- Email: [email protected]
Colic—A Pain Syndrome of Infancy?
Acute Pain in Children
0031-3955/89 $0.00
+ .20
Colic-A Pain Syndrome of Infancyl M. Alex Geertsma, MD, * and Jeffrey S. Hyams, MDt
ASSUMPTION OF PAIN IN INFANTILE COLIC The clinician is most likely to invoke the diagnosis of infantile colic when faced with the frantic sleep-deprived parent who has been dealing ineffectively with prolonged periods of crying in a less than 3-month-old infant. Parents seek a cure, not only for the alleviation of their sleep deprivation, but also because they often infer that their baby is in pain. Why is it that parents infer pain from the behavioral colic constellation? Physical signs such as pulling up of the legs with apparent abdominal cramping and passing of flatus are often observed with colicky crying. Another reason for the assumption of pain stems from the communicative function of crying. Lester and Zeskind27 discuss the evidence that the infant cry often reflects a specific need state or pain. Once the pain is alleviated or the need state fulfilled by the parent, the communication-the cry-stops. It is therefore understandable that prolonged crying that does not abate with the usual parental trial and error interventions, namely feeding, rocking, holding, verbally soothing, and changing, creates an escalating concern that there is a major unmet need or unalleviated pain. For these reasons, a significant amount of literature concerning colic is directed toward the discovery of an organic or physical cause of the presumed pain. However, there is an equal body of literature that proposes a nonorganic etiology for colic. This perspective, largely behavioral in orientation, proposes the absence of actual pain. Taken as a whole, then, the pediatric colic literature is confusing. Is there evidence for pain in colic? If so, what causes it and how should the clinician approach the pain? Are alternative behavioral and psychosocial explanations justified? If so, how are they to be reconciled with the evidence of pain? Finally, why does this disagreement exist? *Ciinical Associate Professor of Pediatrics, Division of General Pediatrics
t Associate Professor of Pediatrics, Division of Pediatric Gastroenterology From the University of Connecticut Health Center, Farmington; St. Francis Hospital and Medical Center and Hartford Hospital, Hartford, Connecticut
Pediatric Clinics of North America-Vol. 36, No.4, August 1989
905
906
M.
ALEX GEERTSMA AND JEFFREY
S.
HYAMS
This article considers the evidence for infantile colic as a pain syndrome. In order to accomplish this, the literature on parental and observer interpretation of cry and pain signals will be reviewed. Then the formal evidence for pain and its causes in the colic syndrome will be considered, as will alternative behavioral explanations of colic. Finally, an attempt will be made to explain the lack of agreement about colic and to propose a treatment approach.
PARENTAL INTERPRETATION OF INFANT CRYING A major difficulty in studying infantile colic is that the presence of pain versus other causes of crying must be inferred by the observer. Because the young infant cannot use language to communicate, the parent or observer is often dependent on interpreting relatively subjective behavioral signs. Given the tendency to assume pain as a cause for colic, how accurate are parents in assessing the actual meaning of their infant's cry? Work addressing this question appears primarily in the child development literature. WolfI52 observed a small sample of 14 infants and recorded cry in different circumstances including hunger, pain, and cold. He noted that cry spectrographs displayed consistent unique patterns for each of the conditions. He also noted that mothers of the infants were largely able to distinguish the circumstance of crying by sound recording playback alone. Wasz-Hockert et al48 found results similar to Wolff, showing cries unequivocably associated with certain situations (hunger, pain, and so forth) had very specific spectrographic patterns. In contrast to Wolfl's limited study, more detailed work by Muller et aP6 cast doubt on the accuracy of maternal cry interpretation. While mothers of infants could correctly identify their childrens' hunger cries from sound alone, they made the judgment of hunger many other times for spectrographic patterns not characteristic of hunger. That is, they overinterpreted crying as reflective of hunger. Work by others 37• 36 largely confirms the latter two studies. Essentially, the evidence suggests that cry characteristics for a given infant do vary according to the cause of the cry, potentially conveying meaning to the parent. However, parents may misinterpret the cry and may tend to err on the side of a particular presumed etiology such as pain or hunger. This tends to support the view that at least some presumed colicky crying may have different meaning than the parent might assume.
POSSIBLE CAUSES OF PAIN IN COLIC Certainly, defined physical illnesses and structural anomalies can cause mysterious pain and, hence, crying in young infants. Incarcerated hernias, otitis media, fractured clavicles, hematomas secondary to inapparent
COLIc-A PAIN SYNDROME OF INFANCY?
907
trauma, and a multitude of other discernible causes of pain could theoretically cause a young infant to cry vigorously and persistently. However, most would agree that these conditions are usually easily diagnosed. Most would also agree that an essential component of the colic syndrome is the occult nature of its causation. Therefore, such easily determined causes of infant crying should exclude the diagnosis of colic.
GASTROINTESTINAL CAUSES OF COLIC For well over 50 years most physicians have considered abdominal pain to be of primary importance in the colic syndrome. In 1923, White SO used the term gastroenterospasm to describe what he considered autonomic imbalance leading to abdominal pain and excessive crying in infants. Abdominal pain is classically separated into visceral pain and parietal pain. Parietal pain arises from noxious stimuli to the parietal peritoneum and is exacerbated by movement, exactly the opposite from what is observed in the colicky infant who often improves with movement. Visceral pain arises from noxious stimuli affecting an abdominal viscus. It is frequently cramping in nature and often accompanied by autonomic symptoms such as sweating and restlessness. This constellation of findings bears some resemblance to the findings in infants with colic. Visceral pain fibers are generally only sensitive to stretching or tension in the wall of the gut or in the capsule of solid organs. Tension must be increased relatively rapidly for pain to appear. Distension of the gut wall is commonly produced by either intraluminal gas or fluid. This has suggested several possible etiologic factors for the production of colic in infants. Malabsorption and Gas Function When dietary carbohydrate is not completely absorbed by the small intestine, it passes into the colon where it is fermented by colonic bacteria. 28 This process results in the production of short chain organic acids and the evolution of several gases. The short chain organic acids are either absorbed by the colonic mucosa or, if present in excess, act as a cathartic and are passed in the stool. 43 Approximately 15 per cent of the gas produced during this reaction (primarily CO 2 and H 2) is absorbed into the bloodstream and the remainder is passed as flatus. 29 Abdominal cramping is a common accompaniment to carbohydrate malabsorption. Most term infants cannot fully absorb the lactose load contained in formula feedings in the first several months of life despite having "adult" levels of mucosal lactase activity.2, 4, 29, 35 Breath H2 levels are quite high following feedings with either breast milk or lactose-containing formula. 4 One could postulate that colicky infants have increased intraluminal gas production following feedings. Liebman,30 using stool examination for pH and reducing substances, was unable to find a relationship between lactose malabsorption and colic. Unfortunately, the methodology for detecting carbohydrate malabsorption was relatively insensitive and future studies using newer techniques will be required to examine this issue further. Another common dietary carbohydrate that is not completely absorbed
908
M.
ALEX GEERTSMA AND JEFFREY
S.
HYAMS
in the first several months of life is starch. Amylase activity in the pancreas of the term neonate is only 10 per cent of that found in adults; one study found no amylase activity in the duodenal fluid of some infants at birth. 26 Thus, the ingestion of cereal in the first few months of life might pose a theoretic problem to the infant. In one study, the prevalence of colic was found to be greater in a group of infants fed cereal in the first 3 months of life compared to infants receiving no solids. 17 What was not clear from that study was whether the infants were receiving the cereal before the colic developed or, in fact, were receiving it as a potential treatment for persistent crying. Deisher and Goers13 found no difference in the incidence of colic between infants who received cereal in the first month of life compared to those who did not. Thus, it is not clear at the present time if colicky infants have increased intraluminal gas production. In 1952, Jorup22 suggested that no excessive gas could be found using radiologic studies in a group of infants with colic. In adults with irritable bowel syndrome characterized by cramping and bloating, no increased colonic gas can be demonstrated either, but these individuals have increased sensitivity to bowel distention caused by a balloon inflated in the colon. 41 Whether infants with colic may be comparable in having this heightened sensitivity to bowel distention remains to be determined. Additionally, it is not clear if infants with colic differ from non colicky infants in their ability to propulse gas along the gastrointestinal tract. Since a small percentage of swallowed air is not lost through eructation and enters the small intestine, it is possible that colicky infants might swallow more air, expel less gas through ineffective burping, or handle the gas which enters the small bowel less effectively. Gastrointestinal Allergy Although allergy to cow's milk has been well documented to result in a variety of gastrointestinal and nongastrointestinal symptoms in some infants,47 several reports also suggest that intolerance to cow's milk containing formula might be associated with infantile colic. 21 , 33 It is important to note that these studies have focused on the protein components of cow's milk rather than the carbohydrate lactose discussed previously. As long ago as 1901,53 a report appeared suggesting that "proteids" are contained in the breast milk of certain women whose infants have colic and that these substances may cause abdominal cramping. In 1983, Jakobsson and Lindberg21 reported that 50 per cent of breast-fed infants with colic improved when the mother eliminated cow's milk protein from her diet, In 35 per cent the symptoms of colic returned on repeat challenges with cow's milk protein. Lothe et aP3 demonstrated that 18 per cent of 60 colicky infants receiving a cow's milk formula improved when switched to a protein hydrolysate formula, The authors postulated that sensitivities to cow's milk and soy protein were important causes of infantile colic. These findings have not been corroborated in other studies. Evans et aP4 found no improvement when the breast-feeding mothers of 20 infants with colic avoided cow's milk in a double-blind, placebo-controlled, crossover study. Liebman30 found no elevation of total IgE levels or radioallergosorbent test (RAST) to cow's milk protein in 56 infants with colic and
COLIc-A PAIN SYNDROME OF INFANCY?
909
concluded that COW'S milk protein allergy was not a significant cause of infantile colic. However, other reports have suggested that normal IgE and RAST levels may be found in infants with cow's milk allergy if they do not have systemic signs of allergy such as eczema or wheezing. 18 Recently, Kahn et al24 suggested that cow's milk allergy may be a cause of infantile insomnia. They reported eight infants who slept extremely poorly and in whom substitution of a protein hydrolysate formula for one with cow's milk protein normalized sleep patterns. When cow's milk protein was reintroduced into the diet, the infants again developed sleeplessness. All had elevated IgE levels and 50 per cent had elevated RAST levels to beta-lactoglobulin. Most had a history of either wheezing or eczema. Another component of infant formula that has been postulated to lead to colicky symptoms is iron. Many physicians will change from an ironcontaining to an iron-free formula in the infant with fussiness, diarrhea, or constipation. However, Oski39 could show no difference in gastrointestinal symptoms between two similar groups of infants fed either an iron-fortified or a conventional cow's milk formula. Gastrointestinal Reflux Spitting and vomiting are common in the first several months of life and only uncommonly lead to significant problems. Yet, a small number of these infants will also manifest extreme irritability and raise the possibility of pathologic gastrointestinal reflux with esophagitis. 19 Clinically, these infants often awaken from sleep with excessive crying, particularly 1 to 2 hours after a feeding. The establishment of a relationship between irritability and gastrointestinal reflux in an infant may be difficult. Therapeutic trials of antacids during these crying spells may be attempted. In the infant with intractable crying, performance of prolonged intraesophageal pH probe monitoring may be helpful. If the clinician can document a fall in intraesophageal pH preceding periods of irritability, a causal relationship can be presumed. Hormonal Abnormalities Since abnormalities in intestinal peristalsis have been presumed to be important in the pathogenesis of infantile colic, it is natural that some attention has focused on gastrointestinal hormones that regulate intestinal motility. In 1963, Clark et aP2 suggested that infants with colic may be deficient in progesterone, a hormone which may relax smooth muscle. When colicky infants were given progesterone, their symptoms improved. Unfortunately, this study was not performed in a blinded or controlled manner. More recently, Lothe et aP2 have evaluated the levels of motilin, vasoactive intestinal peptide (VIP), and gastrin in infants with colic as well as in a group of age-matched controls. No differences were found in VIP and gastrin levels, but the infants with colic had a higher mean motilin level than did noncolicky infants. Motilin has a stimulating effect on gastrointestinal motility. Whether this elevation is a primary contributor to colic or rather a secondary phenomenon remains to be established. In summary, although pain, especially gastrointestinal pain, remains a
910
M.
ALEX GEERTSMA AND JEFFREY
S.
HYAMS
common explanation for colicky infant behavior, it is evident from the preceding discussion that the evidence, at present, is not conclusive.
EVIDENCE FOR BEHAVIOR AND INTERACTION CAUSING COLIC Behavioral and interactive factors have also been suggested as causes of colicky crying. This view holds that there is in fact no physical cause of colic and hence no true pain. Instead, it is postulated that the infant becomes distraught and cries due to problems with interaction and handling by the environment or simply due to intrinsically difficult infant temperament. Intrinsic Temperament and Colic While the concept of temperament and temperamental types has gained some popularity in pediatrics,46 it remains controversial in research circles, especially the assumption of stability over timeY Nevertheless, it is common to assume that the difficult child was colicky as a young infant and, in turn, that the colicky infant usually becomes a difficult temperament older child. Actually, very little exists in the pediatric literature directly addressing stability of temperament or the relationship between colic and subsequent temperament. What does exist is either inconclusive or refutes the suggested relationship. In 1972 Careys attempted to correlate scores on his temperament scale with various clinical conditions including colic. He claimed to show a high correlation between colic documented from office records and difficult temperament as recorded by parental assessments at 4 and 8 months. This conclusion must be viewed cautiously because of the possibility of parental report bias. Parents who experience difficult, colicky young infants might be expected to see subsequent, relatively normal infant behavior as still difficult. In general, the objectivity of parental assessments of temperament is a major point of contention in the research literature. 23, 42 Perhaps due to the problems of parental report reliability and the methodologic problems of trying to avoid them, very little has been published addreSSing the possible relationship between colic and temperament. At best, limited associations have been suggested although the overall results have been inconclusive or negative. 1. 44 Maternal Caretaker Maladaptation Another commonly held view is that colic is caused by maternal caretaker maladaptation or psychological problems. This view originated from episodic clinical observations, published editorial opinion, and older psychiatric case studies noting that a change of environment and caretaker would at times lead to a change in the difficult crying. Specific mechanisms were often unclear. However, an inability or unwillingness of the mother to effectively calm or nurture the crying infant was generally assumed. Lakin,25 for example, in a retrospective interview study, described
COLIc-A PAIN SYNDROME OF INFANCY?
911
mothers of colicky infants as having less adequate marital adjustment, poorer parent-child relationships, and a greater sense of inadequacy as women. In a much quoted early study, Wessel et al 49 claimed to find a high degree of family tension in homes with colicky infants. In both studies, retrospective design leaves open the alternative interpretation that family tension and maternal insecurity are the products of, rather than the causes of, colicky infant crying. Using a prospective assessment of maternal anxiety, Carey9 noted an association between maternal anxiety and colic. However, colic was documented via office records making it uncertain if maternal anxiety simply led to inflated reports of crying. Paradise 40 studied 146 newborns and their families. He documented maternal adjustment and personality factors prospectively using objective and reproducible measures such as the Minnesota Multiphasic Personality Inventory (MMPI). Paradise noted that there was no evidence to support the premise that colic somehow arises from "an unfavorable emotional climate created by an inexperienced, anxious, hostile, or unmotherly mother." The majority of these early behavioral-interactive studies suffered from similar methodologic problems, namely, potential report biasing, retrospective data, imprecise measures, and small study populations. What has been characteristic of behavioral pediatric studies in the 1980s has been the use of sophisticated forms of statistical analysis, computer implementation, and refined research design. Forsyth et al's prospective study 15 of 373 mothers found that colic and excessive cry reports more often came from mothers who displayed one or a combination of a number of psychologic risk factors. Unfortunately, criteria used to qualify for colic were diffuse and evidently heavily dependent on general parental impression. It is, therefore, unclear whether Forsyth et al were documenting an association between psychologic risk factors and the actual occurrence of difficult crying or simply negative parental impressions of infant behavior. If one believes that all colic can be explained by psychosocial and behavioral factors, then it stands to reason that one should be able to intervene and decrease colic by behaviorally oriented intervention. Taubman 45 tested this theory by giving parents of colicky infants a detailed stepwise behavioral intervention to use with their colicky infants. He gave them a rigid set of instructions of what to do whenever their infants cried. He reported a significant decrease in total crying versus a control group. Unfortunately, the study, while impressive in its success, shed little light on the actual mechanism by which the intervention was successful. Therefore, we have no better idea whether colic in his study population was due to interactional problems, problems with maternal interpretation of cry, or ill-defined psychosocial instability. Although not studying colic per se, Barr and Elias 3 came to similar conclusions. They noted that more frequent "contingent" feeding resulted in less fretful crying than longer intervals between feeding generated by a more defined feeding schedule. This is consistent with the classical child development research on crying performed by Bell and Ainsworth. 6 In summary, much like the research on organic causes of colic, the
912
M.
ALEX GEERTSMA AND JEFFREY
S.
HYAMS
behavioral and interactive literature also appears to be inconclusive and contradictory. Therefore, it is not surprising that clinicians and parents alike remain uncertain about how to approach disturbing infant cryingwhether it is reflective of pain, a pattern of normal but misinterpreted crying, or indicative of an interactive disorder. Various writers have speculated on the reasons for the lack of helpful research. 9. 20 Almost all critical reviewers recommend better research, better methodology, and less dependence on clinical impression and anecdotal reports. However, few have actually critiqued the methodology of existing research. Doing so may not only suggest better research design, but also insights into more effective treatment. PROBLEMS WITH DEFINITION AND METHODOLOGY Detailed reviews of both organic and behavioral colic studies reveal several important methodologic points. One broad methodologic deficiency is the dependence on retrospective data. As noted in our discussions, various organic and behavioral studies have used retrospective reports of colic and crying. This approach, along with restrictive definitions, can overly limit the population being studied. It can, in turn, lead to mistaken conclusions about the applicability of the findings. As Carey9 points out, there is also a lack of agreement concerning the definition of colic. In a number of studies discussed in this article, the definitions used were very imprecise or not described in sufficient detail. For example, in Forsyth et aI's study, 15 the presence of colic was determined primarily by impressionistic parental reporting. Other studies have utilized more precise definitions dependent on less subjective information. However, inclusion criteria generated by those definitions may have resulted in overly broad definitions and, hence, excessively large study populations. As an example, Taubman 45 and others have used Wessel et aI's definition of colic. This included a very specific criterion of 3 or more hours of crying per day for more than 3 days in any 1 week. Using this criterion himself, Wessel et al49 identified 49 per cent of a normal nursery population as eventually colicky, clearly far too high a number, thereby raising doubts about the validity of the definition. In fact, Brazelton, 7 in a prospective cry study, documented that the median amount of crying at 6 weeks is 2% hours total per day. Implicit in the extensive work on the spectrographic analysis of crying is the possibility that the severity of cry not only relates to the potential degree of generated parental distress, but also the likely meaning of the cry. It is not inconceivable that a louder, more persistent cry might indicate pain, whereas less intense but frequent crying may indicate something else, such as inadequate parental response to infant needs or simply immature sleep cycling. In many studies-both those supporting pain-gastrointestinal causes and those supporting behavioral-interactive etiologies-little is done to objectifY either the amount or the character of the crying reported. Neither
913
COLIc-A PAIN SYNDROME OF INFANCY?
CareylO nor Paradise 40 required recording of actual cry periods. Instead, they depended on office records of parental impressions of the severity of cry. Similarly, Jakobsson and Lindberg's study21 and Lothe et aI's hormonal study32 did not require quantitative cry records. It is true that some work does pay attention to estimates of the severity and quality of crying, although not requiring precise measures of the severity. Other studies choose to totally ignore certain aspects of infant crying, thereby ignoring possible clues to the etiology of the crying. In general, studies that claim to prove gastrointestinal causes for colic tend to study primarily those children with intense persistent crying and associated gastrointestinal symptoms. Those that support a behavioralinteractive etiology tend to use either diffuse assessments of the quality of cry or none at all. Either approach, given the possibility that quality of crying and associated symptoms might convey meaning, could lead to spurious or biased results favoring one clinical approach over another (Table 1). It is quite likely that this inconsistency, along with dependence on retrospective methodology and use of imprecise definitions, has contributed significantly to the confUSing and apparently contradictory results of colic research.
A PROSPECTIVE STUDY OF COLIC-PATTERNS OF CRY AND RELEVANCE FOR PAIN IN COLIC In view of the methodologic problems in colic research and to better answer the question of the presence of pain and an organic etiology, we undertook a pilot prospective study of crying and colicky behavior in newborns. We asked each of 60 families to record at between 2 and 4 weeks and Table 1. Definitions and Presumed Etiology of Colic DEFINITION
Gastrointestinal Studies Jakobsson and Lindberg (1983) Lothe et al (1982) Lothe et al (1987) Liebman (1981)
Behavior-Interaction Studies Taubman (1984) Wessel et al (1954) Carey (1968) Forsyth et al (1985)
Persistent cry, abdominal distension, flatus Persistent cry, abdominal distension, flatus Intense cry, abdominal distension, flatus 2:3 hr/day crying, no gas symptoms
Retrospective parent report 2:3 hr/day crying
CAUSE
Cow's milk protein Cow's milk protein Increased s-motilin levels No evidence of allergy or malabsorption
Parental misinterpretation of cry >3 hr/day crying for 3 days/week Family tension Retrospective parent report 2:3 Maternal anxiety hr/day crying Retrospective parental impression Maternal psychosocial of crying issues
914
M.
ALEX GEERTSMA AND JEFFREY
S.
HYAMS
again at between 6 and 8 weeks a detailed log of infant crying, associated infant behaviors, and parental responses. The use of infant behavior logs has been studied formally and proven to be a reliable source of behavioral information. 5 Such logs make the systematic collection of extensive behavioral data from a large study population feasible. Furthermore, methodology studies confirm that parents can be accurate observers and reporters if they are asked to limit their observations to specific circumscribed parameters. This is in distinction to retrospective interview data that are dependent on impressionistic recollections of parents. In the case of our study, parents were carefully instructed to record clear assessments of whether their child cried during a given hour and semiquantitative estimates of the amount of crying that occurred each hour. Parents also recorded the actual presence or absence of flatus passage during the same observation periods. The global opinion of parents about their infants' crying is of clinical importance because it is parental distress about the crying that the clinician must address. Therefore, we also asked parents to record the degree of satisfaction with their babies' crying. We accomplished this by asking them to fill out an infant behavior assessment prior to each cry log. Preliminary analysis of the data revealed patterns of great potential relevance to our earlier discussion of past colic research and possible treatment approaches. First, different patterns of crying were evident from analysis of the cry logs. One distinct pattern was noted in infants who cried intensively; that is, when they cried, they cried for most of the hour, regardless of what parents did. This intense crying group also passed flatus significantly more than other infants. We called these babies "classic colic" infants. They made up approximately 9 per cent of the prospective study group. Another pattern was observed in infants who did not cry intensively, but who calmed quickly and stopped crying shortly after their parents intervened with them. They tended to cry frequently, displaying low intensity crying in many of the 24-hour periods on the days observed. We labeled these babies "frequent cry" infants. They made up 10 to 11 per cent of the study population. The remainder of the babies were termed "noncolic" infants. Parental distress and dissatisfaction with infant crying were significantly greater for the "classic colic" intense criers. Parental dissatisfaction was least for the "non colic" babies, whereas dissatisfaction with the "frequent cry" group was significantly greater than the "noncolics" but less than the "classic colics." This preliminary study raises important questions concerning past and future colic research. First, it may be that closer, detailed attention must be given to specific patterns of crying in colic research. Simply summing the amount of crying per 24 hours may not be precise enough. The intensity of cry (the inverse of "soothability") may need to be distinguished from frequency of cry. In studies using Wessel et aI's criterion both types of crying may be lumped together. A child who cries intensely and inconsolably with passage of flatus may be grouped together with infants who cry frequently but are
COLIc-A PAIN SYNDROME OF INFANCY?
915
easily consoled. Moreover, using the criterion of 3 hours of crying per day may define even more infants as colicky than just classic colic and frequent cry infants. In Wessel et al's study49 as much as 49 per cent of the population was defined as colicky. Conversely, the intensely crying infant with signs of gassiness, even though very distressing to a parent, may be excluded altogether from an intervention study if the crying does not total 3 or more hours per day. This may leave frequent but consolable criers and exclude some intense criers from the study. These groups of infants may be quite different from each other, but, as demonstrated in our study, still distressing to their parents. If a given intervention, such as behavioral parent training, is successful with the remaining group, it may be misleading to suggest that the intervention is reliable for all infants whom parents consider difficult criers. It is also possible that these distinct cry groups may have different causes for their crying. It is not inconceivable, for instance, that the high intensity, high flatus "classic colic" infants have true pain from an organic cause, perhaps gastrointestinal in origin. It is possible that the frequent but easily consolable crier group-the "frequent crier" group-is a group of infants who awaken and initiate crying frequently with no true organic causes. These infants may well be more responsive to changes in parental caretaking behavior. Having two relatively distinct cry patterns associated with different causations would be consistent with the previously quoted data on the meaning and interpretation of infant crying. It could also help explain some of the apparent discrepancies in research on colic. In any event, because parents in our study tended to be dissatisfied with both groups, one still needs to consider both groups in any proposed treatment approach. In order to determine the significance of parameters such as the intensity and frequency of cry, more research separating and controlling for these parameters will be necessary. If it is discovered that a particular pattern of distressing cry is associated with pain of a gastrointestinal origin, then it will be possible to direct our decisions on treatment with more assurance. Until that time, we can only suggest a reasonable clinical approach based on suggestive but limited data.
REVIEW OF AND PROPOSED APPROACH TO TREATMENT Over the years, treatment for colic has fallen into two major categories, those assuming pain (usually of a gastrointestinal origin) and those assuming an interactive causation. Gastrointestinal treatment usually involves either pharmacologic agents or formula changes. Most pharmacologic agents that have been used are either sedatives or affect gastrointestinal motility. Because some of these agents have both central nervous system and gastrointestinal effects, it has been difficult to determine the actual mechanism when successful. Furthermore, the safety of certain of these agents has been called into question recently in terms of a possible association with sudden infant death syndrome. 34, 51
916
M.
ALEX GEERTSMA AND JEFFREY
S.
HYAMS
1. History and Examination
~
2. Diagnosable condition?
Ye~~o Treat
3. What is crying like?
~
(cry log if necessary)
~
4a. Frequent but consolable cry
! !
5a. Behavioral intervention
6a. Trial of mechanical rocking-vibrating device
~
4b. High-intensity inconsolable cry ± gas, cramping
t
5b. Dietary intervention
.
t
Go to 5a
Trial of commercial soy formula
~
Trial of protein hydrolysate formula
1
Go to 5a Figure 1. Clinical approach to colic.
Forsyth et aP6 have documented the frequent and relatively indiscriminant use of formula change and experimentation by clinicians in response to maternal concerns about infant crying. Although the incidence of serious physical complications from such formula manipulation is rare, Forsyth et aP6 have noted the major negative effect this has on the mothers' perceptions of their children. In general, these mothers tend to believe their children are medically vulnerable and that their infants suffer from a medical disease or illness. Interventions directed at a presumed interactive cause have incorporated either general supportive counseling or specific behavioral approaches to the infant's crying. As an example, Taubman 45 has outlined the following stepwise set of instructions for parents to use whenever their infants cry excessively: 1. Never let your baby cry. 2. Consider hunger, the desire to suck (pacifier and so forth), the need to be held, boredom (the need for social stimulation), and fatigue (the need to be put down to sleep) as possible causes and give appropriate responses. 3. If the crying continues more than 5 minutes, try another response. 4. Don't be concerned about overfeeding. 5. Don't be concerned about "spoiling" the baby.
COLIC-A PAIN SYNDROME OF INFANCY?
917
As noted in a previous section, it is unclear if Taubman and others who are successful with behavioral approaches have preselected infants and families most likely to benefit from such interventions. Finally, various commercial rocking and vibr/lting devices claiming effectiveness in calming colicky infants have recently flooded the market. 31 Most are expensive and have not been proven to be consistently effective or theoretically sound. As with other behavioral interventions, it is unclear from limited existing research if such devices are applicable to the broader range of colicky infants who are disturbing to parents. Based on the literature that exists and our discussion of our preliminary findings, we believe the clinical approach to colic outlined in Figure 1 is most judicious. SUMMARY
A review of colic research supports the theory that colic may be caused by pain, particularly pain generated by gastrointestinal causes. At the same time, it also supports the view tha,t disturbing crying in the first 3 months of life may be secondary to behavioral-interactive problems or simple parental misinterpretation of cry. In fact, a closer look at the methodology of colic studies along with our preliminary results suggest there may be at least two different patterns of disturbing infant crying. It is possible that one is associated with true pain and the other not. Dependence on retrospective parental reports alone to substantiate the presence of pain is only as reliable as parent interpretation of crying. Although specific cry patterns may represent specific infant conditions, parents may not be accurate interpreters of those patterns. ' In trying to determine the presence of gastrointestinal pain in colicky crying, close attention therefore should be given to the nature of the cry. Specific attention to qualitative and quantitative aspects of crying in colicky infants is identified as important by our pilot work. In the future this may help to explain t1!e apparent discrepancies in the colic literature and to determine to what extent infant colic is a true pain syndrome. REFERENCES 1. Als H, Tronick E, Adamson L, Brazelton TB: The behavior of the full-term yet underweight newborn infant. Devel Med Child Neurf)1 18:590--602, 1976 2. Auricchio S, Rubino A, Murset G: Intestinal glycosidase activities in the human embryo, fetus, and newborn. Pediatrics 35:944-954, 1965 3. Barr RG, Elias MF: Nursing interval and maternal responsivity: Effect on early infant crying. Pediatrics 81:529-536, 1988 4. BaIT RG, Hanley J, Patterson DK, et al: Breath hydrogen excretion in normal newborn infants in response to usual feeding patterns: Evidence for 'functional lactase insufficiency' be¥.pnd the first month of Hfe. J Fediatr 104:527-533, 1984 5. Barr RG, Kramer MS, BOisjoly C, et al: Parental diary of infant cry and fuss behavior. Arch Dis Child 63:380-387, 1988 6. Bell SM, Ainsworth MDS: Infant crying and maternal responsiveness. Child Devel 43:1171-1190, 1972
918
M.
ALEX GEERTSMA AND JEFFREY
S.
HYAMS
7. Brazelton TB: Crying in infancy. Pediatrics 29:579-588, 1962 8. Carey WB: Clinical applications of infant temperament measurements. J Pediatr 81:823828, 1972 9. Carey WB: 'Colic'-primary excessive crying as an infant-environment interaction. Pediatr Clin North Am 31:993-1004, 1984 10. Carey WB: Maternal anxiety and infantile colic: Is there a relationship? Clin Pediatr 7:590-595, 1968 11. Carey WB: Some pitfalls in infant temperament research. Infant Behav Dev 6:247-254, 1983 12. Clark RL, Ganis FM, Bradford WL: A study of the possible relationship of progesterone to colic. Pediatrics 31:65-71, 1963 13. Deisher RW, Goers SS: A study of early and later introduction of solids into the infant diet. J Pediatr 45:191-199, 1954 14. Evans RW, Allardyce RA, Fergusson DM, et al: Maternal diet and infantile colic in breast-fed infants. Lancet 1:1340-1342, 1981 15. Forsyth BWC, Leventhal JM, McCarthy PL: Mothers' perceptions of problems of feeding and crying behaviors. Am J Dis Child 139:269-272, 1985 16. Forsyth BWC, McCarthy PL, Leventhal JM: Problems of early infancy, formula changes, and mothers' beliefs about their infants. J Pediatr 106:1012-1017, 1985 17. Hide DW, Guyer BM: Prevalence of infant colic. Arch Dis Child 57:559-560,1982 18. Hill DJ, Firer MA, Shelton MJ, et al: Manifestations of milk allergy in infancy: Clinical and immunologic findings. J Pediatr 109:270-276, 1986 19. Hyams JS, Ricci A, Leichtner AM: Clinical and laboratory correlates of esophagitis in young children. J Pediatr Gastroenterol Nutr 7:52-56, 1988 20. Illingworth RS: Infantile colic revisited. Arch Dis Child 60:981-985, 1985 21. Jakobsson I, Lindberg T: Cow's milk proteins cause infantile colic in breast-fed infants: A double blind crossover study. Pediatrics 71:268-271, 1983 22. Jorup S: Colonic hyperperistalsis in neurolabile infants. Acta Paediatr Scand 85(suppl):110,1952 23. Kagan J: Comments on the construct of difficult temperament: A critical analysis. MerrillPalmer Q 28:35-40, 1982 24. Kahn A, Mozin MJ, Casimir G, et al: Insomnia and cow's milk allergy in infants. Pediatrics 76:880-884, 1985 25. Lakin M: Personality factors in mothers of excessively crying (coliCky) infants. Monogr Soc Res Child Dev 22:7-15, 1957 26. Lebenthal E, Lee PC: Development of functional response in human exocrine pancreas. Pediatrics 66:556-560, 1980 27. Lester BM, Zeskind PS: A biobehavioral perspective on crying in infancy. In Fitzgerald HE, Lester BM, Yogman MW (eds): Theory and Research in Behavioral Pediatrics, vol 1. New York, Plenum Press, 1982 28. Levitt MD: Production and excretion of hydrogen gas in man. N Engl J Med 281:122127,1969 29. Levitt MD, Donaldson RM: Use of respiratory hydrogen (H.) el'cretion to detect carbohydrate malabsorption. J Lab Clin Med 75:937-945, 1970 30. Liebman WM: Infantile colic association with lactose and milk intolerance. JAMA 245:732733, 1981 31. Loadman WE, Arnold K, Volmer R, et al: Reducing the symptoms of infant colic by introduction of a vibrationlso!lnd based intervention. Pediatr Res 21:182a, 1987 32. Lothe L, Ivarsson SA, Lindberg T: Motilin, vasoactive intestinal peptide and gastrin in infantile colic. Acta Paediatr Scand 76:316-320, 1987 33. Lothe L, Lindberg T, Jakobsson I: Cow's milk formula as a cause of infantile colic: A double blind crossover study. Pediatrics 70:7-10, 1982 34. Merril Dow Pharmaceuticals: Bulletin-New prescribing information on Bentyl (dicyclomine). Nov. 19, 1984 35. Mobassaleh M, Montgomery RK, Biller JA, et al: Development of carbohydrate absorption in the fetus and neonate. Pediatrics 75(suppl):160-166, 1985 36. Muller E, Hollien H, Murry T: Perceptual response to infant crying: Identification of cry types. J Child Language 1:89-95, 1974 37. Murry T, Amundson P, Hollien H: Acoustical characteristics of infant cries: Fundamental frequency. J Child Language 3:321-328, 1977
COLIc-A PAIN SYNDROME OF INFANCY?
919
38. Murry T, Hollien H, Muller E: Perceptual responses to infant crying: Maternal recognition and sex judgments. J Child Language 2:199-204, 1975 39. Oski FA: Iron-fortified formulas and gastrointestinal symptoms in infants: A controlled study. Pediatrics 66: 168-170, 1980 40. Paradise JL: Maternal and other factors in the etiology of infantile colic. JAMA 197:123131, 1966 41. Ritchie J: Pain from distention of the pelvic colon by inflating a balloon in the irritable colon syndrome. Gut 14:125-132, 1973 42. Rothbart MK: The concept of difficult temperament: A critical analysis. Merrill-Palmer Q 28:35-40, 1982 43. Ruppin H, Bar-Melr S, Soergel KH, et al: Absorption of short-chain fatty acids by the colon. Gastroenterology 78:1500-1507, 1980 44. Sostek AM, Anders TF: Relationship among the Brazelton Neonatal Scale, Bayley Infant Scales, and early temperament. Child Devel 48:320-323, 1977 45. Taubman B: Clinical trial of the treatment of colic by modification of parent-infant interaction. Pediatrics 74:998-1003, 1984 46. Thomas A, Chess S, Birch HG, et al: Behavioral Individuality in Early Childhood. New York, New York University Press, 1963 47. Walker-Smith JA, Ford RPK, Phillips AD: The spectrum of gastrointestinal allergies to food. Ann Allergy 53:629-636, 1984 48. Wasz-Hockert 0, Lind J, Vuorenkoski V, et al: The infant cry, a spectrographic and auditory analysis. Clinics in Developmental Medicine no. 29. London, Heinemann, 1968 49. Wessel MA, Cobb JC, Jackson EB, et al: Paroxysmal fussing in infancy, sometimes called 'colic.' Pediatrics 14:421-424, 1954 50. White PJ: Gastro-enterospasm as a manifestation of autonomic imbalance in early infancy. Am J Dis Child 26:91-102, 1923 51. Williams J, Watkins-jones R: Dicyclomine: Worrying symptoms associated with its use in small babies. Br Med J 288:901, 1984 52. Wolff PH: Crying and vocalization in early infancy. In Foss BM (ed): Determinants of Infant Behavior IV. New York, Wiley Press, 1969 53. Zahorsky J: Mixed feeding of infants. Pediatrics 11:208-215, 1901 Department of Pediatrics St. Francis Hospital and Medical Center 114 Woodland Street Hartford, CT 06105
0031-3955/89 $0.00
+ .20
Colic-A Pain Syndrome of Infancyl M. Alex Geertsma, MD, * and Jeffrey S. Hyams, MDt
ASSUMPTION OF PAIN IN INFANTILE COLIC The clinician is most likely to invoke the diagnosis of infantile colic when faced with the frantic sleep-deprived parent who has been dealing ineffectively with prolonged periods of crying in a less than 3-month-old infant. Parents seek a cure, not only for the alleviation of their sleep deprivation, but also because they often infer that their baby is in pain. Why is it that parents infer pain from the behavioral colic constellation? Physical signs such as pulling up of the legs with apparent abdominal cramping and passing of flatus are often observed with colicky crying. Another reason for the assumption of pain stems from the communicative function of crying. Lester and Zeskind27 discuss the evidence that the infant cry often reflects a specific need state or pain. Once the pain is alleviated or the need state fulfilled by the parent, the communication-the cry-stops. It is therefore understandable that prolonged crying that does not abate with the usual parental trial and error interventions, namely feeding, rocking, holding, verbally soothing, and changing, creates an escalating concern that there is a major unmet need or unalleviated pain. For these reasons, a significant amount of literature concerning colic is directed toward the discovery of an organic or physical cause of the presumed pain. However, there is an equal body of literature that proposes a nonorganic etiology for colic. This perspective, largely behavioral in orientation, proposes the absence of actual pain. Taken as a whole, then, the pediatric colic literature is confusing. Is there evidence for pain in colic? If so, what causes it and how should the clinician approach the pain? Are alternative behavioral and psychosocial explanations justified? If so, how are they to be reconciled with the evidence of pain? Finally, why does this disagreement exist? *Ciinical Associate Professor of Pediatrics, Division of General Pediatrics
t Associate Professor of Pediatrics, Division of Pediatric Gastroenterology From the University of Connecticut Health Center, Farmington; St. Francis Hospital and Medical Center and Hartford Hospital, Hartford, Connecticut
Pediatric Clinics of North America-Vol. 36, No.4, August 1989
905
906
M.
ALEX GEERTSMA AND JEFFREY
S.
HYAMS
This article considers the evidence for infantile colic as a pain syndrome. In order to accomplish this, the literature on parental and observer interpretation of cry and pain signals will be reviewed. Then the formal evidence for pain and its causes in the colic syndrome will be considered, as will alternative behavioral explanations of colic. Finally, an attempt will be made to explain the lack of agreement about colic and to propose a treatment approach.
PARENTAL INTERPRETATION OF INFANT CRYING A major difficulty in studying infantile colic is that the presence of pain versus other causes of crying must be inferred by the observer. Because the young infant cannot use language to communicate, the parent or observer is often dependent on interpreting relatively subjective behavioral signs. Given the tendency to assume pain as a cause for colic, how accurate are parents in assessing the actual meaning of their infant's cry? Work addressing this question appears primarily in the child development literature. WolfI52 observed a small sample of 14 infants and recorded cry in different circumstances including hunger, pain, and cold. He noted that cry spectrographs displayed consistent unique patterns for each of the conditions. He also noted that mothers of the infants were largely able to distinguish the circumstance of crying by sound recording playback alone. Wasz-Hockert et al48 found results similar to Wolff, showing cries unequivocably associated with certain situations (hunger, pain, and so forth) had very specific spectrographic patterns. In contrast to Wolfl's limited study, more detailed work by Muller et aP6 cast doubt on the accuracy of maternal cry interpretation. While mothers of infants could correctly identify their childrens' hunger cries from sound alone, they made the judgment of hunger many other times for spectrographic patterns not characteristic of hunger. That is, they overinterpreted crying as reflective of hunger. Work by others 37• 36 largely confirms the latter two studies. Essentially, the evidence suggests that cry characteristics for a given infant do vary according to the cause of the cry, potentially conveying meaning to the parent. However, parents may misinterpret the cry and may tend to err on the side of a particular presumed etiology such as pain or hunger. This tends to support the view that at least some presumed colicky crying may have different meaning than the parent might assume.
POSSIBLE CAUSES OF PAIN IN COLIC Certainly, defined physical illnesses and structural anomalies can cause mysterious pain and, hence, crying in young infants. Incarcerated hernias, otitis media, fractured clavicles, hematomas secondary to inapparent
COLIc-A PAIN SYNDROME OF INFANCY?
907
trauma, and a multitude of other discernible causes of pain could theoretically cause a young infant to cry vigorously and persistently. However, most would agree that these conditions are usually easily diagnosed. Most would also agree that an essential component of the colic syndrome is the occult nature of its causation. Therefore, such easily determined causes of infant crying should exclude the diagnosis of colic.
GASTROINTESTINAL CAUSES OF COLIC For well over 50 years most physicians have considered abdominal pain to be of primary importance in the colic syndrome. In 1923, White SO used the term gastroenterospasm to describe what he considered autonomic imbalance leading to abdominal pain and excessive crying in infants. Abdominal pain is classically separated into visceral pain and parietal pain. Parietal pain arises from noxious stimuli to the parietal peritoneum and is exacerbated by movement, exactly the opposite from what is observed in the colicky infant who often improves with movement. Visceral pain arises from noxious stimuli affecting an abdominal viscus. It is frequently cramping in nature and often accompanied by autonomic symptoms such as sweating and restlessness. This constellation of findings bears some resemblance to the findings in infants with colic. Visceral pain fibers are generally only sensitive to stretching or tension in the wall of the gut or in the capsule of solid organs. Tension must be increased relatively rapidly for pain to appear. Distension of the gut wall is commonly produced by either intraluminal gas or fluid. This has suggested several possible etiologic factors for the production of colic in infants. Malabsorption and Gas Function When dietary carbohydrate is not completely absorbed by the small intestine, it passes into the colon where it is fermented by colonic bacteria. 28 This process results in the production of short chain organic acids and the evolution of several gases. The short chain organic acids are either absorbed by the colonic mucosa or, if present in excess, act as a cathartic and are passed in the stool. 43 Approximately 15 per cent of the gas produced during this reaction (primarily CO 2 and H 2) is absorbed into the bloodstream and the remainder is passed as flatus. 29 Abdominal cramping is a common accompaniment to carbohydrate malabsorption. Most term infants cannot fully absorb the lactose load contained in formula feedings in the first several months of life despite having "adult" levels of mucosal lactase activity.2, 4, 29, 35 Breath H2 levels are quite high following feedings with either breast milk or lactose-containing formula. 4 One could postulate that colicky infants have increased intraluminal gas production following feedings. Liebman,30 using stool examination for pH and reducing substances, was unable to find a relationship between lactose malabsorption and colic. Unfortunately, the methodology for detecting carbohydrate malabsorption was relatively insensitive and future studies using newer techniques will be required to examine this issue further. Another common dietary carbohydrate that is not completely absorbed
908
M.
ALEX GEERTSMA AND JEFFREY
S.
HYAMS
in the first several months of life is starch. Amylase activity in the pancreas of the term neonate is only 10 per cent of that found in adults; one study found no amylase activity in the duodenal fluid of some infants at birth. 26 Thus, the ingestion of cereal in the first few months of life might pose a theoretic problem to the infant. In one study, the prevalence of colic was found to be greater in a group of infants fed cereal in the first 3 months of life compared to infants receiving no solids. 17 What was not clear from that study was whether the infants were receiving the cereal before the colic developed or, in fact, were receiving it as a potential treatment for persistent crying. Deisher and Goers13 found no difference in the incidence of colic between infants who received cereal in the first month of life compared to those who did not. Thus, it is not clear at the present time if colicky infants have increased intraluminal gas production. In 1952, Jorup22 suggested that no excessive gas could be found using radiologic studies in a group of infants with colic. In adults with irritable bowel syndrome characterized by cramping and bloating, no increased colonic gas can be demonstrated either, but these individuals have increased sensitivity to bowel distention caused by a balloon inflated in the colon. 41 Whether infants with colic may be comparable in having this heightened sensitivity to bowel distention remains to be determined. Additionally, it is not clear if infants with colic differ from non colicky infants in their ability to propulse gas along the gastrointestinal tract. Since a small percentage of swallowed air is not lost through eructation and enters the small intestine, it is possible that colicky infants might swallow more air, expel less gas through ineffective burping, or handle the gas which enters the small bowel less effectively. Gastrointestinal Allergy Although allergy to cow's milk has been well documented to result in a variety of gastrointestinal and nongastrointestinal symptoms in some infants,47 several reports also suggest that intolerance to cow's milk containing formula might be associated with infantile colic. 21 , 33 It is important to note that these studies have focused on the protein components of cow's milk rather than the carbohydrate lactose discussed previously. As long ago as 1901,53 a report appeared suggesting that "proteids" are contained in the breast milk of certain women whose infants have colic and that these substances may cause abdominal cramping. In 1983, Jakobsson and Lindberg21 reported that 50 per cent of breast-fed infants with colic improved when the mother eliminated cow's milk protein from her diet, In 35 per cent the symptoms of colic returned on repeat challenges with cow's milk protein. Lothe et aP3 demonstrated that 18 per cent of 60 colicky infants receiving a cow's milk formula improved when switched to a protein hydrolysate formula, The authors postulated that sensitivities to cow's milk and soy protein were important causes of infantile colic. These findings have not been corroborated in other studies. Evans et aP4 found no improvement when the breast-feeding mothers of 20 infants with colic avoided cow's milk in a double-blind, placebo-controlled, crossover study. Liebman30 found no elevation of total IgE levels or radioallergosorbent test (RAST) to cow's milk protein in 56 infants with colic and
COLIc-A PAIN SYNDROME OF INFANCY?
909
concluded that COW'S milk protein allergy was not a significant cause of infantile colic. However, other reports have suggested that normal IgE and RAST levels may be found in infants with cow's milk allergy if they do not have systemic signs of allergy such as eczema or wheezing. 18 Recently, Kahn et al24 suggested that cow's milk allergy may be a cause of infantile insomnia. They reported eight infants who slept extremely poorly and in whom substitution of a protein hydrolysate formula for one with cow's milk protein normalized sleep patterns. When cow's milk protein was reintroduced into the diet, the infants again developed sleeplessness. All had elevated IgE levels and 50 per cent had elevated RAST levels to beta-lactoglobulin. Most had a history of either wheezing or eczema. Another component of infant formula that has been postulated to lead to colicky symptoms is iron. Many physicians will change from an ironcontaining to an iron-free formula in the infant with fussiness, diarrhea, or constipation. However, Oski39 could show no difference in gastrointestinal symptoms between two similar groups of infants fed either an iron-fortified or a conventional cow's milk formula. Gastrointestinal Reflux Spitting and vomiting are common in the first several months of life and only uncommonly lead to significant problems. Yet, a small number of these infants will also manifest extreme irritability and raise the possibility of pathologic gastrointestinal reflux with esophagitis. 19 Clinically, these infants often awaken from sleep with excessive crying, particularly 1 to 2 hours after a feeding. The establishment of a relationship between irritability and gastrointestinal reflux in an infant may be difficult. Therapeutic trials of antacids during these crying spells may be attempted. In the infant with intractable crying, performance of prolonged intraesophageal pH probe monitoring may be helpful. If the clinician can document a fall in intraesophageal pH preceding periods of irritability, a causal relationship can be presumed. Hormonal Abnormalities Since abnormalities in intestinal peristalsis have been presumed to be important in the pathogenesis of infantile colic, it is natural that some attention has focused on gastrointestinal hormones that regulate intestinal motility. In 1963, Clark et aP2 suggested that infants with colic may be deficient in progesterone, a hormone which may relax smooth muscle. When colicky infants were given progesterone, their symptoms improved. Unfortunately, this study was not performed in a blinded or controlled manner. More recently, Lothe et aP2 have evaluated the levels of motilin, vasoactive intestinal peptide (VIP), and gastrin in infants with colic as well as in a group of age-matched controls. No differences were found in VIP and gastrin levels, but the infants with colic had a higher mean motilin level than did noncolicky infants. Motilin has a stimulating effect on gastrointestinal motility. Whether this elevation is a primary contributor to colic or rather a secondary phenomenon remains to be established. In summary, although pain, especially gastrointestinal pain, remains a
910
M.
ALEX GEERTSMA AND JEFFREY
S.
HYAMS
common explanation for colicky infant behavior, it is evident from the preceding discussion that the evidence, at present, is not conclusive.
EVIDENCE FOR BEHAVIOR AND INTERACTION CAUSING COLIC Behavioral and interactive factors have also been suggested as causes of colicky crying. This view holds that there is in fact no physical cause of colic and hence no true pain. Instead, it is postulated that the infant becomes distraught and cries due to problems with interaction and handling by the environment or simply due to intrinsically difficult infant temperament. Intrinsic Temperament and Colic While the concept of temperament and temperamental types has gained some popularity in pediatrics,46 it remains controversial in research circles, especially the assumption of stability over timeY Nevertheless, it is common to assume that the difficult child was colicky as a young infant and, in turn, that the colicky infant usually becomes a difficult temperament older child. Actually, very little exists in the pediatric literature directly addressing stability of temperament or the relationship between colic and subsequent temperament. What does exist is either inconclusive or refutes the suggested relationship. In 1972 Careys attempted to correlate scores on his temperament scale with various clinical conditions including colic. He claimed to show a high correlation between colic documented from office records and difficult temperament as recorded by parental assessments at 4 and 8 months. This conclusion must be viewed cautiously because of the possibility of parental report bias. Parents who experience difficult, colicky young infants might be expected to see subsequent, relatively normal infant behavior as still difficult. In general, the objectivity of parental assessments of temperament is a major point of contention in the research literature. 23, 42 Perhaps due to the problems of parental report reliability and the methodologic problems of trying to avoid them, very little has been published addreSSing the possible relationship between colic and temperament. At best, limited associations have been suggested although the overall results have been inconclusive or negative. 1. 44 Maternal Caretaker Maladaptation Another commonly held view is that colic is caused by maternal caretaker maladaptation or psychological problems. This view originated from episodic clinical observations, published editorial opinion, and older psychiatric case studies noting that a change of environment and caretaker would at times lead to a change in the difficult crying. Specific mechanisms were often unclear. However, an inability or unwillingness of the mother to effectively calm or nurture the crying infant was generally assumed. Lakin,25 for example, in a retrospective interview study, described
COLIc-A PAIN SYNDROME OF INFANCY?
911
mothers of colicky infants as having less adequate marital adjustment, poorer parent-child relationships, and a greater sense of inadequacy as women. In a much quoted early study, Wessel et al 49 claimed to find a high degree of family tension in homes with colicky infants. In both studies, retrospective design leaves open the alternative interpretation that family tension and maternal insecurity are the products of, rather than the causes of, colicky infant crying. Using a prospective assessment of maternal anxiety, Carey9 noted an association between maternal anxiety and colic. However, colic was documented via office records making it uncertain if maternal anxiety simply led to inflated reports of crying. Paradise 40 studied 146 newborns and their families. He documented maternal adjustment and personality factors prospectively using objective and reproducible measures such as the Minnesota Multiphasic Personality Inventory (MMPI). Paradise noted that there was no evidence to support the premise that colic somehow arises from "an unfavorable emotional climate created by an inexperienced, anxious, hostile, or unmotherly mother." The majority of these early behavioral-interactive studies suffered from similar methodologic problems, namely, potential report biasing, retrospective data, imprecise measures, and small study populations. What has been characteristic of behavioral pediatric studies in the 1980s has been the use of sophisticated forms of statistical analysis, computer implementation, and refined research design. Forsyth et al's prospective study 15 of 373 mothers found that colic and excessive cry reports more often came from mothers who displayed one or a combination of a number of psychologic risk factors. Unfortunately, criteria used to qualify for colic were diffuse and evidently heavily dependent on general parental impression. It is, therefore, unclear whether Forsyth et al were documenting an association between psychologic risk factors and the actual occurrence of difficult crying or simply negative parental impressions of infant behavior. If one believes that all colic can be explained by psychosocial and behavioral factors, then it stands to reason that one should be able to intervene and decrease colic by behaviorally oriented intervention. Taubman 45 tested this theory by giving parents of colicky infants a detailed stepwise behavioral intervention to use with their colicky infants. He gave them a rigid set of instructions of what to do whenever their infants cried. He reported a significant decrease in total crying versus a control group. Unfortunately, the study, while impressive in its success, shed little light on the actual mechanism by which the intervention was successful. Therefore, we have no better idea whether colic in his study population was due to interactional problems, problems with maternal interpretation of cry, or ill-defined psychosocial instability. Although not studying colic per se, Barr and Elias 3 came to similar conclusions. They noted that more frequent "contingent" feeding resulted in less fretful crying than longer intervals between feeding generated by a more defined feeding schedule. This is consistent with the classical child development research on crying performed by Bell and Ainsworth. 6 In summary, much like the research on organic causes of colic, the
912
M.
ALEX GEERTSMA AND JEFFREY
S.
HYAMS
behavioral and interactive literature also appears to be inconclusive and contradictory. Therefore, it is not surprising that clinicians and parents alike remain uncertain about how to approach disturbing infant cryingwhether it is reflective of pain, a pattern of normal but misinterpreted crying, or indicative of an interactive disorder. Various writers have speculated on the reasons for the lack of helpful research. 9. 20 Almost all critical reviewers recommend better research, better methodology, and less dependence on clinical impression and anecdotal reports. However, few have actually critiqued the methodology of existing research. Doing so may not only suggest better research design, but also insights into more effective treatment. PROBLEMS WITH DEFINITION AND METHODOLOGY Detailed reviews of both organic and behavioral colic studies reveal several important methodologic points. One broad methodologic deficiency is the dependence on retrospective data. As noted in our discussions, various organic and behavioral studies have used retrospective reports of colic and crying. This approach, along with restrictive definitions, can overly limit the population being studied. It can, in turn, lead to mistaken conclusions about the applicability of the findings. As Carey9 points out, there is also a lack of agreement concerning the definition of colic. In a number of studies discussed in this article, the definitions used were very imprecise or not described in sufficient detail. For example, in Forsyth et aI's study, 15 the presence of colic was determined primarily by impressionistic parental reporting. Other studies have utilized more precise definitions dependent on less subjective information. However, inclusion criteria generated by those definitions may have resulted in overly broad definitions and, hence, excessively large study populations. As an example, Taubman 45 and others have used Wessel et aI's definition of colic. This included a very specific criterion of 3 or more hours of crying per day for more than 3 days in any 1 week. Using this criterion himself, Wessel et al49 identified 49 per cent of a normal nursery population as eventually colicky, clearly far too high a number, thereby raising doubts about the validity of the definition. In fact, Brazelton, 7 in a prospective cry study, documented that the median amount of crying at 6 weeks is 2% hours total per day. Implicit in the extensive work on the spectrographic analysis of crying is the possibility that the severity of cry not only relates to the potential degree of generated parental distress, but also the likely meaning of the cry. It is not inconceivable that a louder, more persistent cry might indicate pain, whereas less intense but frequent crying may indicate something else, such as inadequate parental response to infant needs or simply immature sleep cycling. In many studies-both those supporting pain-gastrointestinal causes and those supporting behavioral-interactive etiologies-little is done to objectifY either the amount or the character of the crying reported. Neither
913
COLIc-A PAIN SYNDROME OF INFANCY?
CareylO nor Paradise 40 required recording of actual cry periods. Instead, they depended on office records of parental impressions of the severity of cry. Similarly, Jakobsson and Lindberg's study21 and Lothe et aI's hormonal study32 did not require quantitative cry records. It is true that some work does pay attention to estimates of the severity and quality of crying, although not requiring precise measures of the severity. Other studies choose to totally ignore certain aspects of infant crying, thereby ignoring possible clues to the etiology of the crying. In general, studies that claim to prove gastrointestinal causes for colic tend to study primarily those children with intense persistent crying and associated gastrointestinal symptoms. Those that support a behavioralinteractive etiology tend to use either diffuse assessments of the quality of cry or none at all. Either approach, given the possibility that quality of crying and associated symptoms might convey meaning, could lead to spurious or biased results favoring one clinical approach over another (Table 1). It is quite likely that this inconsistency, along with dependence on retrospective methodology and use of imprecise definitions, has contributed significantly to the confUSing and apparently contradictory results of colic research.
A PROSPECTIVE STUDY OF COLIC-PATTERNS OF CRY AND RELEVANCE FOR PAIN IN COLIC In view of the methodologic problems in colic research and to better answer the question of the presence of pain and an organic etiology, we undertook a pilot prospective study of crying and colicky behavior in newborns. We asked each of 60 families to record at between 2 and 4 weeks and Table 1. Definitions and Presumed Etiology of Colic DEFINITION
Gastrointestinal Studies Jakobsson and Lindberg (1983) Lothe et al (1982) Lothe et al (1987) Liebman (1981)
Behavior-Interaction Studies Taubman (1984) Wessel et al (1954) Carey (1968) Forsyth et al (1985)
Persistent cry, abdominal distension, flatus Persistent cry, abdominal distension, flatus Intense cry, abdominal distension, flatus 2:3 hr/day crying, no gas symptoms
Retrospective parent report 2:3 hr/day crying
CAUSE
Cow's milk protein Cow's milk protein Increased s-motilin levels No evidence of allergy or malabsorption
Parental misinterpretation of cry >3 hr/day crying for 3 days/week Family tension Retrospective parent report 2:3 Maternal anxiety hr/day crying Retrospective parental impression Maternal psychosocial of crying issues
914
M.
ALEX GEERTSMA AND JEFFREY
S.
HYAMS
again at between 6 and 8 weeks a detailed log of infant crying, associated infant behaviors, and parental responses. The use of infant behavior logs has been studied formally and proven to be a reliable source of behavioral information. 5 Such logs make the systematic collection of extensive behavioral data from a large study population feasible. Furthermore, methodology studies confirm that parents can be accurate observers and reporters if they are asked to limit their observations to specific circumscribed parameters. This is in distinction to retrospective interview data that are dependent on impressionistic recollections of parents. In the case of our study, parents were carefully instructed to record clear assessments of whether their child cried during a given hour and semiquantitative estimates of the amount of crying that occurred each hour. Parents also recorded the actual presence or absence of flatus passage during the same observation periods. The global opinion of parents about their infants' crying is of clinical importance because it is parental distress about the crying that the clinician must address. Therefore, we also asked parents to record the degree of satisfaction with their babies' crying. We accomplished this by asking them to fill out an infant behavior assessment prior to each cry log. Preliminary analysis of the data revealed patterns of great potential relevance to our earlier discussion of past colic research and possible treatment approaches. First, different patterns of crying were evident from analysis of the cry logs. One distinct pattern was noted in infants who cried intensively; that is, when they cried, they cried for most of the hour, regardless of what parents did. This intense crying group also passed flatus significantly more than other infants. We called these babies "classic colic" infants. They made up approximately 9 per cent of the prospective study group. Another pattern was observed in infants who did not cry intensively, but who calmed quickly and stopped crying shortly after their parents intervened with them. They tended to cry frequently, displaying low intensity crying in many of the 24-hour periods on the days observed. We labeled these babies "frequent cry" infants. They made up 10 to 11 per cent of the study population. The remainder of the babies were termed "noncolic" infants. Parental distress and dissatisfaction with infant crying were significantly greater for the "classic colic" intense criers. Parental dissatisfaction was least for the "non colic" babies, whereas dissatisfaction with the "frequent cry" group was significantly greater than the "noncolics" but less than the "classic colics." This preliminary study raises important questions concerning past and future colic research. First, it may be that closer, detailed attention must be given to specific patterns of crying in colic research. Simply summing the amount of crying per 24 hours may not be precise enough. The intensity of cry (the inverse of "soothability") may need to be distinguished from frequency of cry. In studies using Wessel et aI's criterion both types of crying may be lumped together. A child who cries intensely and inconsolably with passage of flatus may be grouped together with infants who cry frequently but are
COLIc-A PAIN SYNDROME OF INFANCY?
915
easily consoled. Moreover, using the criterion of 3 hours of crying per day may define even more infants as colicky than just classic colic and frequent cry infants. In Wessel et al's study49 as much as 49 per cent of the population was defined as colicky. Conversely, the intensely crying infant with signs of gassiness, even though very distressing to a parent, may be excluded altogether from an intervention study if the crying does not total 3 or more hours per day. This may leave frequent but consolable criers and exclude some intense criers from the study. These groups of infants may be quite different from each other, but, as demonstrated in our study, still distressing to their parents. If a given intervention, such as behavioral parent training, is successful with the remaining group, it may be misleading to suggest that the intervention is reliable for all infants whom parents consider difficult criers. It is also possible that these distinct cry groups may have different causes for their crying. It is not inconceivable, for instance, that the high intensity, high flatus "classic colic" infants have true pain from an organic cause, perhaps gastrointestinal in origin. It is possible that the frequent but easily consolable crier group-the "frequent crier" group-is a group of infants who awaken and initiate crying frequently with no true organic causes. These infants may well be more responsive to changes in parental caretaking behavior. Having two relatively distinct cry patterns associated with different causations would be consistent with the previously quoted data on the meaning and interpretation of infant crying. It could also help explain some of the apparent discrepancies in research on colic. In any event, because parents in our study tended to be dissatisfied with both groups, one still needs to consider both groups in any proposed treatment approach. In order to determine the significance of parameters such as the intensity and frequency of cry, more research separating and controlling for these parameters will be necessary. If it is discovered that a particular pattern of distressing cry is associated with pain of a gastrointestinal origin, then it will be possible to direct our decisions on treatment with more assurance. Until that time, we can only suggest a reasonable clinical approach based on suggestive but limited data.
REVIEW OF AND PROPOSED APPROACH TO TREATMENT Over the years, treatment for colic has fallen into two major categories, those assuming pain (usually of a gastrointestinal origin) and those assuming an interactive causation. Gastrointestinal treatment usually involves either pharmacologic agents or formula changes. Most pharmacologic agents that have been used are either sedatives or affect gastrointestinal motility. Because some of these agents have both central nervous system and gastrointestinal effects, it has been difficult to determine the actual mechanism when successful. Furthermore, the safety of certain of these agents has been called into question recently in terms of a possible association with sudden infant death syndrome. 34, 51
916
M.
ALEX GEERTSMA AND JEFFREY
S.
HYAMS
1. History and Examination
~
2. Diagnosable condition?
Ye~~o Treat
3. What is crying like?
~
(cry log if necessary)
~
4a. Frequent but consolable cry
! !
5a. Behavioral intervention
6a. Trial of mechanical rocking-vibrating device
~
4b. High-intensity inconsolable cry ± gas, cramping
t
5b. Dietary intervention
.
t
Go to 5a
Trial of commercial soy formula
~
Trial of protein hydrolysate formula
1
Go to 5a Figure 1. Clinical approach to colic.
Forsyth et aP6 have documented the frequent and relatively indiscriminant use of formula change and experimentation by clinicians in response to maternal concerns about infant crying. Although the incidence of serious physical complications from such formula manipulation is rare, Forsyth et aP6 have noted the major negative effect this has on the mothers' perceptions of their children. In general, these mothers tend to believe their children are medically vulnerable and that their infants suffer from a medical disease or illness. Interventions directed at a presumed interactive cause have incorporated either general supportive counseling or specific behavioral approaches to the infant's crying. As an example, Taubman 45 has outlined the following stepwise set of instructions for parents to use whenever their infants cry excessively: 1. Never let your baby cry. 2. Consider hunger, the desire to suck (pacifier and so forth), the need to be held, boredom (the need for social stimulation), and fatigue (the need to be put down to sleep) as possible causes and give appropriate responses. 3. If the crying continues more than 5 minutes, try another response. 4. Don't be concerned about overfeeding. 5. Don't be concerned about "spoiling" the baby.
COLIC-A PAIN SYNDROME OF INFANCY?
917
As noted in a previous section, it is unclear if Taubman and others who are successful with behavioral approaches have preselected infants and families most likely to benefit from such interventions. Finally, various commercial rocking and vibr/lting devices claiming effectiveness in calming colicky infants have recently flooded the market. 31 Most are expensive and have not been proven to be consistently effective or theoretically sound. As with other behavioral interventions, it is unclear from limited existing research if such devices are applicable to the broader range of colicky infants who are disturbing to parents. Based on the literature that exists and our discussion of our preliminary findings, we believe the clinical approach to colic outlined in Figure 1 is most judicious. SUMMARY
A review of colic research supports the theory that colic may be caused by pain, particularly pain generated by gastrointestinal causes. At the same time, it also supports the view tha,t disturbing crying in the first 3 months of life may be secondary to behavioral-interactive problems or simple parental misinterpretation of cry. In fact, a closer look at the methodology of colic studies along with our preliminary results suggest there may be at least two different patterns of disturbing infant crying. It is possible that one is associated with true pain and the other not. Dependence on retrospective parental reports alone to substantiate the presence of pain is only as reliable as parent interpretation of crying. Although specific cry patterns may represent specific infant conditions, parents may not be accurate interpreters of those patterns. ' In trying to determine the presence of gastrointestinal pain in colicky crying, close attention therefore should be given to the nature of the cry. Specific attention to qualitative and quantitative aspects of crying in colicky infants is identified as important by our pilot work. In the future this may help to explain t1!e apparent discrepancies in the colic literature and to determine to what extent infant colic is a true pain syndrome. REFERENCES 1. Als H, Tronick E, Adamson L, Brazelton TB: The behavior of the full-term yet underweight newborn infant. Devel Med Child Neurf)1 18:590--602, 1976 2. Auricchio S, Rubino A, Murset G: Intestinal glycosidase activities in the human embryo, fetus, and newborn. Pediatrics 35:944-954, 1965 3. Barr RG, Elias MF: Nursing interval and maternal responsivity: Effect on early infant crying. Pediatrics 81:529-536, 1988 4. BaIT RG, Hanley J, Patterson DK, et al: Breath hydrogen excretion in normal newborn infants in response to usual feeding patterns: Evidence for 'functional lactase insufficiency' be¥.pnd the first month of Hfe. J Fediatr 104:527-533, 1984 5. Barr RG, Kramer MS, BOisjoly C, et al: Parental diary of infant cry and fuss behavior. Arch Dis Child 63:380-387, 1988 6. Bell SM, Ainsworth MDS: Infant crying and maternal responsiveness. Child Devel 43:1171-1190, 1972
918
M.
ALEX GEERTSMA AND JEFFREY
S.
HYAMS
7. Brazelton TB: Crying in infancy. Pediatrics 29:579-588, 1962 8. Carey WB: Clinical applications of infant temperament measurements. J Pediatr 81:823828, 1972 9. Carey WB: 'Colic'-primary excessive crying as an infant-environment interaction. Pediatr Clin North Am 31:993-1004, 1984 10. Carey WB: Maternal anxiety and infantile colic: Is there a relationship? Clin Pediatr 7:590-595, 1968 11. Carey WB: Some pitfalls in infant temperament research. Infant Behav Dev 6:247-254, 1983 12. Clark RL, Ganis FM, Bradford WL: A study of the possible relationship of progesterone to colic. Pediatrics 31:65-71, 1963 13. Deisher RW, Goers SS: A study of early and later introduction of solids into the infant diet. J Pediatr 45:191-199, 1954 14. Evans RW, Allardyce RA, Fergusson DM, et al: Maternal diet and infantile colic in breast-fed infants. Lancet 1:1340-1342, 1981 15. Forsyth BWC, Leventhal JM, McCarthy PL: Mothers' perceptions of problems of feeding and crying behaviors. Am J Dis Child 139:269-272, 1985 16. Forsyth BWC, McCarthy PL, Leventhal JM: Problems of early infancy, formula changes, and mothers' beliefs about their infants. J Pediatr 106:1012-1017, 1985 17. Hide DW, Guyer BM: Prevalence of infant colic. Arch Dis Child 57:559-560,1982 18. Hill DJ, Firer MA, Shelton MJ, et al: Manifestations of milk allergy in infancy: Clinical and immunologic findings. J Pediatr 109:270-276, 1986 19. Hyams JS, Ricci A, Leichtner AM: Clinical and laboratory correlates of esophagitis in young children. J Pediatr Gastroenterol Nutr 7:52-56, 1988 20. Illingworth RS: Infantile colic revisited. Arch Dis Child 60:981-985, 1985 21. Jakobsson I, Lindberg T: Cow's milk proteins cause infantile colic in breast-fed infants: A double blind crossover study. Pediatrics 71:268-271, 1983 22. Jorup S: Colonic hyperperistalsis in neurolabile infants. Acta Paediatr Scand 85(suppl):110,1952 23. Kagan J: Comments on the construct of difficult temperament: A critical analysis. MerrillPalmer Q 28:35-40, 1982 24. Kahn A, Mozin MJ, Casimir G, et al: Insomnia and cow's milk allergy in infants. Pediatrics 76:880-884, 1985 25. Lakin M: Personality factors in mothers of excessively crying (coliCky) infants. Monogr Soc Res Child Dev 22:7-15, 1957 26. Lebenthal E, Lee PC: Development of functional response in human exocrine pancreas. Pediatrics 66:556-560, 1980 27. Lester BM, Zeskind PS: A biobehavioral perspective on crying in infancy. In Fitzgerald HE, Lester BM, Yogman MW (eds): Theory and Research in Behavioral Pediatrics, vol 1. New York, Plenum Press, 1982 28. Levitt MD: Production and excretion of hydrogen gas in man. N Engl J Med 281:122127,1969 29. Levitt MD, Donaldson RM: Use of respiratory hydrogen (H.) el'cretion to detect carbohydrate malabsorption. J Lab Clin Med 75:937-945, 1970 30. Liebman WM: Infantile colic association with lactose and milk intolerance. JAMA 245:732733, 1981 31. Loadman WE, Arnold K, Volmer R, et al: Reducing the symptoms of infant colic by introduction of a vibrationlso!lnd based intervention. Pediatr Res 21:182a, 1987 32. Lothe L, Ivarsson SA, Lindberg T: Motilin, vasoactive intestinal peptide and gastrin in infantile colic. Acta Paediatr Scand 76:316-320, 1987 33. Lothe L, Lindberg T, Jakobsson I: Cow's milk formula as a cause of infantile colic: A double blind crossover study. Pediatrics 70:7-10, 1982 34. Merril Dow Pharmaceuticals: Bulletin-New prescribing information on Bentyl (dicyclomine). Nov. 19, 1984 35. Mobassaleh M, Montgomery RK, Biller JA, et al: Development of carbohydrate absorption in the fetus and neonate. Pediatrics 75(suppl):160-166, 1985 36. Muller E, Hollien H, Murry T: Perceptual response to infant crying: Identification of cry types. J Child Language 1:89-95, 1974 37. Murry T, Amundson P, Hollien H: Acoustical characteristics of infant cries: Fundamental frequency. J Child Language 3:321-328, 1977
COLIc-A PAIN SYNDROME OF INFANCY?
919
38. Murry T, Hollien H, Muller E: Perceptual responses to infant crying: Maternal recognition and sex judgments. J Child Language 2:199-204, 1975 39. Oski FA: Iron-fortified formulas and gastrointestinal symptoms in infants: A controlled study. Pediatrics 66: 168-170, 1980 40. Paradise JL: Maternal and other factors in the etiology of infantile colic. JAMA 197:123131, 1966 41. Ritchie J: Pain from distention of the pelvic colon by inflating a balloon in the irritable colon syndrome. Gut 14:125-132, 1973 42. Rothbart MK: The concept of difficult temperament: A critical analysis. Merrill-Palmer Q 28:35-40, 1982 43. Ruppin H, Bar-Melr S, Soergel KH, et al: Absorption of short-chain fatty acids by the colon. Gastroenterology 78:1500-1507, 1980 44. Sostek AM, Anders TF: Relationship among the Brazelton Neonatal Scale, Bayley Infant Scales, and early temperament. Child Devel 48:320-323, 1977 45. Taubman B: Clinical trial of the treatment of colic by modification of parent-infant interaction. Pediatrics 74:998-1003, 1984 46. Thomas A, Chess S, Birch HG, et al: Behavioral Individuality in Early Childhood. New York, New York University Press, 1963 47. Walker-Smith JA, Ford RPK, Phillips AD: The spectrum of gastrointestinal allergies to food. Ann Allergy 53:629-636, 1984 48. Wasz-Hockert 0, Lind J, Vuorenkoski V, et al: The infant cry, a spectrographic and auditory analysis. Clinics in Developmental Medicine no. 29. London, Heinemann, 1968 49. Wessel MA, Cobb JC, Jackson EB, et al: Paroxysmal fussing in infancy, sometimes called 'colic.' Pediatrics 14:421-424, 1954 50. White PJ: Gastro-enterospasm as a manifestation of autonomic imbalance in early infancy. Am J Dis Child 26:91-102, 1923 51. Williams J, Watkins-jones R: Dicyclomine: Worrying symptoms associated with its use in small babies. Br Med J 288:901, 1984 52. Wolff PH: Crying and vocalization in early infancy. In Foss BM (ed): Determinants of Infant Behavior IV. New York, Wiley Press, 1969 53. Zahorsky J: Mixed feeding of infants. Pediatrics 11:208-215, 1901 Department of Pediatrics St. Francis Hospital and Medical Center 114 Woodland Street Hartford, CT 06105