On Wednesday, Steve Novella published a post discussing a recently-published study investigating the use of acupuncture for infant colic, as did a friend of SBM back on January 19th. As expected, they did a masterful job pointing out weaknesses in the study design and the absurdity of the authors’ conclusion when put into the context of a science-based understanding of acupuncture’s plausibility and research base. The study is unethical and doesn’t actually support claims that acupuncture has a specific effect on the amount of daily infant crying.
But as the pediatrician on the team, I couldn’t resist adding my two cents to the discussion because it involved infant colic. Colic, and infant crying in general, is an extremely common concern and it can result in significant caregiver and family stress. Many parents feel overwhelmed when caring for a new baby and seek out the help of medical professionals when they believe that their baby is crying too much. Although this complaint rarely results in the diagnosis of a serious medical condition, these babies are at higher risk for abuse and neglect as well as excessive medical diagnostic and treatment interventions.
I doubt many days go by where a general pediatric medical professional doesn’t discuss crying with at least one family. And the way we handle these conversations is incredibly important because we can change the way that caregivers perceive the health of their child. Our responses can also negatively impact how parents feel about the medical system in the future.
Back in 2013 I compared postthe science-based medical approach to colic with that of chiropractic. In that post, I go into a bit more detail on possible causes of colic as well as various treatment approaches, so I encourage readers to check it out. In this post, I’ll cover the important highlights along with a few new thoughts before coming back around to the aforementioned study.
What is colic?
All babies cry to varying degrees, usually only intermittently and in response to their environment. And babies are no different than the rest of us in that they have a range of temperaments. There is a normal distribution of crying in healthy babies with some falling on either end of a spectrum, particularly in the first 3-4 months of life when most crying takes place.
Sometimes crying in infants crosses what is essentially an arbitrary threshold of “too much.” That is all colic really is. We have traditionally used the Wessel criteria, or “rule of three”, which states that colic is an appropriate designation when an otherwise-healthy infant cries for more than three hours per day, more than three days a week, and it lasts longer than three weeks. But widespread use as a nebulous descriptor for any young baby who is fussy for unclear reasons has essentially rendered the label meaningless.
There is also a normal distribution of caregiver responses to crying. Some take it in stride while others are much more sensitive. Some see any crying beyond some arbitrary personal cutoff in duration or severity as a sign of pain or illness. Perception is reality and some parents subjectively assess their children as crying both longer and more intensely than a more objective determination would show. Being forced to rely on parental report can result in some infants being given a medical label, which can serve as a self-fulfilling prophecy of sorts.
We often focus on anecdotal descriptions of colic that many believe to be specific to it, such as the belief that episodes of colic start and stop suddenly, with no clear inciting event or environmental variable. Many believe that colic is more common in the evening (“the witching hour“) and is subjectively more intense, urgent, or painful than “normal” crying. Many believe that colic spells don’t respond to soothing as easily as regular crying. All of this may be true, but then again many caregivers and pediatricians also falsely believe that they can predict when a tooth will emerge based on a particular constellation of symptoms.
Ultimately, we don’t really know if colic represents actual pathology or if it is primarily a psychosocial construct. I tend to lean towards the latter and believe that many babies diagnosed with colic have temperaments that are a poor fit with their caregivers’ tolerance levels. But as I’ll discuss shortly, I believe that the subjective nature of the assessment of crying and the symptom overlap with a very short list of possible medical ailments, forces me to leave some cerebral wiggle room.
A very little colic epidemiology
According to the lore passed down from one generation of medical professionals to the next, about one out of every three infants will develop colic in the first few months of life. But we don’t really know if this is true. Depending on criteria used, the incidence is anywhere from 8 to almost 50 percent in various surveys.
Most babies labeled as colicky don’t fit any strict criteria. No matter how you slice it, the determination is entirely based on subjective parental report. Perhaps this is a question that big data and the “internet of things” will be able to eventually answer as baby monitors are fairly ubiquitous these days.
Nothing has been shown to reliably increase the likelihood of developing colic, not even when comparing formula and breast fed babies. Probably the closest we have to a solid association are psychosocial factors such as parental stress and a lack of parental self-confidence. But does stress somehow impact parental interactions with the infant, leading to more crying? Does it alter the intrauterine environment? Or does stress and parental insecurity alter the caregiver’s perception of how much their baby is crying?
What causes colic?
By definition, babies diagnosed with colic cry a lot but are healthy as far as we can tell based on a thorough history and physical exam. But not every possible underlying pathology is obvious and easy to diagnose. It is extremely likely that there is a self-limited but treatable organic etiology in a small subset of these babies that has gone undiagnosed.
Despite the plausibility of some cases of colic having an organic cause, we don’t have a lot of good evidence to support this. What we have, such as it is, breaks potential causes down into three general categories: the gut, biologic, and psychosocial. The gut is by far the category which has garnered the most attention. Again, for a much more detailed discussion of these categories please read my 2013 post on the subject.
Allergy to cow milk protein is probably the most plausible potential disease process leading to a diagnosis of colic. It is fairly common and can result in a persistently fussy baby. Most babies will develop visibly bloody stools, but milder cases might not. I actually agree with the authors of the acupuncture study when they recommend that a specialized formula, or maternal dairy restriction if breastfeeding, should be considered when crying is felt to be excessive.
Other proposed causes involving the gut are fruit juice intolerance (don’t give babies juice), differences in intestinal microflora (isn’t everything these days?), and gas.
Conditions in the biologic category include migraines (not likely), feeding difficulties (almost certainly plays a role in some cases), immaturity (brain and/or gut), and prenatal/postnatal nicotine and/or tobacco smoke exposure (plausible). There are some correlation-versus-causation concerns in this category, but it isn’t hard to imagine some of these playing a role in excessive crying in some infants.
The most important category in my opinion is psychosocial. This includes infant temperament, hypersensitivity to stimuli, and the parental factors I’ve mentioned several times already. I say it is the most important not just because of the causal role I believe it plays in labeling a baby as colicky but also because caregiver stress and their ability to manage it is important regardless of whether or not the child is actually crying a lot.
How is colic managed?
When a caregiver is worried about crying, it is of extreme importance to evaluate for any condition or environment that could put the child at risk of harm. It really doesn’t matter if a caregiver’s description meets any potential diagnostic criteria for infant colic. Even if the baby only cries for 30 minutes a week, it is vital to take any parent asking for help seriously.
To make sure that a baby is safe, and does not suffer from a serious illness or live in a home where caregiver stress is a possible risk factor for neglect or abuse, there are many important questions to ask besides those that cover the nature of the crying itself. It is perhaps equally if not more important to ask how the caregiver, and the family as a whole, feels about the crying and how they have reacted to it. This can serve as an entry into a broader discussion about stress and anxiety and reveal areas of childcare that may be amenable to reassurance and education.
After discussing the problem, a thorough exam is necessary. Are there any concerning physical finding that could be a clue to illness or abuse? Are there any signs of pain that might not be related to possible colic? Again, infant colic is a diagnosis of exclusion.
There is no one-size-fits-all treatment approach to infant colic. But, if after a thorough evaluation and assessment of the need for parental support, a diagnosis of colic is made, then a reasonable approach in many cases may be to simply reassure and educate as needed. Infant colic will stop on its own, often by 3 months of age, and nearly all infants have significantly improved by 4 months. This is why infant colic is such a perfect condition for practitioners of irregular medicine. Crying duration and severity is subjective and colic will get better on its own, so it is easy to assume incorrectly that an intervention has worked.
So what is the appropriate approach when a pediatrician and family decide together that an intervention is necessary? I do not think anyone would describe any of the proposed treatments as well-supported in the pediatric literature, but some have a better evidence base than others. First and foremost is addressing and managing parental stress. This may lead to a new perspective on the crying and remove the desire to intervene medically. Teaching simple soothing techniques can be an instant cure in many cases.
Based on the available evidence, the additional intervention that has the best chance of helping decrease the duration and intensity of crying episodes is initiating a trial of specialized formula containing no intact cow’s milk protein, or having a breastfeeding mother remove dairy from her diet. All other interventions likely only seem to work because of parental placebo and colic’s natural course. These include feeding techniques to reduce swallowed air, probiotics, and the use of pacifiers, baby swings and infant massage.
No medication is both effective and safe for use in the treatment of infant colic and drugs are rarely prescribed. Even the extremely popular OTC drug simethicone (Mylicon) has failed to demonstrate benefit. Other potential treatment modalities that have failed the test of science are numerous herbal remedies, homeopathy, various manipulative techniques such as chiropractic and craniosacral therapy, and acupuncture.
Now back to that study
As I mentioned in the introduction, Steve and Orac did a damn good job dismantling the conclusions made by the authors of this abysmal study. Just in case you didn’t read their posts, I’ll provide a quick review. Acupuncture research methodology is generally pretty awful and in many cases seemingly designed to be positive by avoiding the use of proper blinding and placebo controls. When evaluating the acupuncture literature, the only rational conclusion is that acupuncture doesn’t work for any condition. It doesn’t matter where the needles are inserted, or even if they were inserted. What matters is that subjects, parents, or in some cases the researchers themselves, believe that they were.
Specifically regarding the use of acupuncture for colic, Steve and Orac correctly point out that there is no plausible mechanism for this to work. I’ll add that this includes so-called medical acupuncture which attempts to shed the vestiges of its prescientific past by replacing beliefs about the flow of Qi with scientific concepts such as neuromodulation. Sounding more like science doesn’t make it more like science, however.
They also pointed out that p-hacking almost certainly played a role in the positive outcome, such as it was, concluded by the study authors. That outcomes were cherry picked was painfully obvious. The results represent noise in the data, and if critically appraised this study is actually negative. But there are a few additional interesting details that I wanted to mention.
More excuses from biased researchers
The primary outcome being investigated in the study was the difference in total crying time, as reported by caregivers, between two groups of infants receiving different types of acupuncture. Data was collected multiple times during the first and second week of the study, and again during a follow-up phone call six days after the last treatment. The initial secondary outcome was how many of the infants in either acupuncture group still met criteria for colic during the same intervals. They planned on recruiting 192 patients in order for the study to be sufficiently powered but stopped recruitment well short of that because, according to the authors, “acupuncture for colic became available in other clinics without the need for randomization.”
In order to continue the study at that point, they decided to pool the two acupuncture groups together and compare them to babies that did not undergo any acupuncture treatments. This was after they had already been collecting data for over a year. This is highly suspicious of p-hacking, even if unintentional. In their conclusion, they mention that a comparison of the two acupuncture groups showed no difference in any outcome, which they claim was because of the lack of power, not appearing to even entertain the possibility that it is because they are both equally ineffective beyond placebo. It is clear when reading this paper that the authors had already decided that acupuncture works.
An embarrassment of outcomes
As I mentioned above, the primary and secondary outcomes of the study as initially developed involved comparing two acupuncture groups and looking at total crying time and resolution of colic. They hedged their bets by breaking crying down into three categories: fussing, crying, and colicky crying. There is no reason to do this as the diagnostic criteria for colic does not and these are extremely subjective/meaningless distinctions.
So after a post hoc switch to comparing the combined acupuncture groups to a no acupuncture group, they then looked at more than 30 different possible tertiary outcomes and found four to be barely statistically significant. Considering the questions regarding p-hacking, the small number of patients, and the numerous potential outcomes, these are likely falsely positives. Then there is the issue of blinding.
The issue of blinding
There is no sham acupuncture group in this study, something that actually doesn’t bother me all that much although it would have been a nice addition. That’s a study design that must be used if measuring acupuncture outcomes in older patients but it isn’t that important in babies. But blinding is extremely important because parental placebo effects can completely account for the meager benefits claimed in studies like this.
Blinding was attempted in this study by completely separating the acupuncture provider from the caregivers. Even the nurse who brought the subject into and out of the treatment room was supposedly unaware of whether acupuncture was administered or the baby just sat in the room for a few minutes untouched. They even included an assessment of parental blinding at multiple visits and at follow-up.
One would have expected that parents would guess right about 50% of the time if truly blinded. This was roughly true at the first assessment. However, what the authors found was that parents of children who received some form of acupuncture increasingly believed that their child received acupuncture, with 72% correctly guessing at the follow-up phone call. Less than half of parents of children not being poked guessed correctly.
The authors, again showing bias, chalked this discrepancy up to the assumption that the acupuncture was working. So either acupuncture worked really well, which we know from the actual data would be an unlikely parental conclusion, or blinding was not successful. Perhaps the acupuncturist forgot to put the needles away.
Conclusion: Acupuncture can’t fix colic
Colic is a troublesome entity in the world of pediatric medicine as well as for parents. It is an essentially arbitrary distinction that in many cases appears to result in more harm than benefit. Rather than being seen a simply a child who tends to cry a bit more than others, or in many cases is simply perceived to do so, these babies are given a label which promoters of all manner of bogus treatments can use to effectively manipulate tired and stressed out caregivers.
Believers in acupuncture, even if well-meaning, are doing this. They are denying the evidence when they recommend their pet therapy and, in the case of the authors of the acupuncture for colic study, allowing their bias to influence methodology such that the ultimate conclusion is known before the first subject is recruited. Treating an infant like a pincushion is not an effective means of reducing their crying or learning how to cope with it.