard of a case in which argyria has been attributed to eating with a silver utensil.

    I have been told by many silver promoters and even read some overviews written by reputable doctors in which they say that the reason aristocrats were called “blue bloods” is that they developed argyria by eating from silver utensils, but even though I’ve asked, i’ve never been given a solid reference to substantiate that aristocrats or royalty were ever, much less frequently, discolored or even that they commonly used silver utentsils.

    Silver promoters go so far as to claim that blue aristorcrats were much healthier than ordinary people because of the silver in their bodies. However, if aristorctats really were healtier and longer lived than others, there are many other plausible explanations such as having more food, and since the time doctors have studies and reported argyria, no one has noticed that argyric people are healthier or sicker than the average person.

    You can reach me off list at

  15. durvit says:

    Gracely RH, Dubner R, Deeter WR, Wolskee PJ. Clinicians’ expectations influence placebo analgesia. Lancet. 1985 Jan 5;1(8419):43.

    Gracely studied patients scheduled for removal of wisdom teeth. He randomised them into three treatment groups:
    *salt water (an inert placebo in these circumstances);
    *fentanyl (opiate painkiller);
    *naloxone (opiate receptor blocker).

    This wasn’t the obvious assessment of the 3 substances – Gracely was evaluating the impact of the beliefs of those who were administering the substances. There were an additional two groups. In all cases, the people adminstering the substances were blinded to which one they were giving to each patient, but:
    *in one group, those administering the treatment were told that it was placebo or naloxone: in effect, they believed that whatever they were giving, it would do nothing at best, or worsen the pain through the blocking action (however – some patients did receive fentanyl).
    *in the other group, those administering the treatment were told that it was either placebo, naloxone, or fentanyl: in effect, they believed that there was chance that they were administering some pain relief.

    Gracely demonstrated that manipulating the beliefs of those administering the substances delivered different outcomes – even though they did not speak about what they thought that they knew to the patients on the receiving end.

    Unsurprisingly for this sort of placebo study, the second group (the ones where those administering the substance knew that there was a chance it would be a painkiller) reported substantially less pain.

    The difference in outcome did not reflect the actual substance that had been administered, nor even what the patients thought that they knew: it did track what those adminstering the substance thought that they knew.

  16. daedalus2u says:

    Mark, it would be very unlikely that physiological pain pathways would be “maxed out” during childbirth. Childbirth was an extremely common occurrence during evolutionary times. Virtually every female became pregnant multiple times and gave birth multiple times. For the population to remain stable, the average woman had to have two children survive and reproduce.

    Pain can evolve as a successful trait only if it has survival benefits. Pain only has survival benefits if the experience of pain modifies the behavior of the organism experiencing it. As a common occurrence where pain is important, childbirth would be an extremely important event to have proper pain modulation of behavior. In the absence of modern delivery practices, the maternal death rate per pregnancy is about 1%.

    If pain during childbirth is important (and it must be if humans evolved to experience pain during childbirth), then it is doing “something”, and what ever “something” it is doing is only useful if there is differential regulation of what ever “something” it is doing.

    Presumably the degree of pain is related to how serious the situation is and how likely the situation is to result in death. If humans evolved to have pain during childbirth with a 1% mortality rate, childbirth under circumstances where the death rate is much less is likely to be a lot more benign than delivery in “the wild”.

    In other words, pain is only useful as a signaling mechanism, and that signaling mechanism is only useful if it remains in the dynamic range where differential behaviors can be induced by differential levels of pain.

  17. Harriet Hall says:


    I thought you said you had read Bausell’s “Snake Oil Science” – among many other studies on placebo, he references

    Bausell’s description in the book is much more extensive than the abstract.

    Another interesting study I found while looking that one up is

    If you have read Bausell and he didn’t change your mind about placebos, I’d like to hear why. I thought his analysis was very convincing, but maybe I’m too eaily convinced. :-)

  18. Erica says:

    @daedalus2u — This logic — If pain during childbirth is important (and it must be if humans evolved to experience pain during childbirth) — is intriguing, but flawed.

    Humans also evolved with an appendix, a spleen, “extras” of organs like lungs and kidneys… Evolution is demonstrably NOT optimization, so it is inaccurate to assume that childbirth pain was somehow selected for. Indeed, the theory I usually hear holds that the larger skulls of homo sapiens are a significant birth obstacle — evolution of big brains had the side effect of more difficult (and more painful) deliveries, while evolution of upright locomotion (with a smaller pelvic girdles) exacerbated the problem.

    Some types, locations, or intensities of pain can indicate that a problem is occurring — but when you’re living without any medical interventions, meaning that problems are vastly more likely to lead to death, the advantage of knowing there’s a complication and the mother is about to die seems very limited.

  19. daedalus2u says:

    Erica, evolution only minimizes non-survival and non-reproduction from all mechanisms simultaneously. If a trait increases death in some circumstances and decreases it in others, evolution will configure the trait to minimize the sum of deaths (in the optimized limit).

    You are exactly correct, a large head at birth does cause significant maternal deaths in the wild. The advantage of a large brain at birth is so large that it outweighs the increased deaths due to cephalopelvic disproportion. In other words, a large brain is such an advantage that brain size at birth has increased faster than has maternal pelvis size. The mechanism for evolution to increase the size of the maternal pelvis is for women with small pelvises to die in childbirth and not pass on their small pelvis genes.

    I don’t dispute that knowing there is a problem doesn’t necessarily result in a solution. Extreme pain in childbirth might occur to influence other women in the birthing woman’s group. Extreme pain might elicit sufficient sympathy such that a lactating woman might foster the newborn infant if the infant’s mother dies. Even if that occurred rarely, that could have an effect over evolutionary time.

    I am not saying that eliciting sympathy is the “reason” for pain during childbirth. It could be one of many reasons. Severe pain during a first birth might program a woman’s physiology so that subsequent births will be easier. There are dozens of things it might do under different circumstances depending on the idiosyncratic details of that woman’s physiology and the circumstances of the specific birth she is experiencing.

    The only reason that women are capable of feeling pain during childbirth is because they have neural structures that support the generation of signals that other neural structures interpret as pain. Making the assumption that such pain signals serve no purpose, or that such signals are saturated is completely unwarranted.

    As to Mark’s question why pain during childbirth would be susceptible to a placebo effect, if a lactating female did agree to foster a birthing woman’s infant should she die, that may produce such an obvious placebo effect that the foster mother’s willingness to follow through on her promise might be greatly reinforced. Fostering may not be limited to the infant being born. If the birthing woman had other small children, who is going to care for them? Perhaps the woman who’s promise to her dying sister gave her sister such relief.

    There are potentially many different reasons why pain during birthing and why a placebo effect on that pain might have survival advantages to humans living in “the wild”. Such things cannot be studied now because replicating the conditions that resulted in a 1% maternal death rate is unacceptable.

  20. wertys says:


    You might find this of interest to bring you up to date with current myofascial pain research, ie the scientific use of the much-abused term ‘trigger points’.
    Arch Phys Med Rehabil. 2008 Jan;89(1):16-23.

    The placebo analgesia effect is mediated to a large degree through attentional mechanisms in the frontal lobe. A functional MRI study has illustrated that diversion of attention deliberately from the bodily pain experienced in the experiment causes activation of the periaqueductal grey matter, a crucial area in pain suppression. Benedetti and colleagues in Italy further illustrated the importance of the frontal descending projections in a study in Pain in 2006 (I think) which showed that patients with dementia specifically affecting the frontal lobes had a diminished degree of placebo analgesia proportionate to the degree of frontal disconnection.

    A presentation I saw of unpublished research last year suggested that individual subjects differ in the degree to which they can activate these descending projections, which would explain why the amount of placebo analgesia varies considerably from one person to the next.

    With regard to the topic of this post I think the counterirritation explanation is the only plausible one, and I agree with Harriett that there have got to be better and less heartless ways to bring this effect about…

  21. Mark Crislip says:

    My reponse to Dr. Halls question

    By the time I get my thoughts in order and find time to get them committed to electrons, the post has long passed me by. Its why I do not participate much in the comment sections. But better never than late.

    I am hesitant to put a lot of credit into small studies, never repeated, by proponents of a therapy that make little biologic sense and have effects that seem of marginal clinical significance. No I not talking about homeopathy or acupuncture. I am talking about studies of the placebo effect.

    I really liked Snake Oil Science. It should be required reading in every Med School. One of the major points of his book is that studies should be large, multicenter, repeated and done by people with no dogs in the fight. Then he goes on to cite three studies that fail to meet all the criteria above in defense of a placebo effect.

    Sorry. Not impressed with the data. More later as to Dr. Halls specific study.

    I will admit that most of my readings in the placebo effect are in relation to quackery, so my mastery of the information is not on par as, say, my understanding of Infectious Diseases or quackery in general. So I expect to be schooled by them what know better. I know that as a reader of blogs that the ethos of a blogger is one of perfect knowledge :), but I find I learn most when I stick my neck out and look like an ignorant fool.

    There are, in the context of placebo effects, three kinds of studies with pain.

    The first is comparing an intervention to placebo with acute pain to look for a clinical effect.

    The second is comparing an intervention to placebo to chronic pain to look for a clinical effect.

    The third are studies comparing interventions that look to shed light on the mechanism of the placebo effect.

    In considering the first two, when these studies have an objective component, like pain medication use in a study comparing water injection to acupuncture, while the subjective component says the pain is less, by objective criteria it is not.

    Them will argue that the pain is therefor decreased, and it is. But in trying to understand what is going on with pain control and why the pain is less, it is not because less pain is getting to the brain.

    In most studies I have seen where there is a functional measurement of pain, there is no improvement with placebo. So there is, at one level of understanding, no change in the pain.

    If pain were a swimming pool, and I jump in right after a round of golf, it would seem warm and refreshing. If I get in the hot tub first, then the pool is cold cold cold. But in both cases, the temperature of the water is unchanged. Placebo does not change the temperature of the water, it just makes me think it is colder.

    Picky picky, one might say. But the placebo effect doesn’t add to my understanding of the response to disease nor is it of any clinical utility, as in the first two kinds of studies the effect, if any, is marginal. And at the end of the day I am a doctor who needs to use these tools to help sick people. Sorry, I am a tool of the medical industrial complex who wants only to make scads of money at the expense of the ill. I have to keep my story straight.

    I like to understand why things are the way they are. In Infectious Diseases, and a lot of internal medicine, I have that understanding. When I kill a staphylococcus I know how in boring detail how it dies and the physiologic responses to the infection. I know the evolutionary history and the correlates in other animals of, say, the fever response.

    I get none of this with the placebo effect.

    Pain exists, as best I can tell, to prevent damage. Those who cannot feel, like diabetics, lepers, paraplegics and neocons, end up destroying the body (or the body politic). Most of the time you hurt to prevent injury.

    The endogenous pain control system is, at best, a weak system. For the pain I inflict, treat and have experienced, I can say that most of the time the endogenous pain control systems are marginal and probably maxed out.

    And I would expect this. In the old days, when we were hunter gatherers, I doubt we evolved a pain control system to deal with a leg fracture or being gored by a stegosaurus (I saw the Flinstsones, so I know man and dinosaurs coexisted). I would bet, and I really expect to get smacked down for this, that acute pain control exists so that when we had a fight or flight response, we could respond without being distracted or slowed down by minor trauma. Certainly you see that clinically in patients who have acute trauma like a car accident and do not notice the pain until several hours later. The same can happen to solders in battle.

    So the pain control system is minor and ineffective for major pain and with good reason. And in all clinical situations it is probably maxed out. (I will hazard a guess here to bet the daedalus2u has no children). Which is why when functional status is measured in pain, there is no change.

    But why oh why would we have evolved a system that only works best when we are lied to. I understand how the toll system evolved and the complement system and the febrile response. I understand how flagellum evolved and the eye. Why the placebo effect? I have yet to hear even a good just so story. And if it were real, I would expect it to have coorelates in apes at least.

    I bet the placebo effect has nothing to do with the pain control system, but some epiphenomena of another brain system like the belief or other altered states of mind. Anyone know a neurologist with an interest in evolution who could weigh in? Probably not.

    And this gets us to the studies that look at inflicting pain in an attempt to tease out the placebo effect referred to my Dr. Hall, specifically Pain. 2001 Feb 15;90(3):205-15. Response variability to analgesics: a role for non-specific activation of endogenous opioids.

    Also in Snake Oil Medicine prior to this he discusses an interesting study where researchers burned patients to test a new drug or placebo where in fact both were placebo. An interesting summary of that study in the book is “they remembered the no placebo pain as greater than it was and they remembered the pain of the strong placebo point as less than it actually was” (page 151). Perception changes not reality. But on to the Amanzio study

    First the pain inflicted: the patients increased pressure on a blood pressure cuff every four seconds until it was intolerable.

    Very. Minor. Pain. Try it. Personally I have pumped up a blood pressure cuff. It hurts. A little. Not as much as an inflammatory knee, a broken bone, a chronic c6 disc or a (surgical) disemboweling, all of which I have had. Am I sharing too much? This on face value has little if any similarities to any clinical process. So right up front I am not impressed.

    This study then showed that for very acute, completely reversible pain under controlled situations that placebo probably induces a slight surge of endogenous opioids and allows one to tolerate pain longer. Sort of the equivalent of a sudden fight or flight where you do not want to be distracted by a little owie when you twist your ankle. But I do not see what this model has to do with any clinical situation and I question it applicability to clinical situations.

    Again: one study, small numbers, not repeated, done by a one center: all the hallmarks of a study to be taken with a grain of salt. Interesting and well done, but of what applicability?

    In every clinical (emphasis on clinical) study, while pain is subjectively decreased, when functional/objective endpoints are evaluated there are no changes and since pain exists to prevent increased function from causing damage, no change in function, no change in pain.

    If the focus is the few studies of questionable clinical relevance in chapter 10, then, yeah, there is a placebo effect in those limited, artifical situations. If that is your belief, then that chapter provides a nice confirmation bias.

    When I read the literature from the bias of no placebo effect, then the studies show no improved objective changes validate my confirmation bias that there is no placebo effect.

    It is way past my bedtime and I must sleep. Goodnight.

  22. daedalus2u says:

    Mark, as it happens I do have children, but through circumstances beyond my control am unable to interact with them.

    I appreciate that because pain is subjective, only the person experiencing it can tell how severe it is. People do misreport their pain levels for a variety of reasons, conscious and unconscious. To get drugs, to elicit sympathy, to manipulate parents.

    I don’t doubt that the communication of the expression of pain maxes out. The number of adjectives available to express severity of pain is limited. On a 0 to 10 scale, once you hit 10, you can’t go any higher. Unless you make your scale go to 11 (as in Spinal Tap). When a patient reports pain of 10, are they actually incapable of feeling additional pain? If they were subjected to burning cigarettes, electric shocks or bamboo splints would their pain level stay the same? I appreciate that the experiment can’t be done, but I don’t think so. I think they would experience increased pain because that pain would signal damage that could elicit action to mitigate the damage. That was the sense of “saturation” that I was using. Where additional injury produces no increment in pain. The only plausible circumstance I can imagine where the pain system would “saturate” would be when an additional pain signal cannot produce behavioral changes to mitigate the additional injury the pain is signaling.

    You agree that there is physiology that prevents pain from being disabling or distracting during or in anticipation of life threatening trauma. The example I use is running from a bear. Those are the circumstances where pain might “saturate”. But “saturate” is not the right term. Damage and injury is still being signaled, you can’t run from a bear on a broken leg if you treat the leg as uninjured. Even when a leg is broken with a compound fracture, the pain increases when the leg is moved. If there is any action that increases the pain, the pain is not “saturated”. The pain of a compound fracture may be intolerable and may be more intolerable when the leg is moved, but that indicates the system is not saturated.

    The conceptualization of separating the physiological pathways that mediate pain into a “pain system” and a separate “pain control system” is (I think) unnecessarily complicating and confusing. For me, it is easier to think of it as one system, but an exceedingly complex system. The modulation of pain has to be very closely coupled to the systems that produce pain, for what reasons and to elicit which behaviors depending on the organism’s physiological state. I think it is non-physiologic to arbitrarily assume a separate “pain control system” that is somehow independent.

    Organisms experience pain for good physiological reasons. I appreciate that people don’t like to experience pain; that is exactly why physiology causes pain to be experienced, to get the organism’s attention and to elicit actions to reduce the pain. Until the organism has done everything it can to mitigate the damage the pain is signaling, there is no physiological reason, or evolved reason for the pain to diminish. There are good reasons for the pain to not diminish if there is more that can be done to mitigate the damage.

    Physiology can turn off the pain if there is something more important to do, such as run from a bear. My understanding of the placebo effect is that it is the turning down of pain because the organism is satisfied that it is doing or has done something effective to mitigate the damage and/or prevent more damage from occurring. I see the reduction in pain while running from a bear as quite different than the placebo effect.

  23. Fifi says:

    Dr Crislip – While understanding the placebo effect is probably less interesting and important in terms of general medicine, it has quite a lot of bearing on understanding pain and potential usefulness/interest in terms of chronic pain management. Pain is harder to nail down and pinpoint than a disease – particularly chronic pain, there’s a distinction to be made between passing acute pain and chronic pain). There is no objective system of measurement and even if there was everyone experiences pain differently (we have different levels of sensitivity to pain) and manages it differently psychologically and culturally. (Please excuse me if I’m stating the obvious or things you’ve already considered!) I’d be much more interested in studies that looked at and investigated the neurobiology of placebos than ones that rely upon subjective reporting (not that the patient’s subjectivity and experience isn’t important in a clinical setting, it’s just that in an experimental session the results potentially change each set of test subjects you use since it’s all about subjective reporting). And things like creating competing pain in another area or that shift the focus from the primary pain onto something else will reduce the experience of pain (since changing the focus of attention always does) but that’s not an effect of a placebo but of changed focus of attention. Pain behavior is learned – how much we tolerate, how well we understand the difference between good and bad pain, how we express it or even how much we admit to feeling so these kinds of studies would be heavily influenced by where they’re conducted, the features of the individuals involved and also how they’re physically done. It would be pretty hard to manage the confounding factors when combined with subjective reporting of pain.

  24. durvit says:

    Dr Ben’s Goldacre has a 2-part radio exploration of the placebo. The BBC is currently offering Listen Again for Part 1: Part 2 shall be broadcast next week and they have a programme information page (Goldacre links to it).

  25. Harriet Hall says:

    “But why oh why would we have evolved a system that only works best when we are lied to.”

    Spandrels. Not everything in evolution benefits the organism; some things just got a free ride. I think there are a lot of defects in the way our minds work, but our minds probably couldn’t accomplish all the wonderful things they do without traits that produce those defects as side effects. For instance, if we couldn’t be fooled by optical illusions, our visual systems would be less functional for everyday uses.

  26. daedalus2u says:

    Harriet, you are exactly right, it is a trade-off of type 1 errors for type 2 errors. False positives for false negatives for not being able to decide quickly enough.

    There was a recent blog

    that mentioned a recent PLOS article that showed the placebo effect is stronger in children than in adults.

    In a follow-up blog the author mentioned that when working with preemies, giving them something to suck on that has been sweetened makes them a lot calmer and seems to have a pain relieving effect.

    I made the comment that in the wild, the only sweet thing to suck on is mom, and if your mom is holding you and feels safe enough to nurse, then as far as an infant is concerned, all is right with the world. If there is nothing you can do about a situation, there is no need to divert resources to try and deal with it.

  27. durvit says:

    We need to look for ways to harness the power of the placebo without lying to patients. One way to do that might be to tell them it’s a placebo and get their informed consent. Studies have shown that knowing you’re getting a placebo doesn’t necessarily destroy the placebo effect.

    There was a good discussion of various forms of words that might be used to achieve this in Dr Ben Goldacre’s placebo series on BBC Radio 4, part 2. Professor Walter Brown argued strongly that it was possible to use placebo in an ethical manner and discussed some interesting recent studies that seem to demonstrate that this is possible (I’m linking to this discussion for the links to avoid the filter).

Comments are closed.