The Role of Inflammation in Infant Colic—A Systematic Review and Meta‐Analysis

ABSTRACT

Aim

Despite being a self‐limiting condition, infant colic is associated with negative consequences for caregivers and infants. Gut inflammation is hypothesised as a contributing factor. We collected and meta‐analysed the most recent evidence on the association between infant colic and gut inflammation.

Methods

A systematic PubMed search was conducted to identify peer‐reviewed English language studies about infants with colic regarding their immunological profile. The search spanned from inception to the 24th of February 2024, identifying 124 articles. After exclusions, eight articles were included for analysis. Statistics on faecal calprotectin were extracted from four studies (infant colic = 227, healthy controls = 166) and meta‐analysed.

Results

Faecal calprotectin was significantly elevated in infant colic cases in five of six studies covering faecal calprotectin with two studies finding a mean difference exceeding 200 μg/g. The meta‐analysis showed a 49.5 μg/g mean difference in faecal calprotectin levels (95% CI = 43–56, p < 0.0001). MCP‐1, MIP‐1β and IL‐8 were elevated in one study with no differences for other biomarkers.

Conclusion

We found evidence of an inflammatory response in infant colic supporting the hypothesis that gut inflammation is involved in the pathogenesis of infant colic. Faecal calprotectin may become a valuable diagnostic tool for infant colic.

Summary

• This is to our knowledge the first review and meta‐analysis summarising the association between gut inflammation and infant colic.

• Faecal calprotectin was found as the only inflammatory biomarker that was significantly elevated in infant colic cases across studies.

• Elevated faecal calprotectin in infants with colic supports the hypothesis of infant colic being linked to gut inflammation, and faecal calprotectin may become a diagnostic tool for infant colic.

Abbreviations

FC

faecal calprotectin

HC

healthy controls

IC

infant colic

MeSH

medical subheadings

NHS

National Health Service

1. Introduction and Background

Crying is an essential part of neonatal communication, allowing nurture and safety to be obtained from their caregivers. In the first 6 weeks of life, daily crying time for healthy babies is approximately 2 h [1, 2], decreasing to just over 1 h in weeks 10 to 12 [3], clustered mostly in the evening and afternoon [4, 5]. When crying significantly exceeds these normal crying patterns for infants, it has long been referred to as infant colic [6]. Infant colic is characterised by excessive crying or fussing, also described as hard‐to‐calm behaviour, in a well‐fed infant with no signs of organic pathology and normal age‐related physical examination, including normal weight gain and growth acceleration [7, 8]. Infant colic is commonly diagnosed by the 2016 Rome IV criteria [9], or the Wessel criteria: 3 h of crying on at least 3 days over 3 weeks [6, 8, 10].

Infant colic is a common reason for visits to primary care physicians [11] and has an estimated incidence of 3%–40% among all infants [7, 12, 13, 14, 15]. Furthermore, the economic cost for the individual and society due to excessive infant crying is substantial [16, 17]. A study from the United Kingdom from 2001 by Morris et al. reported a cost of approximately £65 million to the National Health Service (NHS) as a result of infant crying and sleeping problems in the first 12 weeks [17].

Although infant colic has been recognised as a paediatric syndrome for decades, its aetiology is still unclear and continues to be a topic of discussion [3, 7, 18, 19].

However, gut microbiota dysbiosis does appear to be involved in pathogenesis, while inflammatory marker studies comparing infants with colic to infants without colic show varying results [3, 20, 21, 22, 23, 24, 25, 26, 27].

Infant colic is a self‐limiting condition, which usually resolves between 3 and 6 months of age [18, 28]. However, being a parent of an infant with infant colic can be overwhelming, both physically and mentally. Infant colic is associated with sleep deprivation, aggressive thoughts, disruption of social contacts, and depression for parents. Those caring for infants with infant colic experience more stress, anxiety, and bonding problems compared with parents of infants without colic [34, 35, 36, 37, 38, 39] Infants with infant colic are at increased risk of experiencing abuse and shaken‐baby‐syndrome [7, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40].

It remains unclear if infant colic has long‐term adverse effects on the infant. Development of recurrent abdominal pain syndrome, irritable bowel syndrome and other functional gastrointestinal disorders (FGIDs) has been associated with previous infant colic [19, 34, 40, 41].

To date, no effective treatment for infant colic has been found. Providing parental support and reassurance is the only general recommendation [3, 18, 38, 42, 43, 44, 45]. Lack of effective treatment, along with limited knowledge of the pathogenesis of infant colic, highlights the need for further research into its aetiology and potential treatment options. We aim to investigate how infants with infant colic differ in the immunological biomarker profile. To the best of our knowledge, this is the first study to gather, summarise and meta‐analyse changes in immunological biomarkers in relation to infant colic.

2. Methods

2.1. Data Sources

We evaluated peer‐reviewed articles on PubMed by conducting a search for IC associated with different markers of inflammation. The inflammatory markers selected were based on published literature on the subject. The search terms contained ‘infant colic’ and ‘inflammation’ or ‘inflammatory’ or one of the specific markers of inflammation: ‘calprotectin’, ‘T‐cells’, ‘neutrophils’, ‘leukocytes’, ‘C‐reactive protein’, ‘basophils’, ‘eosinophils’, ‘lymphocytes’, ‘interleukin’, ‘TNF‐α’, ‘cytokines’, ‘chemokines’, ‘interferons’, ‘monocytes’, ‘autoantigens’, ‘plasma cells’, and ‘B‐cells’.

2.2. Study Selection

Our search was constructed based on the literature on IC and the inflammatory processes in humans and related biomarkers, cells, and associated proteins [50]. The PubMed database was searched from inception until the date of search on February 26, 2024. We aimed to solely include peer‐reviewed original articles to ensure the validity of the included evidence. To maximise article output while minimising articles not relevant to the subject Medical Subject Headings (MeSH) were used to narrow the search to specific subcategories of the immunologically associated molecules mentioned. The final input on PubMed being: (infant OR infantile) AND (Colic) AND (inflammation OR inflammatory OR calprotectin OR T‐cells OR neutrophils OR B‐cells OR leukocytes OR CRP OR basophils OR eosinophils OR lymphocytes OR interleukin OR TNF‐α OR cytokines OR interferons OR monocytes OR autoantigens OR plasma cells OR chemokines).

Appendix S1 provides full advanced PubMed search including every entry term from MeSH and word‐tense‐variation.

Following a step‐by‐step approach, we aimed to identify the maximum number of relevant articles for final evaluation. The obtained articles were reviewed by title and abstract. Articles were excluded if they did not focus on infant colic, with infant colic required to have at least its own dedicated subsection. Furthermore, articles were excluded if they did not focus on human infants or failed to mention any inflammatory markers or processes. Articles that did not compare infants with colic with healthy controls regarding their immunological profile were also excluded.

Duplicates, small opinion pieces, clinical guidelines, books and case reports were excluded. Thus, only original peer‐reviewed full‐text studies about both IC and inflammation were included. Studies published in other languages than English were excluded.

3. Results

3.1. Search Results

In total, 124 articles were identified in PubMed. After exclusion based on title and abstract, eight articles remained for full‐text analysis.

Appendix S2 provides a full list of all 124 obtained articles obtained from a PubMed search with individual reasons for inclusion or exclusion for full‐text analysis. The primary reason for excluding articles was insufficient focus on infant colic. Included studies were thoroughly read as full texts and summarised based on their design, characteristics, measurements, objectives, data, results and conclusion (Table 1).

TABLE 1.

Summary and characteristics of included studies.

Author, country and year	Study design and evidence quality rating	Diagnosis	Measurements	Participants (n)	Characteristics	Objective	Results	Conclusion
Pärtty et al. [24] Finland 2017	Case‐control. Evidence score*: 3	IC by Wessel criteria recorded by diary	Blood and faecal samples collected at the age of 3 weeks and 5‐weeks measuring IL‐1β, IL‐6, IL‐8, IL10‐ TNF‐α, Interferon‐γ, MCP‐1, MIP‐1β, CXCL16, and zonulin. Faecal samples were used for comparing microbiota composition	Infants with colic (n = 28), healthy controls (n = 12)	Clinical characteristics comparable between groups. Infants with IC cried for 116 min per day on average while non‐IC controls cried for 62 min per day on average	To assess whether IC is associated with systemic inflammation, infants with IC were compared with infants without IC on their profile of inflammatory markers	Mean MCP‐1 values (pg/mL): Case: 1461 (95% CI: 1242–1681) Control: 971 (95% CI: 737–1205) p = 0.026. Mean MIP‐1β values (pg/mL): Case: 99 (95% CI: 81–118) Control: 65 (95% CI: 51–79) p = 0.003. Mean IL‐8 values (pg/mL): Case: 17.5 (95% CI: 14.9–20.0) Control: 9.9 (95% CI: 6.6–13.3) p = 0.001. All other measurements in blood showed no significant changes	IC is associated with low‐grade systemic inflammation
Savino et al. [22] Italy 2018	RCT and case‐control. Evidence score: 1	IC based on Wessel criteria	Crying time by diary. FC by faecal sample	Infants with colic (n = 32), healthy controls (n = 27)	Similar between groups (feeding type: predominantly breast‐fed)	To compare infants with colic with healthy controls regarding RORγ and FOXP3 mRNA levels, gut microbiota, and faecal calprotectin	Infants with colic had a higher median FC concentration of 500 μg/g [IQR = 415] vs. 145 μg/g [IQR = 106] in babies without colic	Higher median FC values in infants with colic than infants without colic
Rhoads et al. [52] USA 2009	Case‐control. Evidence score: 3	Rome III by Barr diary	FC and microorganisms by stool sample	Infants with colic (n = 19), healthy controls (n = 17)	14–81 days old. Similar characteristics	To determine if gut microflora and/or inflammation takes part in IC pathology. This was done by measuring FC, breath H2, and bacterial species from infants with colic and infants without colic	FC was elevated in IC: 413 ± 71 μg/g in infants with colic. 197 ± 46 μg/g in infants without colic (p = 0.042). Differences in FC cannot be attributed to feeding type	Evidence of intestinal inflammation and neutrophil infiltration in infants with colic compared with controls
Rhoads et al. [26] USA 2018	Nested case‐control. Evidence score*: 3	IC based on 3 h of crying each day on average	Barr diary for crying time. FC and gut bacterial composition was measured from stool samples	Infants with colic (n = 37), healthy controls (n = 28)	Infants aged 21–90 days. Controls were 4 weeks older than cases on average and there was a slight difference in the ethnic composition of the two groups. Otherwise, similar group characteristics	To determine differences in crying time, FC, and gut microbiota between infants with IC and controls, and if such differences are due to different feeding methods	Significantly increased FC in IC, regardless of feeding type (p < 0.001). Median difference in FC values between groups: 204.9 μg/g stool. p = 0.016. FC was found to be the only significant predictor of IC (β = 9.74; 95% CI: 2.96–16.53). Feeding type was not a significant predictor of IC. Seven‐fold higher FC levels in breast‐fed infants compared with formula fed infants: median FC: 183.9 μg/g; IQR, 234 in breast fed infants. Median FC: 23.7 μg/g IQR, 97.7 in formula fed infants. Median for mixed formula and breast‐fed infants: 107 μg/g; IQR, 109	IC is associated with gut inflammation and dysbiosis—no difference was found by feeding type
Sommermeyer et al. [25] Poland 2022	Case‐control. Evidence score: 3	IC by Wessel criteria	FC through point‐of‐care measurement	Infants with colic (n = 95), infants without colic (n = 100)	Similar characteristics between groups. Infants were aged 3–6 weeks at study start	To assess whether FC can be used in the diagnosis of IC by comparing infants with colic and infants without colic according to their FC levels	Significantly increased FC in infants with IC. Findings were not influenced by feeding type, gestation age, gender of the baby, or weight. Average FC levels were significantly higher in IC cases by 46.7 μg/g. p < 0.0001. Caesarean‐born babies had higher FC levels than vaginally delivered babies. Increased by 14.5 μg/g. p < 0.0001. A level of FC over 100 μg/g (prediction model) was associated with IC, while FC levels under 100 μg/g were associated with non‐IC (error margin 11.2% and 13.2%, respectively). An accuracy of 97.8% for prediction of IC in vaginally born infants was reported	FC may be useful in helping physicians diagnose IC
Bernatek et al. [56] Poland 2022	Open‐Label Randomised Study. Evidence score: 2	IC by Wessel criteria	Crying time by diary. FC from stool samples on day 1 and 28	Infants with colic (n = 93), healthy controls (n = 20)	Similar characteristics between groups. Infants aged 3 to 6 weeks at study start. All participants were treated with probiotics	To evaluate probiotic effects on IC, partly by measuring and evaluating FC values between groups at day 1 and 28	FC was significantly higher in IC in both treatment groups compared with infants without colic at both timepoints	Suggests that FC may prove valuable in supporting a diagnosis or monitoring treatment for IC. FC is found elevated at all timepoints in IC cases compared to non‐IC controls. Calls for longer duration studies
Savino et al. [59] Italy 2017	RCT and case‐control. Evidence score: 1	IC by Wessel criteria	Blood samples at day 0 and day 28	Infants with colic (n = 34) 18 of which received treatment and 16 received placebo, healthy controls (n = 25)	Under 60 days old. Similar characteristics	To investigate if infants with colic significantly differ from healthy infants regarding their TLR2, TLR4 and T‐reg profile (FoxP3 as proxy for T‐reg cells)	T‐reg (FoxP3): 1.23 ± 1.29 arbitrary units (AU) in infants with colic, 1.10 ± 0.61 AU in healthy infants (p = 0.286). Not significant. TLR2: 1.48 ± 1.05 AU in IC and 1.40 ± 1.18 AU without colic (P=0.583). Not significant. TLR4: 1.23 ± 0.74 AU in IC and 1.26 ± 0.87 without colic (p = 0.915). Not significant	No significant differences in FoxP3, TLR2 or TLR4 values between groups
Olafsdottir et al. [48] Norway 2002	Case‐control. Evidence score: 3	IC by Wessel criteria	Faecal sample for FC. Blood samples in older infants (not healthy) for immune markers	Part I of study: Infants with colic (n = 76), infants with transient lactose intolerance (n = 7), healthy controls (n = 27). Part II: infants with recurrent abdominal pain (n = 19), infants with IBD (n = 10 Crohn's disease) and (n = 7 ulcerative colitis), healthy controls (n = 27)		To investigate FC values in different gastrointestinal diseases in childhood and infancy, compared with healthy controls	FC values: healthy controls: 277 ± 109 mg/kg−1. Infants with colic: 278 ± 105 mg/kg−1. p = 0.97. Similar FC in boys and girls. FC significantly decreased in healthy controls with age: p = 0.04 from week 2 to 10	No significant difference in FC values from IC cases to healthy controls. FC decreased over time

^*Evidence score is based on the Oxford Centre of Evidence‐based Medicine rating system. 1 = randomised controlled trial/systematic review and meta‐analysis, 2 = unrandomised controlled trial/cohort study, 3 = case‐control, 4 = cross sectional study, 5 = opinion piece, case report.

3.2. Data Extraction and Meta‐Analysis

All analyses were conducted using R (version 4.4.0).

Summary statistics on FC values in infants with and without colic were extracted from the studies. This was done only for original articles with non‐overlapping study populations.

The association between FC and age was examined to determine whether studies with different age groups could be combined in a single meta‐analysis.

We extracted the mean and standard deviation (SD) of FC and the number of participants (n) from each study for both infants with colic (IC) and healthy controls (HC). If only median and interquartile range (IQR) were provided, they were transformed assuming a Gaussian distribution so that $\text{mean}=\text{median}$ and $\mathrm{SD}=\mathrm{IQR}/1.35$.

Mean differences in FC between IC and HC for each study were calculated as $\mathrm{mea}{\mathrm{n}}_{\text{difference}}=\mathrm{mea}{\mathrm{n}}_{\mathrm{IC}}-\mathrm{mea}{\mathrm{n}}_{\mathrm{HC}}$. The corresponding standard errors (SE) were calculated as $\mathrm{S}{\mathrm{E}}_{\text{difference}}=\sqrt{\frac{\left(n_{\mathrm{IC}}-1\right)\mathrm{S}{\mathrm{D}}_{\mathrm{IC}}^2+\left(N_{\mathrm{HC}}-1\right)\mathrm{S}{\mathrm{D}}_{\mathrm{HC}}^2}{n_{\mathrm{IC}}+n_{\mathrm{HC}}-2}}\times \sqrt{\frac{1}{n_{\mathrm{IC}}}+\frac{1}{n_{\mathrm{HC}}}}$.

p‐values of mean differences were calculated using a two‐tailed t‐test.

A fixed‐effects meta‐analysis of the combined difference in FC between IC and HC was conducted using the meta package (version 8.0.1). Plots were generated using the ggplot2 (version 3.5.1), forestploter (version 1.1.2) and ggpubr (version 0.6.0) packages.

Prospero ID: 10111257.

3.3. Article Summaries

Summaries of the eight articles included for full‐text evaluation and analysis are provided in Table 1.

3.4. Review Findings—Number of Measurements

A wide spectrum of immunological biomarkers was measured in the included articles (Table 2) by comparing infants with and without colic. FC was the most measured biomarker of inflammation and was included in 6 articles. Other measured biomarkers include IL‐1β, IL‐6, IL‐8, IL10, TNF‐α. Interferon‐γ, MCP‐1, MIP‐1β, CXCL16 and zonulin. Blood and faecal samples were collected at 3 and 5 weeks, measuring IL‐1β, IL‐6, IL‐8, IL10‐ TNF‐α, Interferon‐γ, MCP‐1, MIP‐1β, CXCL16, zonulin, RORγ, FOXp3, TLR2 and TLR.

TABLE 2.

Summary of results on included inflammatory biomarkers in IC compared to HC.

Immunological biomarker	Significantly increased in infants with IC	No significant difference between case and control
Faecal calprotectin	5 articles	1 article
IL‐1β		1 article
IL‐6		1 article
IL‐8	1 article
IL10		1 article
TNF‐α		1 article
Interferon‐γ		1 article
MCP‐1	1 article
MIP‐1β	1 article
CXCL16		1 article
Zonulin		1 article
RORγ (Th17)		1 article
FOXP3 (Treg‐cells)		2 articles
RORγ/FOXP3 ratio		1 article
TLR2		1 article
TLR4		1 article

Note: Results in articles comparing inflammatory biomarker values among infants with IC and healthy controls.

3.5. Review Findings—Faecal Calprotectin Values

Six articles compared FC values in infants with and without infant colic. Five studies found significantly elevated FC in infant colic cases [22, 25, 26, 52, 56] (Table 2). Olafsdottir et al. [48] found no significant association between FC values and infant colic. According to Sommermeyer et al., FC may be a useful marker for diagnosing infant colic [25].

Since both articles from Rhoads et al. originated from the same study sample, we only included the original study in this meta‐analysis because data on mean difference and SD was stated [2, 26]. Since Bernatek et al. [56] also recruited participants from the same recruitment group as Sommermeyer et al. [25], this article too was ineligible for meta‐analysis. This resulted in four independent studies on FC values in infants with infant colic for the meta‐analysis (Figure 1). No clear pattern emerged when evaluating the FC levels in infants with colic and HC in relation to infant age (Figure 1A). The meta‐analysis found a mean FC difference of 49.5 μg/g (95% CI = 43–56, p < 0.0001) between IC and non‐IC cases (Figure 1).

FIGURE 1.

Summary of independent studies investigating FC in infants with IC compared to HC. (A) Mean and 95% CI of FC for infants with (IC) and (HC) for each included study. The age is the mean age reported in the studies. Best lines were approximated using the geom_smooth function. (B) Forest plot showing results of FC values in IC versus HC for the four included studies and the result of the meta‐analytic synthesis.

3.6. Other Immunological Biomarkers

Values measured by blood samples with no significant differences between infants with and without infant colic included: IL‐6, IL‐10, interferon‐γ, TNF‐α, I‐FABP, CXCL16, Zonulin, FOXP3, RORγ, FOXP3/RORγ, TLR2 and TLR4 [22, 24, 57] (Table 2). By measuring immunological biomarkers Pärtty et al. [24] found elevated MCP‐1 (mean_IC = 1461 pg/mL, 95% CI_IC = 1242–1681, mean_HC = 971 pg/mL, 95% CI_HC = 737–1205, p = 0.003), MIP‐1β (mean_IC = 99 pg/mL, 95% CI_IC = 81–118, mean_HC = 65 pg/mL, 95% CI_HC = 51–79, p = 0.026), and IL‐8 (mean_IC = 17.5 pg/mL, 95% CI_IC = 14.9–20.0, mean_HC = 9.9 pg/mL, 95% CI_HC = 6.6–13.3, p = 0.001) in IC cases.

4. Discussion

Studies analysing faecal samples provide substantial evidence of underlying gut inflammation in infant colic, with significantly increased faecal calprotectin (FC) values found in infant colic cases in five out of six articles on the topic. The meta‐analysis further supports this finding, showing a highly significant increase in FC levels in infants with infant colic (p < 0.0001). The study by Olafsdottir et al. is the only included study that did not find a difference in FC between healthy controls and infants with colic. The reason appears to be the unexpectedly high mean FC levels in the healthy controls of 277 μg/g [48].

We did not observe a clear pattern of association between age and FC in infants with and without infant colic. Instead, it appeared that there was a significant batch effect in the study. This may be due to differences in infant recruitment. Thus, we decided to conduct a meta‐analysis of the mean difference in FC between the groups within each study. Studies with larger population groups and similar recruitment might find an association with age [48, 54].

There is limited evidence on any other inflammatory markers other than FC associated with infant colic, possibly due to the lack of studies on the topic. Pärtty et al. [24] found no significant change in IL‐1‐β, IL‐6, IL‐1, interferon γ, I‐FABP, TNF‐α, CXCL16 or Zonulin values in infants with IC compared with their healthy counterparts. Evidence regarding a strong systemic response in infant colic is limited. The increase of monocyte chemoattractant protein‐1 (MCP‐1/CCL2), macrophage inflammatory protein 1 (MIP‐1β/CCL4), and Interleukin‐8 (IL‐8) in the same study supports evidence of a minor systemic response limited to immune cell activation and mobilisation. The finding that infants with infant colic have elevated levels of FC across studies supports the hypothesis that the immune system and gut inflammation, characterised by neutrophilic infiltration, are involved in the pathogenesis of infant colic.

Gut inflammation may be associated with gut dysbiosis as has been reported in the infant colic literature [23, 26, 52]. Studies have established that infants with colic have a distinct microbiota compared with healthy infants, as it appears that infants with colic have lower bacterial diversity with higher levels of some Proteobacteria and lower levels of Lactobacilli and Bifidobacterium [59, 60]. These findings have led researchers to hypothesise that probiotics may have beneficial effects in infants with colic. Systematic reviews and meta‐analyses conducted by Harb et al. from 2017 and Xu et al. from 2015 found that Lactobacillus reuteri strains may be effective in reducing symptoms of infant colic, but both reviews state that limitations to these findings weaken results [61, 62].

FC, a marker of neutrophil infiltration, is a widely used biomarker for measuring gut inflammation in older patients [46]. FC is already of great value in the clinical setting, aiding in diagnosing and ruling out inflammatory bowel diseases in patients [47, 52].

Furthermore, FC is used to monitor and evaluate the effect of therapeutic interventions in both adults and children with inflammatory bowel disease and can be measured noninvasively in faecal matter [51, 53]. These findings suggest that FC may serve as a potential tool for diagnosing, ruling out, or monitoring infant colic.

Currently the diagnosis of infant colic is based on parental self‐reporting and characteristic symptoms, which introduces potential bias and susceptibility to reporting errors. The potential use of FC as a diagnostic aid underscores the need for more rigorous and standardised diagnostic criteria based on objective measures, benefiting both families affected by infant colic and researchers.

More primary clinical studies on larger groups of infants with infant colic are needed to establish sensitivity and specificity of FC as a diagnostic tool for infant colic in the future. The cause‐effect relationship between infant colic and FC also requires further investigation. Likewise, more studies are needed to provide knowledge regarding inflammatory responses in infants with colic to help understand the pathogenesis of this common and consequential disorder.

5. Limitations

There are several limitations to this study, the most notable being the small sample size with a limited number of articles covering FC in infant colic. It is also worth noting that while the meta‐analysis showed statistically significant elevated FC in children with infant colic, one article by Olafsottir et al. did not find any association between FC and infant colic. Some clinical characteristics of study subjects that were not accounted for by all articles may distort results. Neonates born by caesarean section may have higher levels of FC compared to neonates born vaginally [55]. Some studies have found that different FC values can be detected between types of feeding, with FC levels higher in breast‐fed infants compared to formula‐fed infants [55]. However, Rhoads et al. found that the differences in FC levels between cases with infant colic and healthy controls was not explained by feeding type [26, 55].

There were certain differences in the clinical characteristics of participants between groups, but these were not found to be significant. Recruitment problems in obtained articles introduce potential biases regarding participant selection. In addition, information on infant well‐being was often reported by the Barr diary. Even though this diary has a well‐established role as a valid and reliable measurement of crying time, it remains subject to reporting biases and errors as crying time is measured solely by self‐report [58]. Some parents were non‐compliant in completing the diary [26]. In a clinical setting, diaries are seldom used for diagnosis of infant colic, and diagnosis is often based on current reports of infant behaviour from parents, varying between the use of Wessel, Rome III or Rome IV criteria [58]. Therefore, there is a gap in the objective clinical diagnosis of infant colic and how studies on infant colic are conducted. Another limitation is the lack of reference intervals for measured values, making it difficult to evaluate if the included infants had values representative of the general population. Study findings would optimally need validation in future studies with larger sample sizes and more objective means of measurement. Despite our efforts to conduct an extensive yet focused search, some literature may have been overlooked. The review could have benefitted from retrieving searches in other databases in addition to PubMed, including grey literature. Furthermore, global generalisation of findings may not be possible since the included articles are based on participants from either Europe or the United States.

6. Conclusion

To our knowledge, this is the first review and meta‐analysis of the evidence for an inflammatory response in infant colic. Our findings highlight a substantial need for further research on the subject. Despite conducting a comprehensive search covering a variety of inflammatory biomarkers, we identified a limited number of relevant articles. Although infant colic is a self‐limiting condition, it is likely associated with several consequences and an unclear aetiology [19, 32, 34, 35, 36, 37, 38, 39, 40, 41]. Our meta‐analysis found significantly increased levels of FC (49.5 μg/g) in infants with colic. Two of the included studies found mean differences of more than 200 μg/g in infants with colic, indicating that FC measurement in infants with excessive crying could serve as a diagnostic tool for infant colic and aid in monitoring disease resolution. We propose that future research on infant colic considers the use of FC as a supplement for the current diagnostic guidelines.

Author Contributions

All authors contributed to the study design, data extraction, writing and revision of the manuscript. Dan Boëtius Andersen worked on conceptualising and designing the study and drafted the initial manuscript. Including screening of records of inclusion and extraction of summary statistics from the studies. Julius Fog critically reviewed and revised the manuscript and was responsible for fitting the draft to correspond to author guidelines. Betina Hebbelstrup Jensen critically reviewed and revised the manuscript as well as aiding in choosing the study design and organizing the writing process. Philip Wilson, Ida Scheel Rasmussen and Jakob Kragstrup worked on conceptualising the study and critically reviewed and revised the manuscript. Gritt Overbeck worked on ensuring the validity of the search criteria as well as critically reviewing and revising the manuscript. Marie V. Vestergaard carried out the statistical meta‐analysis. Louise Bidstrup Jørgensen critically reviewed the manuscript and was responsible for optimizing layout. All authors approved the final manuscript as submitted and agreed to be accountable for all aspects of the work.

Funding

The writing and publication of this paper was in part supported by Lundbeckfonden who funded the writing of the original draft of this manuscript by Dan Boëtius Andersen.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Appendix S1: Full advanced Pubmed search.

Appendix S2: Reasons for article exclusion.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analysed during the current study.