Infantile Colic
The Evidence Base For Spinal Manipulation
by John Reggars DC, MChiroSc

A report from World Federation of Chiropractic 8th Biennial Congress Sydney June 2005
Having looked at this topic a few years ago and finding the evidence wanting, I was most interested to hear what noted Danish chiropractic researcher, Niels Grunnet-Nilsson had to say on this subject¹.

Dr. Grunnet-Nilsson described the existing three published research trials, plus the results of an unpublished trial. He expanded on their strengths and weaknesses, as well as adding some anecdotal evidence to support his assertion that chiropractic manipulation was an effective treatment for infantile colic.

The first study cited was a prospective single cohort observational study, in which Dr. Grunnet-Nilsson and his co-workers followed 316 colicky children, who received chiropractic manipulation somewhere between 2-3 times per week for 2-3 weeks, with the average number of manipulations being 3 in 2 weeks.² The manipulation was usually applied with one finger and the adjustive force was very modest, if performed at all. Analysis of the region treated revealed 53% of the infants received upper cervical manipulation only, while another 41% received upper cervical and thoracic manipulation.

The results of this study, according to Grunnet-Nilsson, were “quite dramatic” and warranted a randomized control trial (RCT). According to the published data, at day 14, 34% of the children improved and 60% ceased to be colicky. Overall the mean number of daily hours of crying reduced from 2.5 at day one to 0.65 at day 14, a 74% reduction. Interestingly, colic hours per day reduced from a pre-treatment estimate of 5.2 hours per day to 2.5 hours per day after the first day of treatment.

As the authors point out in their article, the limitations of this type of study include a lack of a comparable control group to avoid such problems as reporting bias and non-specific treatment effects, like the placebo effect.

The second study, aimed at avoiding the limitations of a prospective case series, was a randomized controlled trial comparing spinal manipulation to the drug, dimethicone³. According to Nilsson et al, although dimethicone has been shown to be no better than placebo, it is still used in the treatment of colic and therefore in essence this study would be classified as a placebo controlled trial.

Fifty infants were randomly assigned to either receive manipulation or dimethicone for a period of up to 13 days. The manipulative technique employed was a specific light fingertip pressure applied to restricted spinal segments which were primarily found in the upper and mid thoracic regions and applied on average 3.8 times over the study period. All 25 infants in the manipulation group completed the trial however there were 9 dropouts in the dimethicone group. By day 12 of the study the manipulation group showed a 67% reduction in crying hours while the dimethicone group only showed a 38% reduction. The greatest improvement occurred after day 5 in the manipulation group, with the average number of manipulations being 3.8 over the study period.

Trial three was another RCT in which 15 infants received spinal manipulation, six times over 2 weeks, and 15 infants received detuned ultrasound.⁴ The results of this study indicated that there was a statistically significant difference in favour of manipulation, with 93% of the spinal manipulation group being colic free within 3 weeks.

Unfortunately, this study has not been published in the peer reviewed literature with only the abstract from the conference proceedings being available. This does not provide any detail of the study regarding statistical analysis, inclusion criteria, methodology etc.

Finally, the last study, and the best designed of the four studies, was a randomized controlled double blind study conducted in 2001 by Olafsdottir et al.⁵ In this trial 86 infants with colic were randomized into two groups to receive either spinal manipulation or no treatment. The infants were taken from their mother by a nurse and then either given to the chiropractor for manipulation or was taken into the consultation room and merely held for 10 minutes by the nurse. An important strength of this study was that the mother was blinded to what intervention the child received. The infants received three intervention sessions over 8 days, followed by an observation period of 8-14 days.

At the end of the trial there were no statistical differences in crying hours between both groups with the treatment group achieving a reduction in crying hours from 5.1 hours to 3.1 hours while the control group decreased crying hours from 5.4 to 3.1 hours.

The authors concluded, “The study emphasises the need to investigate similar alternative methods of treatment by placebo controlled and blinded studies in order to document whether these treatment regimes are effective or not”. It must be noted that this study has drawn some minor criticisms in that the authors failed to publish post-treatment follow up results, confidence intervals and relative risks associated with the treatment.⁷

In his description of the studies, Dr. Grunnet-Nilsson was critical of Olafsdottir et al study⁶ in that he argued that this study limited the treatment group to a maximum of 3 treatments over 8 days, whereas in the other two RCT’s the infants received either between 4-7 treatments³ or up to 6 sessions⁴. According to Dr. Grunnet-Nilsson the lack of response to manipulation by the infants in the double blinded trial⁵ could be attributed to a dose response problem. In other words, more treatment was required to get an effective response. I find this assertion less than convincing in light of the published data. In the Danish RCT³ where the results favoured manipulation, the infants received on average 3.8 manipulations, compared to 3 manipulations in the less favourable trial.⁵ From clinical experience one would not expect a significant difference in outcomes between these two amounts of treatment. Furthermore, in the Danish RCT the greatest treatment effect occurred within the first 5 days of the 2 week treatment period. On the details provided, each infant received between 3-5 treatments and therefore, although not expressly stated, it would appear that each infant would not have received the maximum number of treatments, i.e. 5, during that period, but more realistically between 2-3 treatments. Also, in the Danish case series study² the authors stated that there was a significant improvement after just one treatment and that the average number of treatments was three. In my opinion, the argument that the unfavourable outcome in the Olafsdottir study⁵ was due to a dose response problem is not supported by the published data.

Dr. Grunnet-Nilsson also stated that the 9 dropouts in his RCT³ were “primarily because of worsening symptoms” and had these subjects been included the drug treatment group results would have been far worse. While this maybe so, without objective evidence to support this statement, this is pure conjecture. In fact, as reported in the published study, 5 infants dropped out before even the first week’s diary had been completed and the conclusion that the infants’ condition had worsened was a subjective evaluation by the mothers. Appropriate clinical research design and analysis makes allowance for dropouts by the use of an “intention to treat analysis”. As others have rightly pointed out, had this method been applied to this study, rather than showing a worsening in the drug group, the positive effects for the manipulation group would almost certainly not be as beneficial.⁷

The inclusion, by Grunnet-Nilsson, of the Mercer and Nook study⁴ in support of the argument that spinal manipulation is an effective treatment for infantile colic is also questioned. This small trial was published in an abstract form, within a conference proceedings, and as such lacks detail on important aspects of its methodology as well as preventing scientific scrutiny. Hughes and Bolton⁷ in their review of manipulation for infantile colic included this study but concluded that “The small sample group without well defined inclusion data and the lack of detail in methodology and recorded data seriously undermines the contribution of this study to the evidence base”.

Dr. Grunnet-Nilsson, in his concluding remarks, also stated that according to a recent Cochrane Review there was no evidence to support the placebo effect⁸ and on that basis the benefit derived from manipulation is due to the intervention and not placebo. An earlier study by the authors of this review⁹ has been widely criticized by others,^10,11 it should be remembered that the review relates purely to the “true” placebo effect and not the many other non-specific treatment effects, which can have a significant effect on outcomes in the studies like those cited above. It’s also worth noting that in one study on Dimethicone and colic the researchers attributed a 67% improvement to a high-grade placebo-effect.¹² Studies that involve self-limiting diseases, with a short natural history, and subjective reporting, such as hours of crying as reported by the mother as in this case, are prone to be influenced by many non-specific treatment effects. The use of double blinded RCT’s avoids many of these pitfalls and in the only study of this type investigating the effect of manipulation on infantile colic fails to show a therapeutic benefit. I also found Dr. Grunnet-Nilsson’s refutation of the placebo effect to be at odds with his own conclusion, “that if the mother is in favour of her infant receiving spinal manipulation, this treatment is likely to be successful”. If the “placebo effect”, as Dr. Grunnet-Nilsson describes it, does not exist why should the mother’s preference for this treatment have any bearing on the outcome?

While reviewing the two Danish studies, both of which included a description of the manipulative technique and regions treated, I noted some significant differences. In the case series² 53% of the infants received upper cervical manipulation only, while another 41% received upper cervical and thoracic manipulation. In the RCT³ manipulation was applied to primarily the upper and mid thoracic regions. The question thus arises if colic is due to some sort of spinal segmental dysfunction how and why could such different spinal regions be involved? Furthermore, as previously noted, in one study³ the described manipulation was usually applied with one finger and the adjustive force was very modest, if performed at all. I would argue that light finger pressure without a manipulative thrust could hardly be classified as spinal manipulation and would more appropriately be described as “laying on of the hands”?

In summary, I found the evidence presented does not provide sufficient support for the treatment of infantile colic with spinal manipulation and that any beneficial effect from the treatment is not due the manipulation per se, but rather non-specific treatment effects. However, notwithstanding the absence of evidence in support of this intervention, I am moved by the conclusion of Hughes and Bolton,⁷ “In this clinical scenario where the family is under significant strain, where the infant may be at risk of harm and possible long-term repercussions, where there are limited alternative effective interventions, and where the mother has confidence in a chiropractor from other experiences, the advice is to seek chiropractic treatment”.

The bottom-line, based on the evidence, I would not advocate the referral of all infants with colic to seek chiropractic manipulation but in the absence of any potential risk from the treatment and importantly where the mother thinks the treatment will work a short trial of 2-3 treatments maybe appropriate.

References:

1. Grunnet-Nilsson N. Infantile colic-The evidence base for spinal manipulation. In Haldeman S, Bolton P, Rosner AL eds, 8th Proceedings of the Biennial Congress World Federation of Chiropractic –Sydney, Australia 2005:69-73.
2. Klougart N, Nilsson N, Jacobsen J. Infantile colic treated by chiropractors: a prospective study of 316 cases. J Manipulative Physiol Ther 1989;12(4):281-8.
3. Wiberg JMM, Nordsteen J, Nilsson N. The short-term effect of spinal manipulation in the treatment of infantile colic: A randomized controlled trial with a blind observer. J Manipulative Physiol Ther 1999;22(8):517-22.
4. Mercer C, Nook B. The efficacy of chiropractic spinal adjustments as a treatment protocol in the management of infantile colic. In Haldeman S, Murphy B, eds, 5th Biennal Congress of the World federation of Chiropractic., Auckland, New Zealand 1999:170-1.
5. Olafsdottir E, Forshei S, Fluge G, et al. Randomized controlled trial of infantile colic treated with chiropractic spinal manipulation. Arch Dis Child 2001;84(2):138-41.
6. Nilsson N, Wiberg J. Infantile colic and chiropractic spinal manipulation. (Letter to the Editor) Arch Dis Child 2002;85:268.
7. Hughes S, Bolton J. Chiropractic may be an effective treatment in infantile colic? Arch Dis Child 2002;86:382-4.
8. Hróbjartsson A, Gøtzsche PC. Placebo interventions for all clinical conditions. The Cochrane Database of Systematic Reviews 2004, Issue 2: CD003974.
9. Hrobjartsson A, Gøtzsche P. Is the placebo Powerless? Analysis of clinical trials comparing placebo with no treatment. N Engl J Med. 2001;344:1594-602.
10. Bailar JC. The powerful placebo and the wizard of Oz. N Engl J Med. 2001; 344:1630-2.
11. Kaptchuk TJ. Is the placebo powerless? N Engl J Med. 2001;345:1276-9.
12. Danielsson B, Hwang CP. Treatment of infantile colic with surface active substance (simethicone). Acta Paediatr Scand. 1985;74(3):446-50.