Recurrent abdominal pain is one of the most commonly encountered symptoms in childhood and adolescence. More than 90% of children with chronic abdominal pain were diagnosed as having a “functional gastrointestinal disorder”(1,2). This relatively nonspecific diagnosis assumes the absence of associated structural, infectious, inflammatory, or biochemical findings (3–5). An abnormality of gastrointestinal motility has been hypothesized (6–8). Whether these are abnormalities of the enteric nervous system or of the extrinsic autonomic nervous system is unanswered.
At least in adults with irritable bowel syndrome (IBS), which may bear some pathophysiologic relation to functional abdominal pain in children, generalized abnormalities of autonomic function have been found (9,10). The purpose of this study was to determine whether similar generalized autonomic dysfunction is also present in children with functional abdominal pain. We tested the autonomic nervous system prospectively in eight children with functional abdominal pain (FAP), evaluating cholinergic parasympathetic, adrenergic sympathetic, and cholinergic sympathetic functions.
Eight children, aged 10 to 17 years (mean, 12.6 years), six girls and two boys, with recurrent abdominal pain (at least three bouts of abdominal pain severe enough to affect activities over a period of at least 3 months) (1) were recruited from the Rainbow Babies and Children's Pediatric Gastroenterology clinic. Inclusion criteria for the study included normal physical examination results, guaiac-negative stools, and normal results of endoscopic, radiologic, and biochemical studies when performed as part of the routine evaluation of abdominal pain. An autonomic and gastrointestinal symptom questionnaire was filled out at the time of autonomic testing. In the gastrointestinal questionnaire, the symptoms were rated by the guardian or child from 0 (no pain) to 5 (severe pain). The University Hospitals of Cleveland Institutional Review Board approved the study. Informed consent was obtained from the legal guardian.
To avoid dehydration, autonomic testing was performed after the participants a usual meal. Three tests of cardiovascular function and two tests of sudomotor function were performed in all participants (11). The normal pediatric values for cardiovascular autonomic tests obtained by others (12) were validated in nine healthy children in our laboratory. Our results matched those of previously published findings. We were unable to find enough healthy children willing to undergo the sweat testing to draw meaningful conclusions regarding these tests and, therefore, used the published norms, which are based on identical testing conditions. Heart rate (HR) response to deep breathing and to the Valsalva maneuver assessed parasympathetic cholinergic function. Tilt-table testing and the Valsalva maneuver evaluated sympathetic adrenergic function. The quantitative sudomotor axon-reflex test (QSART) and thermoregulatory sweat test (TST) tested cholinergic sympathetic function (13).
During deep breathing, participants breathed smoothly and deeply 6 times/min. The average HR variation was calculated from the best 6 of 12 respiratory cycles (normal values in this age group, > 22 beats/min) (12). In the Valsalva maneuver, HR was electrocardiographically monitored while the subject blew into a tube connected to a mercury manometer to maintain a 40-mm Hg pressure gradient above atmospheric pressure for 15 seconds. The Valsalva ratio is the ratio of the fastest HR during the 15 seconds of pressure exertion (termed phase II, sympathetically mediated) to the lowest HR after pressure release (termed phase IV, parasympathetically mediated; normal values for this age group, >1.6 beats/min) (12). For the tilt-table test, the subject lay supine for 20 minutes on a motorized tilt-table before a rapid upright tilt to 90° in less than 5 seconds. Continuous blood pressure and HR were monitored noninvasively with the patient in the supine position for 3 minutes and upright for 10 minutes. Diastolic blood pressure normally decreases less than 10 mm Hg, the systolic blood pressure, less than 20 mm Hg, and HR should increase less than 28 beats/min (14).
Sweat output response to the iontophoresis of acetylcholine (10% with a 2 mA current for 5 minutes, recording sweat output for 10 minutes, “QSART”) across the skin of the feet, calves, hands, upper arms, and forehead was measured by standard methods (15). Normal values vary by body site and gender. For axon-reflex sweating to be considered abnormal, more than one site must be outside the range of normal. Thermoregulatory sweat testing was performed using the standard technique. The participant's skin was coated with a powder containing Alizarin red-S, which changes color on contact with sweat (water). The subject was placed supine in a thermal environment (50°C and 50% humidity) until the oral temperature increased by 2°C or to a maximum of 38.5°C, or for a maximum of 30 minutes (modified from Fealey et al. (16)). Overhead infrared heat lamps maintained skin temperature between 39.5°C and 40.5°C. A photograph of the patient was scanned and digitized. The ratio of sweat area to body surface area was calculated using Image software (National Institutes of Health, Bethesda, MD).
The clinical interpretation of the combined sweat test results (listed in the last column of Table 1) is based on the pattern of reduction in the axon-reflex test and the geographic sweat distribution in the thermoregulatory sweat test. A consistent distal-to-proximal gradient of sweat loss in both tests was interpreted as peripheral neuropathy. Patchy anhidrosis in the thermoregulatory sweat test with nondistal loss of output in the axon reflex resulted in a diagnosis of radiculopathy. Finally, thermoregulatory anhidrosis with normal axon-reflex sweating was thought to suggest a central process.
The gastrointestinal symptoms are summarized in Table 2. No particular relations or patterns emerged across participants. Four patients reported epigastric pain with dyspepsia, three had diffuse abdominal pain with altered bowel pattern, and one had isolated periumbilical pain. None of the participants reported well-localized pain or pain radiation. Two reported daily abdominal pain and six had sporadic reports. Four participants had significant school absenteeism. The children had no history of weight loss or growth deceleration, joint pains, skin rashes of unknown origin, mouth ulcers, unexplained fever, bloody stools, pain awakening them from sleep, or consistent sleepiness after the pain attacks. Four patients demonstrated lactose intolerance, determined by lactose hydrogen testing, but their symptoms did not improve with lactose withdrawal. In none of the participants were antroduodenal motility studies or gastric emptying tests ordered by the treating physician. Fifty percent of the girls were prepubertal. Pelvic ultrasound or gynecologic referral was done in two of the three menstruating girls (the other girl did not pursue recommended follow-up) and in one of the three prepubertal girls. These evaluations did not elucidate the cause of the abdominal pain in any patient. Five of the patients underwent psychologic treatment and evaluation, without improvement. In addition, five of the eight patients received H2 blockers or proton-pump inhibitors without benefit; prokinetic agents provided no relief in two of these five.
Table 3 lists the autonomic symptoms. No patient spontaneously reported autonomic symptoms. On the autonomic systems review questionnaire, five of eight participants mentioned specific problems, with four involving orthostasis or flushing. In one patient (AP) the symptoms were extensive.
The results of autonomic cardiovascular testing in patients and controls are shown in Tables 4 and 5. Axon-reflex sweating and thermoregulatory sweating are shown in Figure 1. Orthostatic intolerance was common (six of eight participants) and was reflected as excessive heart rate increase or blood pressure decrease during the tilt test (Table 4). In two patients, the Valsalva ratio was low for age. In both of these patients, the phase II tachycardia (sympathetically mediated) seemed to be lower than that of the other participants, whereas the phase IV bradycardia (parasympathetically mediated) was comparable. The low Valsalva ratio therefore probably arises from an inadequate cardiac sympathetic response, with parasympathetic function preserved. Cardiac response to deep breathing, another marker of cardiac parasympathetic function, was normal in all participants.
Sweat testing results were abnormal in seven of eight participants (Table 1). In five participants, a peripheral nervous system process (small-fiber neuropathy or radiculopathy) was most likely, whereas, in two others, a central nervous system process (at the spinal cord level or higher) appeared more probable (Table 6). In one participant (AP), the sweating pattern was equivocal, either consistent with a “light sweater”(16) or radicular anhidrosis caused by slight extension of anhidrosis into the left anterior calf. Interestingly, in all participants with significant thermoregulatory anhidrosis, the abdomen was always involved.
Generalized autonomic dysfunction occurred in seven of eight children with functional abdominal pain in this study. Sympathetic function was reduced in every system assessed, including cardiac (reduced tachycardia in phase II of the Valsalva maneuver in some participants), vasomotor (postural tachycardia in most participants), and sudomotor (reduced sweating in most participants). In contrast, parasympathetic function was normal, as assessed by the cardiac response to deep breathing and the bradycardia in phase IV of the Valsalva maneuver.
The testing methods used in this study comprise the primary method of assessment of parasympathetic and sympathetic functions in the diagnosis of autonomic failure in neurologic practice (16,17). Using similar methods, studies have repeatedly found abnormal cardiac parasympathetic function early in classic neuropathies, such as diabetes mellitus (18). Sympathetic cardiac and vasomotor abnormalities usually follow later in the course of illness. This sequence is compatible with a nerve length–dependent process, with the longest nerves being affected earliest, as one expects in degenerative and metabolic neuropathies. The findings in the current study are completely opposite. Sympathetic function is reduced, whereas parasympathetic function is preserved. Hence, a nerve length–dependent process is unlikely.
The most likely explanation is a constitutional reduction in sympathetic overflow. These data do not discern whether this reduction arises from a similar physiology in all of the patients or whether differing underlying abnormalities led to a similar expression. Also, we cannot tell whether this reduction in sympathetic outflow directly produced the gastrointestinal symptoms or whether this observation simply constituted an association. Pain and anxiety are unlikely contributors to our results because one would have expected a finding of increased rather than decreased sympathetic function. It is also unlikely that other unknown factors influenced the results because the tests used in the current study are well-accepted, robust, and reproducible measures of autonomic function. Most of the factors influencing the results have been worked-out extremely well (19), and care was taken in this study to avoid these confounding variables.
On theoretical grounds, it is easy to understand how reduced sympathetic function, if also reduced in the gastrointestinal tract, could be associated with abdominal pain. Unrestrained parasympathetic activity would produce abnormally strong peristaltic contractions against an unrelaxed postperistaltic segment. Based on animal studies, sympathetic function has been termed the “brake” of the gastrointestinal tract. After celiac and mesenteric gangliectomy in dogs, the duration of the interdigestive motor complexes is increased (20). With administration of sympatholytic agents, vagus nerve stimulation evokes strong contractions throughout the small intestine, whereas in the absence of those drugs, the vagal response is minimal. This response is more intense in the ileum than in the jejunum (21). Unfortunately, our study did not include motility assessment, and this speculative argument awaits further work, in particular, measurement of antroduodenal motility and colonic transit in these patients. Also, because the autonomic reflexes we assessed are primarily noradrenergic and cholinergic, our study does not address the possible involvement of other transmitters in the production of the symptoms of FAP, such as dopamine, which is known to delay gastric emptying and prolong orocecal transit time (22).
In contrast to our study, studies in adults with IBS have generally shown parasympathetic abnormalities (9,10,23). Some adult studies evaluated parasympathetic function only, using heart rate variation with deep breathing or spectral analysis (24–26). Obviously, only studies that evaluated both functions (9,10) can address the question of which branch is affected. Further study is necessary to determine whether parasympathetic function is truly unaffected in children with FAP. Such a finding would suggest that the autonomic abnormalities in children with FAP differ from those of adults with IBS. Another difference between our findings and those in adults, at least in one study (9), was the lack of association between specific gastrointestinal symptom type (e.g., upper vs. lower gastrointestinal symptoms) and type of autonomic symptoms or autonomic laboratory findings. This is not surprising because the label “functional abdominal pain” defines a syndrome, which probably has many pathophysiologic mechanisms. Current autonomic nervous system evaluation focuses primarily on efferent systems. Yet pain is mainly an afferent function, and visceral hyperalgesia plays an important role in adults in the generation of pain in patients with IBS (27,28). A more complete picture will emerge with increased understanding of the interaction between the efferent autonomic nervous system and afferent sensory information from the bowel.
Interestingly, the abnormalities in our patients, orthostasis or orthostatic tachycardia in most cases, match those seen in “postural orthostatic tachycardia syndrome” (POTS). Gastrointestinal symptoms occur commonly in this syndrome (14). Also, as in the cases reported by Schondorf and Low (29), most in our study were females. This syndrome probably also has multiple causes, including an autoimmune etiology, and these authors considered it an attenuated form of pandysautonomia. Treatment for postural orthostatic tachycardia syndrome often involves use of a β-adrenergic receptor blocker, such as propranolol, or low-dose fludrocortisone (14). In this regard, it is intriguing that some of our patients have responded to each of these medications (11,30) with prolonged resolution of gastrointestinal symptoms. Fludrocortisone in low doses is known to up-regulate adrenergic receptors (31). Increasing the number of receptors will counterbalance a reduction in sympathetic outflow. Response to this agent therefore supports the thesis of this study, namely, that reduced sympathetic outflow to the gastrointestinal tract may play a major role in the pathogenesis of FAP.
The tests of sweating nearly uniformly also showed a reduction in sympathetic output. The pattern of sweat loss, however, differed among patients, implicating different neuraxis locations. Nearly every possible neuraxis location occurred in at least one patient. This diversity is in keeping with the current understanding of functional abdominal pain as a syndrome with many potential etiologies. It should be kept in mind that sweat-test interpretation is based on the very limited normal sweat test controls available for this age group. Nonetheless, the findings in the sudomotor tests support the more robust cardiovascular abnormalities, namely, that sympathetic outflow is reduced.
It is also possible that the sudomotor findings have a more specific meaning in FAP, reflecting specific involvement of cholinergic sympathetic nerve fibers. Sweat glands have a unique sympathetic cholinergic innervation. In addition, sympathetic cholinergic postganglionic nerves are present in the intestine, although their function is unknown (32). In this respect, it is intriguing that most patients with significant anhidrosis had loss of sweating over the abdomen. Although this finding could clearly be nonspecific, one must also consider the interesting possibility that an abnormality of cholinergic sympathetic nerve fibers underlies the findings and the symptoms.
Functional abdominal pain has long been ascribed a psychoemotional basis. Our findings do not directly influence this understanding. Rather, they provide insight into the possible chain of events leading to the symptoms. The autonomic abnormalities could be the primary cause, with psychologic abnormalities as a result of chronic symptomology, as is probably the case in chronic pain syndromes (33). Conversely, the reverse could easily be postulated, with psychologic factors primary, and autonomic abnormalities associated with the same underlying pathologic process or as a result of the psychologic disturbance itself. These findings only constitute a preliminary set of observations. Future work will address etiologic questions only through well-controlled prospective trials with careful patient selection and comprehensive psychologic evaluation.
1. Apley J, Naish N. Recurrent abdominal pain: A field survey of 1,000 school children. Arch Dis Child 1958; 33: 169–70.
2. Oster J. Recurrent abdominal pain, headache, and limb pains in children and adolescents. Pediatrics 1972; 50: 429–36.
3. Stickler G, Murphy D. Recurrent abdominal pain. Am J Dis Child 1979; 133: 486–9.
4. Liebman W. Recurrent abdominal pain in children: a retrospective survey of 119 patients. Clin Pediatr 1978; 17: 149–53.
5. Stone R, Barbero G. Recurrent abdominal pain in childhood. Pediatrics 1970; 45: 732–8.
6. Jian R, Ducrot A, Ruskone A, et al. Symptomatic, radionuclide and therapeutic assessment of chronic idiopathic dyspepsia: A double-blind placebo-controlled evaluation of cisapride. Dig Dis Sci 1989; 34: 657–64.
7. Stanghellini V, Ghidini C, Maccarini M, et al. Fasting and postprandial gastrointestinal motility in ulcer and non-ulcer dyspepsia. Gut 1992; 33: 184–90.
8. Snape W, Carlson G, Cohen S. Colonic myoelectric activity in the irritable bowel syndrome. Gastroenterology 1976; 70: 326–30.
9. Aggarwal A, Cutts T, Abell T, et al. Predominant symptoms in irritable bowel syndrome correlate with specific nervous system abnormalities. Gastroenterology 1994; 106: 945–50.
10. Camilleri M, Fealey R. Idiopathic autonomic denervation in eight patients presenting with functional gastrointestinal disease. Dig Dis Sci 1990; 35: 609–16.
11. Chelimsky G, Hupertz VF, Chelimsky TC. Abdominal pain as the presenting symptom of autonomic dysfunction in a child. Clin Pediatr 1999; 38: 725–9.
12. Ingall TJ, McLeod JG, O'Brien PC. The effects of ageing on the autonomic nervous system function. Aust N Z J Med 1990; 20: 570–7.
13. Low P. Laboratory evaluation of autonomic failure. In: Low P, ed. Clinical Autonomic Disorders. Boston: Little Brown & Company; 1993: 169–96.
14. Schondorf R, Low P. Idiopathic postural tachycardia syndrome. In: Low P, ed. Clinical Autonomic Disorders. Boston: Little, Brown & Company; 1993: 641–52.
15. Low PA, Caskey PE, Tuck RR, et al. Quantitative sudomotor axon reflex in normal and neuropathic subjects. Ann Neurol 1983; 14: 573–80.
16. Fealey R, Low P, Thomas J. Thermoregulatory sweating abnormalities in diabetes mellitus. Mayo Clin Proc 1989; 64: 617–28.
17. Cohen J, Low P, Fealey R, et al. Somatic and autonomic function in progressive autonomic failure and multiple system atrophy. Ann Neurol 1987; 22: 692–9.
18. Kristensson K, Nordborg C, Olsson Y, et al. Changes in the vagus nerve in diabetes mellitus. Acta Pathol Microbiol Scand 1971; 79: 684–5.
19. Low P. Laboratory evaluation of autonomic failure. In: Low P, ed. Clinical Autonomic Disorders. Boston: Little Brown & Company; 1997: 169–96.
20. Marlett J, Code C. Effects of celiac and superior mesenteric ganglionectomy on interdigestive myoelectric complex in dogs. Am J Physiol 1979; 237: E432–43.
21. Kewenter J. The vagal control of the duodenal and ileal motility and blood flow. Acta Physiol Scand (Suppl)
22. Levein NG, Thorn SE, Wattwil M. Dopamine delays gastric emptying and prolongs orocaecal transit time in volunteers. Eur J Anaesthesiol 1999; 16: 246–50.
23. Lee GW, Weeks PM. The role of bone scintigraphy in diagnosing reflex sympathetic dystrophy [see comments; review] J Hand Surg [Am] 1995; 20: 458–63.
24. Orr W. Regulation of cardiac function during sleep in patients with irritable bowel syndrome [abstract]. Gastroenterology 1999; 116: A1054.
25. Harnish M, Elsenbruch S, Orr W. Autonomic responses to visceral stimuli in the irritable bowel syndrome [abstract]. Gastroenterology 1999; 116: A1002.
26. Heitkemper M, Burr R, Jarrett M, et al. Evidence of autonomic nervous system imbalance in women with irritable bowel syndrome. Dig Dis Sci 1998; 43: 2093–8.
27. Chun A, Desautels S, Slivka A, et al. Visceral algesia in irritable bowel syndrome, fibromyalgia, and sphincter of Oddi dysfunction, Type III. Dig Dis Sci 1999; 44: 631–6.
28. Munakata J, Naliboff B, Harraf F, et al. Repetitive sigmoid stimulation induces rectal hyperalgesia in patients with irritable bowel syndrome. Gastroenterology 1997; 112: 55–63.
29. Schondorf R, Low P. Idiopathic postural orthostatic tachycardia syndrome: an attenuated form of acute pandysautonomia? Neurology 1993; 43: 132–6.
30. Chelimsky G, Chelimsky TC. Treatment of autonomic dysfunction resolving gastrointestinal symptoms in a parent and child. J Auton Nerv Syst 1999; 9: 238.
31. Davies B, Bannester R, Sever P, et al. The pressor actions of noradrenaline, angiotensine II and saralasin in chronic autonomic failure treated with fludrocortisone. Br J Clin Pharmacol 1979; 8: 253–60.
32. Roman C, Gonella J. Extrinsic control of digestive tract motility. In: Johnson L, ed. Physiology of the Gastrointestinal Tract.
Vol. 1. New York: Raven Press; 1987: 507–53.
33. Pain and behavioral medicine: a cognitive-behavioral perspective. In: Turk D, Meichenbaum D, Genest M, eds. New York: The Guilford Press; 1983.