- Email: [email protected]
Atypical gastrointestinal symptoms are not associated with gallstones in patients with spinal cord injury
1085
Atypical Gastrointestinal Symptoms Are Not Associated With Gallstones in Patients With Spinal Cord Injury Ravi Moonka, MD, Steven A. Stiens, MD, Matthias Stelzner, MD ABSTRACT. Moonka R, Stiens SA, Stelzner M. Atypical gastrointestinal symptoms are not associated with gallstones in patients with spinal cord injury. Arch Phys Med Rehabil 2000;81:1085-9. Objective: To determine if nonspecific gastrointestinal (GI) symptoms justify cholecystectomy in patients with spinal cord injury (SCI). Design: The frequency of GI symptoms was determined in a sample of patients with SCI in whom the presence or absence of gallstones had been previously determined by screening ultrasonography or a known history of cholecystectomy. The prevalence of various symptoms in patients with and without gallstones was compared. Setting: The Spinal Cord Injury Unit of the Veterans Affairs Puget Sound Health Care System, which provides rehabilitation and longitudinal primary care for SCI veterans. Patients: Two hundred ninety-four patients who had undergone either right upper quadrant ultrasonography or cholecystectomy in the past, and who completed a questionnaire concerning GI symptoms. Main Outcome Measure: Bivariate logistic regression was used to calculate odds ratios (ORs) to determine the strength of associations between the presence of each symptom and the presence of gallstones. Results: Pain in the right upper quadrant or epigastrium that occurred after meals or at night was significantly associated with gallstones (OR: 3.5; 95% confidence interval [CI] 1.0211.73). Abdominal pain in other locations and nonspecific symptoms such as bloating and nausea, were not predictive of the presence of gallstones. Conclusions: Nonspecific symptoms in patients with SCI are not associated with gallstones and do not justify cholecystectomy in patients with otherwise asymptomatic gallstones. Key Words: Gallstones; Spinal cord injuries; Biliary colic; Cholecystectomy; Rehabilitation. r 2000 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation
A
PPROXIMATELY ONE IN 1000 people in the United States lives with the paralysis and sensory deficits of a spinal cord injury (SCI), and approximately 10,000 new injuries occur every year.1 Presently, the long-term survival of these patients closely approximates that of the able-bodied population.2 The objective of ongoing medical and rehabilitative management is to minimize the deterioration in the quality of
From the Department of Surgery (Moonka, Stelzner) and Department of Rehabilitation Medicine (Stiens), The Seattle Division of the Veterans Affairs Puget Sound Health Care System, The University of Washington School of Medicine, Seattle, WA. Submitted July 7, 1999. Accepted in revised form December 13, 1999. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors or upon any organization with which the authors are associated. Reprint requests to Ravi Moonka, MD, C6-GSUR, 1100 Ninth Avenue, PO Box 900, Seattle, WA 98111. 0003-9993/00/8108-5723$3.00/0 doi: 10.1053/apmr.2000.6288
life associated with these injuries. A common challenge to this management strategy is neurogenic gastrointestinal (GI) dysfunction, which causes symptoms of abdominal pain, bloating, and nausea,3,4 and is often refractory to therapeutic intervention. One potential etiology of these complaints is gallstones. Our group and others have shown that almost one-third of SCI patients develop biliary calculi subsequent to their injury.5,6 The risk of gallstone formation in this population is 3 to 6 times higher than is seen in age- and sex-matched controls. Neurologically intact patients who develop symptoms related to gallstones classically complain of biliary colic, in which severe pain occurs in the right upper quadrant (RUQ) or epigastrium, generally after meals or at night.7-14 Because the ability of patients with SCI to localize visceral pain is potentially impaired, the pattern of symptoms they experience may differ. Gallstones could instead cause more nonspecific findings such as generalized abdominal pain, nausea, or bloating. If such an association could be demonstrated, cholecystectomy would therefore be indicated for dyspeptic symptoms in SCI patients with gallstones, even though these symptoms would generally not lead to removal of the gallbladder in the population without SCI. Increasingly over the past decade, SCI patients enrolled in the Spinal Cord Injury Unit of the Veterans Affairs Puget Sound Health Care System have undergone screening RUQ ultrasonography to detect gallstones. Calculi are frequently seen in patients who do not complain of classic biliary colic, but who do note frequent and uncomfortable episodes of dyspepsia. To determine the role of cholecystectomy in such patients, the prevalence of symptoms such as atypical pain, bloating, and nausea was compared between SCI patients with and without gallstones. MATERIALS AND METHODS Subjects A total of 509 patients were seen at the Seattle VAMC Spinal Cord Injury Unit as either inpatients or outpatients from January 1, 1995, to December 31, 1997, all of whom were considered for enrollment in this study. Annual RUQ ultrasonography was performed during this interval on all patients as part of a standard yearly evaluation. As a result, 449 of these 509 patients (88%) had either undergone an informative right RUQ ultrasound or were known to have a surgically absent gallbladder. The remaining 60 patients had only transient contact with the SCI unit, and so were not enrolled in the standard screening evaluation. Because the purpose of the study was to compare symptoms between SCI patients with or without gallstones, patients in whom the presence or absence of gallstones had not been established were not included in the study. In the early part of 1998, these 449 patients were surveyed to determine the frequency with which they experienced abdominal pain, bloating, and nausea, and the general location and pattern of their pain (see Appendix). Patients were initially surveyed by a questionnaire mailed to the patients’ last known Arch Phys Med Rehabil Vol 81, August 2000
1086
GALLSTONE SYMPTOMS IN SCI PATIENTS, Moonka
address. Nonresponders were then contacted by phone and the survey was administered by reading the same questionnaire. Thirty-five patients had died between the time they had been last seen at the SCI unit and the time during which the questionnaires were made available. Two hundred ninety-four of the remaining 414 patients (71%) completed a questionnaire; 196 patients responded by mail, 97 were contacted by phone, and 1 patient was interviewed in person. The study population comprises these 294 patients. Of the 120 patients who were not surveyed, approximately 10% chose not to participate, and the remainder could not be reached at their listed address or telephone number. Survey and Data Collection Through the questionnaire, patients were asked to quantitate the frequency of their abdominal pain, and to identify its location. Pain location was classified into five categories: (1) RUQ and epigastric; (2) left lower quadrant or bilateral lower quadrant; (3) right lower quadrant; (4) left upper quadrant; and (5) diffuse, variable, difficult to localize, or periumbilical. Patients were considered to have abdominal pain if they experienced discomfort at least once a month. RUQ and epigastric pain were grouped together, because both areas are typical sites of biliary pain. Left lower quadrant and bilateral lower quadrant pain were also combined, because these sites are generally not associated with biliary pain.7 The survey and all other elements of the study were approved by the Human Subjects Review Committee of the University of Washington and the Research and Development Committee of the Veterans Affairs Puget Sound Health Care System. Biliary colic is classically described as either nocturnal or postprandial, so patients were also asked if their pain often occurred at night or soon after meals. For purposes of convenience, patients with a predominantly postprandial or nocturnal component to their pain, regardless of its location, were described as having a ‘‘biliary’’ pain pattern. Finally, patients were questioned about the frequency of abdominal bloating or nausea, and were considered to have these symptoms if they reported episodes more often than once a month. Demographic and injury-related information, including age, gender, body mass index, and duration, level, and severity (complete vs incomplete) of injury, was obtained from the medical record and from additional questions on the survey. RUQ ultrasound results of surveyed patients were compiled from computer-based radiologic archives. Statistical Analysis Binary logistic regression was used to assess the association between various symptoms and the presence or absence of gallstones. Symptoms that appeared to be associated with gallstones by this analysis were then included in a full multivariable model that included statistically significant demographic risk factors for gallstones as determined by t test and 2 analysis. All statistical analysis was performed using SPSS Version 9.0 for Windows.a RESULTS Demographic and Injury-Related Characteristics Compared Between Patients With and Without Gallstones Sixty of the 294 patients had gallstones seen on RUQ ultrasonography. In the remaining 234 patients, the ultrasound study demonstrated the absence of gallstones, or the medical record clearly documented a prior cholecystectomy. Patients with gallstones were older and were more likely to be women, Arch Phys Med Rehabil Vol 81, August 2000
Table 1: Comparison of Demographic Factors Between Patients With and Without Gallstones Factor
Age (yrs ⫾SD) Body mass index (⫾SD) Injury duration (yrs ⫾ SD) Female (n, %) Injury level above T6 (n, %) Complete deficit (vs incomplete)
Stones (n ⫽ 60)
No Stones (n ⫽ 234)
t Statistic or 2 Value
p
55.9 ⫾ 11.3
51.6 ⫾ 13.4
⫺2.46
.02
26.4 ⫾ 4.8
25.7 ⫾ 5.9
⫺.99
.32
18.0 ⫾ 11.7 5 (8.3%)
18.1 ⫾ 13.0 4 (1.7%)
.06 7.32
.95 .01
35 (58.3%)
145 (62.0%)
0.25
.55
39 (65%)
133 (57.8%)
1.31
.23
compared with patients without gallstones (table 1). High level of injury, duration of injury, body mass index, and injury severity were not associated with gallstones. Female gender and age are well-established risk factors for gallstones, and it is not surprising that patients with SCI would demonstrate these same associations. Symptom Complexes Associated With Gallstones Using binary logistic regression, various combinations of symptom locations and patterns were studied to determine their association with the presence of gallstones. Patients were categorized based on the location of pain, the ‘‘biliary’’ pattern of the pain, the presence or absence of bloating or nausea, and combinations of these attributes (table 2). The only pain pattern that was statistically associated with gallstones (95% confidence intervals do not include the value 1) was ‘‘biliary’’ RUQ pain. No other pain location was associated with gallstones, even when confined to patients whose pain occurred in a Table 2: Symptom Complexes as Risk Factors for Gallstones Frequency of Symptom Complex
Symptoms
Pain (any location) Biliary* pain (any location) Right upper quadrant pain Biliary* right upper quadrant pain Left upper quadrant pain Biliary* left upper quadrant pain Right lower quadrant pain Biliary* right lower quadrant pain Left lower quadrant pain or lower abdominal pain Biliary* left lower quadrant or lower abdominal pain Diffuse pain Biliary* diffuse pain Postprandial pain Nocturnal pain Bloating Nausea
Patients With Stones (n ⫽ 60)
Patients Without Bivariate 95% Stones Odds Confidence (n ⫽ 234) Ratio Intervals
51.7% 30.5% 10.0%
54.3% 24.2% 6.4%
0.9 1.4 1.6
.51-1.59 .73-2.58 .60-4.38
8.3% 10.0%
2.6% 6.0%
3.5 1.8
1.02-11.73 .64-4.75
6.7% 10.0%
2.6% 6.4%
2.7 1.6
0.7-9.9 0.6-4.4
5.1%
2.1%
2.5
0.6-10.6
8.3%
13.2%
0.6
0.2-1.6
3.3% 13.3% 6.7% 20.3% 13.6% 45.6% 16.9%
5.6% 22.2% 11.6% 12.6% 19.5% 46.9% 11.2%
0.6 0.5 0.5 1.8 0.7 1.6 1.0
0.1-2.7 0.2-1.2 0.2-1.6 0.9-3.7 0.3-1.5 0.7-3.6 0.5-1.7
* The adjective ‘‘biliary’’ denotes inclusion only of patients whose pain occurred after meals or at night.
GALLSTONE SYMPTOMS IN SCI PATIENTS, Moonka
‘‘biliary’’ fashion. ‘‘Biliary’’ characteristics, in which pain is postprandial or nocturnal, were not in and of themselves predictive of stones, nor were isolated nausea or bloating. Of the 294 SCI patients in the study, 11 had classic biliary colic, describing RUQ or epigastric pain that occurred following meals or at night. Five of these 11 patients had gallstones. The association between this classic biliary pattern of pain and the presence of gallstones might be expected to be perceived only by patients with SCIs below T10, in whom innervation of the gallbladder and the upper abdominal wall is essentially intact. However, this subgroup included 1 patient with a C7 level, 2 patients with T5 levels, and 2 patients with L2 levels, suggesting that this symptom complex is seen equally often in patients with ‘‘high’’ and ‘‘low’’ SCIs. Similarly, symptoms of biliary colic might be expect to be relatively less common in patients with complete injuries compared with those with incomplete injuries. However, 3 of these 5 patients with gallstones and biliary colic had a complete loss of sensory function below their level of injury. Multivariable Analysis To determine if RUQ pain in a biliary pattern is an independent risk factor for gallstones in patients with SCI, multivariable logistic regression was performed on a full model, including age and gender as potential risk factors in combination with this particular pain pattern. Age and gender were chosen because they were statistically significant demographic risk factors by bivariable analysis. In this analysis, ‘‘biliary’’ RUQ or epigastric pain remained a risk factor for gallstones, though this association was not as statistically significance as it had been in the bivariable analysis (table 3). DISCUSSION Our data suggest an association of gallstones with ‘‘biliary’’ RUQ pain in patients with SCI. This association was not confined to patients with low or incomplete SCIs. Prior reports suggest spinal cord patients with cholecystitis are often able to localize their pain to the RUQ, and this ability to localize pain is, to some extent, independent of injury level.15,16 Typically, sensory innervation from the gallbladder is thought to be carried by sympathetic afferent fibers entering the spinal cord at the T6 to T10 level.17 As has been previously suggested, in patients with lesions above these levels, the ability to experience biliary colic suggests that alternate pain pathways are in place or are recruited.18,19 The association between a biliary pattern of pain and gallstones may be underestimated as a result of the study design. The health care providers who staff the Veterans Affairs Puget Sound Health Care System Spinal Cord Injury Unit are aware of the propensity of patients with SCIs to develop gallstones and the usual symptoms associated with biliary disease. Consequently, these symptoms are elicited as part of the annual evaluation that all patients enrolled in the SCI unit regularly complete. Most patients who have a combination of gallstones and RUQ pain will be referred for surgery. Patients Table 3: Independent Risk Factors for Gallstones Factor
Odds Ratio
95% Confidence Intervals
p
Age* Male gender Biliary pain†
1.03 0.2 2.7
1.01-1.05 .04-0.8 0.8-9.9
.02 .02 .12
* Odds ratio per year of age difference. Right upper quadrant pain or epigastric pain that occurs after meals or at night.
†
1087
who have had their gallbladder removed are segregated into the ‘‘no stones’’ cohort, diminishing the number of patients with gallstones who suffer from this particular symptom complex, and weakening the association observed between biliary pain and gallstones. Similarly, placement of cholecystectomy patients in the ‘‘no stones’’ group also renders the study ill-suited to identifying demographic risk factors for gallstones, a goal that was been better taken up by other studies.6 The association of pain in a biliary pattern and gallstones in patients with SCI is not completely unexpected. Of greater practical concern to clinicians is the degree of association between gallstones and nonspecific symptoms such as generalized pain and dyspepsia in this population. A variety of survey studies have documented the wide prevalence of abdominal complaints among SCI patients. While many of these problems are the result of difficulties with elimination, poorly localized pain and dyspepsia are also prominent problems interfering with a high quality of life.3,20 If certain patterns of symptoms could be ascribed to gallstones, cholecystectomy could be offered to subpopulations of patients in hopes of providing symptomatic relief. The presence of such an association between nonspecific ‘‘dyspeptic’’ symptoms with gallstones in neurologically intact patients has been studied for the past 50 years. Unfortunately, the great variety of methodologies used to investigate this question does not easily lend itself to definitive statements. Even generally well accepted characteristics of biliary pain, such as its tendency to occur after meals and to be exacerbated by fatty meals, are not always clearly associated with gallstones or a satisfactory response to cholecystectomy.7,21,22 It appears that dyspeptic symptoms such as bloating and nausea will often respond to cholecystectomy, if they occur in conjunction with ‘‘biliary’’ pain in the RUQ and epigastrium.10,12,23 However, these symptoms do not dissipate as reliably as RUQ pain, and their presence often predicts a general dissatisfaction with the operative results on the part of the patient.12,13,24 Dyspepsia in isolation or diffuse or atypical pain generally is not accepted as an indication for surgery, and so the response of these symptoms to cholecystectomy has been only sporadically studied, and in small numbers of patients. Although some studies have suggested a tendency of these symptoms to resolve after surgery,25,26 their response rate is difficult to discriminate from a placebo effect of surgery in other reports.8,13 What has been definitively shown is that patients with and without gallstones demonstrate nonspecific symptoms with an equal frequency, suggesting that gallstones do not directly cause symptoms of dyspepsia.26-28 Similar to the findings observed in patients without SCI, we found no increase in the prevalence of dyspeptic symptoms in SCI patients with gallstones compared with those without gallstones, nor was pain in general or diffuse pain associated with the presence of gallstones. The lack of association between gallstones and atypical symptoms suggests that gallstones do not generally cause these symptoms in SCI patients. The study design does create several potential biases. Because the study was not blinded, patients who were aware that they had gallstones might have been more likely to complain of pain compared with their acalculus cohort. However, the absolute percentage of patients describing pain was the same in patients with and without stones, and patients were not explicitly told on the survey whether or not they had gallstones. Similarly, the tendency of the physicians who administered the survey to elicit symptoms of biliary colic from patients known to have gallstones was minimized through the use of a standard survey instrument. Finally, because only 71% of patients we Arch Phys Med Rehabil Vol 81, August 2000
1088
GALLSTONE SYMPTOMS IN SCI PATIENTS, Moonka
had hoped to question were actually surveyed, it raises the possibility that the conclusions of the study may have differed if all patients had been contacted. However, it is unlikely that the exclusion of patients based on our inability to contact them would result in any systematic bias that could significantly change the results. Despite the absence of a general association, it remains possible that in some fraction of patients with SCI, these atypical symptoms are being caused by gallstones and will respond to cholecystectomy, much as they sometimes do in patients without SCI. Further defining various aspects of reported symptoms not included in our survey, such as the frequency and duration of attacks and the oscillatory nature of any particular episode, might further stratify patients into groups more or less likely to benefit from a cholecystectomy.7 Similarly, a survey is an inadequate substitute for the live interview and subsequent physical examination, which usually underlies any impression that a given patient will or will not benefit from surgery. Nonetheless, in patients without SCI, the likelihood that a patient will find relief of subjective symptoms after a cholecystectomy declines proportionately to the degree that these symptoms deviate from what is generally considered to be classic biliary colic. It appears unlikely, based on our results, that a different pattern will be observed in patients with SCI. References 1. Blumer CE, Quine S. Prevalence of spinal cord injury: an international comparison. Neuroepidemiology 1995;14:258-68. 2. Samsa GP, Patrick CH, Feussner JR. Long-term survival of veterans with traumatic spinal cord injury. Arch Neurol 1993;50: 909-14. 3. Stone JM, Nino-Murcia M, Wolfe VA, Perkash I. Chronic gastrointestinal problems in spinal cord injury patients: a prospective analysis. Am J Gastroenterol 1990;85:1114-9. 4. Stiens SA, Bergman SB, Goetz LL. Neurogenic bowel dysfunction after spinal cord injury: clinical evaluation and rehabilitative management. Arch Phys Med Rehabil 1997;78 Suppl:S86-S102. 5. Apstein MD, Dalecki-Chipperfield K. Spinal cord injury is a risk factor for gallstone disease. Gastroenterology 1987;92:966-8. 6. Moonka R, Steins SA, Resnick WJ, McDonald JM, Eubank WB, Dominitz JA, et al. The incidence and natural history of gallstones in the spinal cord injured population. J Am Coll Surg 1999;189:27481. 7. Diehl AK, Sugarek NJ, Todd KH. Clinical evaluation for gallstone disease: usefulness of symptoms and signs in diagnosis. Am J Med 1990;89:29-33. 8. Fenster LF, Lonborg R, Thirlby RC, Traverso LW. What symptoms does cholecystectomy cure? Insights from an outcomes measurement project and review of the literature. Am J Surg 1995;169: 533-8. 9. Gilliland TM, Traverso LW. Modern standards for comparison of cholecystectomy with alternatives treatments for symptomatic cholelithiasis with emphasis on long term relief of symptoms. Surg Gynecol Obstet 1990;170:39-44. 10. Gunn A, Keddie N. Some clinical observations on patients with gallstones. Lancet 1972;2:239-41. 11. Gui GPH, Cheruvu CVN, West N, Sivaniah K, Fiennes AGTW. Is cholecystectomy effective treatment for symptomatic gallstones? Clinical outcome after long-term follow up. Ann R Coll Surg Engl 1998;80:25-32. 12. Konsten J, Gouma DJ, von Meyenfeldt MF, Menheere P. Longterm follow-up after open cholecystectomy. Br J Surg 1993;80: 100-2. 13. Ros E, Zambon D. Postcholecystectomy symptoms. A prospective study of gallstone patients before and two years after surgery. Gut 1987;28:1500-4. 14. Traverso LW. Clinical manifestations and impact of gallstone disease. Am J Surg 1993;165:405-7. Arch Phys Med Rehabil Vol 81, August 2000
15. Charney KJ, Juler GL, Comarr E. General surgery problems in patients with spinal cord injuries. Arch Surg 1975;110:1083-8. 16. Moonka R, Steins SA, Eubank WB, Stelzner M. The presentation of gallstones and results of biliary surgery in a spinal cord injured population. Am J Surg 1999;178:246-50. 17. Banfield WJ. Physiology of the gallbladder. Gastroenterology 1975;69:770-7. 18. Ingberg HO, Prust FW. The diagnosis of abdominal emergencies in patients with spinal cord injuries. Arch Phys Med Rehabil 1968;49: 343-8. 19. Neumayer LA, Bull DA, Mohr JD, Putnam CW. The acutely affected abdomen in paraplegic spinal cord injury patients. Ann Surg 1990;212:561-6. 20. Han TR, Kim JH, Kwon BS. Chronic gastrointestinal problems and bowel dysfunction in patients with spinal cord injury. Spinal Cord 1998;36:485-90. 21. Koch JP, Donaldson RM. A survey of food intolerances in hospitalized patients. N Engl J Med 1964;271:657-60. 22. Talley NJ, McNeil D, Piper DW. Discriminant value of dyspeptic symptoms: a study of the clinical presentation of 221 patients with dyspepsia of unknown cause, peptic ulceration, and cholelithiasis. Gut 1987;28:40-6. 23. Kingston RD, Windsor CWO. Flatulent dyspepsia in patients with gallstones undergoing cholecystectomy. Br J Surg 1975;62:231-3. 24. Bates T, Ebbs SR, Harrison M, A’Hern RP. Influence of cholecystectomy on symptoms. Br J Surg 1991;78:964-7. 25. Rhind JA, Watson L. Gallstone dyspepsia. Br Med J 1968;1:32. 26. Hinkel CL, Moller GA. Correlation of symptoms, age, sex, and habitus with cholecystographic findings in 1,000 consecutive exams. Gastroenterology 1957;32:807-15. ¨ stberg H. 27. Janzon L, Aspelin S, Eriksson J, Hildell J, Trell E, O Ultrasonographic screening for gallstone disease in middle-aged women. Scand J Gastroenterol 1985;20:706-10. 28. Price WH. Gallbladder dyspepsia. Br Med J 1963;2:138-41. Supplier a. SPSS, 233 S Wacker Dr, Chicago, IL 60606.
APPENDIX Gallstone Survey Name: SSN: Mark the choice that best applies to you: 1. How often do you have abdominal pain? Never Once or twice a year Once or twice a month Once or twice a week Every day 2. Where in your abdomen is your pain? All over In the upper half, close to my rib cage, on the right In the upper half, close to my rib cage, on the left In the lower half, below my belly button, on the right In the lower half, below my belly button, on the left Location varies with each episode Hard to say 3. Do you find that your pain usually occurs after meals? Yes No 4. Does this pain ever wake you up at night? Yes No 5. How often do you have nausea or vomiting that you can’t explain? Never Once or twice a year Once or twice month
GALLSTONE SYMPTOMS IN SCI PATIENTS, Moonka
Once or twice a week Every day 6. How often do you have a feeling of your abdomen being overly full or bloated? Never Once or twice a year Once or twice a month Once or twice a week Every day 7. Have you ever had an x-ray suggesting you have gallstones? Yes No 8. If this x-ray was done at a facility besides the Seattle VA Medical Center, what was the name of the hospital, what city is it in, and as best as you can remember, when was it done?
1089
9. Have you had your gallbladder removed? Yes No If yes, any details of the operation you can recall would be helpful. (Which hospital, in which city, the approximate date of the surgery, whether it was emergency surgery, and how well you recovered.) 10. Have your parents, brothers and sisters, or children been diagnosed with gallstones? Yes No 11. How tall are you? 12. How much do you presently weigh? 13. Do you have diabetes? 14. If you have abdominal pain at least once a month, please mark on this diagram where it is. (Accompanied by anterior and posterior views of a homunculus).
Arch Phys Med Rehabil Vol 81, August 2000
Atypical Gastrointestinal Symptoms Are Not Associated With Gallstones in Patients With Spinal Cord Injury Ravi Moonka, MD, Steven A. Stiens, MD, Matthias Stelzner, MD ABSTRACT. Moonka R, Stiens SA, Stelzner M. Atypical gastrointestinal symptoms are not associated with gallstones in patients with spinal cord injury. Arch Phys Med Rehabil 2000;81:1085-9. Objective: To determine if nonspecific gastrointestinal (GI) symptoms justify cholecystectomy in patients with spinal cord injury (SCI). Design: The frequency of GI symptoms was determined in a sample of patients with SCI in whom the presence or absence of gallstones had been previously determined by screening ultrasonography or a known history of cholecystectomy. The prevalence of various symptoms in patients with and without gallstones was compared. Setting: The Spinal Cord Injury Unit of the Veterans Affairs Puget Sound Health Care System, which provides rehabilitation and longitudinal primary care for SCI veterans. Patients: Two hundred ninety-four patients who had undergone either right upper quadrant ultrasonography or cholecystectomy in the past, and who completed a questionnaire concerning GI symptoms. Main Outcome Measure: Bivariate logistic regression was used to calculate odds ratios (ORs) to determine the strength of associations between the presence of each symptom and the presence of gallstones. Results: Pain in the right upper quadrant or epigastrium that occurred after meals or at night was significantly associated with gallstones (OR: 3.5; 95% confidence interval [CI] 1.0211.73). Abdominal pain in other locations and nonspecific symptoms such as bloating and nausea, were not predictive of the presence of gallstones. Conclusions: Nonspecific symptoms in patients with SCI are not associated with gallstones and do not justify cholecystectomy in patients with otherwise asymptomatic gallstones. Key Words: Gallstones; Spinal cord injuries; Biliary colic; Cholecystectomy; Rehabilitation. r 2000 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation
A
PPROXIMATELY ONE IN 1000 people in the United States lives with the paralysis and sensory deficits of a spinal cord injury (SCI), and approximately 10,000 new injuries occur every year.1 Presently, the long-term survival of these patients closely approximates that of the able-bodied population.2 The objective of ongoing medical and rehabilitative management is to minimize the deterioration in the quality of
From the Department of Surgery (Moonka, Stelzner) and Department of Rehabilitation Medicine (Stiens), The Seattle Division of the Veterans Affairs Puget Sound Health Care System, The University of Washington School of Medicine, Seattle, WA. Submitted July 7, 1999. Accepted in revised form December 13, 1999. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors or upon any organization with which the authors are associated. Reprint requests to Ravi Moonka, MD, C6-GSUR, 1100 Ninth Avenue, PO Box 900, Seattle, WA 98111. 0003-9993/00/8108-5723$3.00/0 doi: 10.1053/apmr.2000.6288
life associated with these injuries. A common challenge to this management strategy is neurogenic gastrointestinal (GI) dysfunction, which causes symptoms of abdominal pain, bloating, and nausea,3,4 and is often refractory to therapeutic intervention. One potential etiology of these complaints is gallstones. Our group and others have shown that almost one-third of SCI patients develop biliary calculi subsequent to their injury.5,6 The risk of gallstone formation in this population is 3 to 6 times higher than is seen in age- and sex-matched controls. Neurologically intact patients who develop symptoms related to gallstones classically complain of biliary colic, in which severe pain occurs in the right upper quadrant (RUQ) or epigastrium, generally after meals or at night.7-14 Because the ability of patients with SCI to localize visceral pain is potentially impaired, the pattern of symptoms they experience may differ. Gallstones could instead cause more nonspecific findings such as generalized abdominal pain, nausea, or bloating. If such an association could be demonstrated, cholecystectomy would therefore be indicated for dyspeptic symptoms in SCI patients with gallstones, even though these symptoms would generally not lead to removal of the gallbladder in the population without SCI. Increasingly over the past decade, SCI patients enrolled in the Spinal Cord Injury Unit of the Veterans Affairs Puget Sound Health Care System have undergone screening RUQ ultrasonography to detect gallstones. Calculi are frequently seen in patients who do not complain of classic biliary colic, but who do note frequent and uncomfortable episodes of dyspepsia. To determine the role of cholecystectomy in such patients, the prevalence of symptoms such as atypical pain, bloating, and nausea was compared between SCI patients with and without gallstones. MATERIALS AND METHODS Subjects A total of 509 patients were seen at the Seattle VAMC Spinal Cord Injury Unit as either inpatients or outpatients from January 1, 1995, to December 31, 1997, all of whom were considered for enrollment in this study. Annual RUQ ultrasonography was performed during this interval on all patients as part of a standard yearly evaluation. As a result, 449 of these 509 patients (88%) had either undergone an informative right RUQ ultrasound or were known to have a surgically absent gallbladder. The remaining 60 patients had only transient contact with the SCI unit, and so were not enrolled in the standard screening evaluation. Because the purpose of the study was to compare symptoms between SCI patients with or without gallstones, patients in whom the presence or absence of gallstones had not been established were not included in the study. In the early part of 1998, these 449 patients were surveyed to determine the frequency with which they experienced abdominal pain, bloating, and nausea, and the general location and pattern of their pain (see Appendix). Patients were initially surveyed by a questionnaire mailed to the patients’ last known Arch Phys Med Rehabil Vol 81, August 2000
1086
GALLSTONE SYMPTOMS IN SCI PATIENTS, Moonka
address. Nonresponders were then contacted by phone and the survey was administered by reading the same questionnaire. Thirty-five patients had died between the time they had been last seen at the SCI unit and the time during which the questionnaires were made available. Two hundred ninety-four of the remaining 414 patients (71%) completed a questionnaire; 196 patients responded by mail, 97 were contacted by phone, and 1 patient was interviewed in person. The study population comprises these 294 patients. Of the 120 patients who were not surveyed, approximately 10% chose not to participate, and the remainder could not be reached at their listed address or telephone number. Survey and Data Collection Through the questionnaire, patients were asked to quantitate the frequency of their abdominal pain, and to identify its location. Pain location was classified into five categories: (1) RUQ and epigastric; (2) left lower quadrant or bilateral lower quadrant; (3) right lower quadrant; (4) left upper quadrant; and (5) diffuse, variable, difficult to localize, or periumbilical. Patients were considered to have abdominal pain if they experienced discomfort at least once a month. RUQ and epigastric pain were grouped together, because both areas are typical sites of biliary pain. Left lower quadrant and bilateral lower quadrant pain were also combined, because these sites are generally not associated with biliary pain.7 The survey and all other elements of the study were approved by the Human Subjects Review Committee of the University of Washington and the Research and Development Committee of the Veterans Affairs Puget Sound Health Care System. Biliary colic is classically described as either nocturnal or postprandial, so patients were also asked if their pain often occurred at night or soon after meals. For purposes of convenience, patients with a predominantly postprandial or nocturnal component to their pain, regardless of its location, were described as having a ‘‘biliary’’ pain pattern. Finally, patients were questioned about the frequency of abdominal bloating or nausea, and were considered to have these symptoms if they reported episodes more often than once a month. Demographic and injury-related information, including age, gender, body mass index, and duration, level, and severity (complete vs incomplete) of injury, was obtained from the medical record and from additional questions on the survey. RUQ ultrasound results of surveyed patients were compiled from computer-based radiologic archives. Statistical Analysis Binary logistic regression was used to assess the association between various symptoms and the presence or absence of gallstones. Symptoms that appeared to be associated with gallstones by this analysis were then included in a full multivariable model that included statistically significant demographic risk factors for gallstones as determined by t test and 2 analysis. All statistical analysis was performed using SPSS Version 9.0 for Windows.a RESULTS Demographic and Injury-Related Characteristics Compared Between Patients With and Without Gallstones Sixty of the 294 patients had gallstones seen on RUQ ultrasonography. In the remaining 234 patients, the ultrasound study demonstrated the absence of gallstones, or the medical record clearly documented a prior cholecystectomy. Patients with gallstones were older and were more likely to be women, Arch Phys Med Rehabil Vol 81, August 2000
Table 1: Comparison of Demographic Factors Between Patients With and Without Gallstones Factor
Age (yrs ⫾SD) Body mass index (⫾SD) Injury duration (yrs ⫾ SD) Female (n, %) Injury level above T6 (n, %) Complete deficit (vs incomplete)
Stones (n ⫽ 60)
No Stones (n ⫽ 234)
t Statistic or 2 Value
p
55.9 ⫾ 11.3
51.6 ⫾ 13.4
⫺2.46
.02
26.4 ⫾ 4.8
25.7 ⫾ 5.9
⫺.99
.32
18.0 ⫾ 11.7 5 (8.3%)
18.1 ⫾ 13.0 4 (1.7%)
.06 7.32
.95 .01
35 (58.3%)
145 (62.0%)
0.25
.55
39 (65%)
133 (57.8%)
1.31
.23
compared with patients without gallstones (table 1). High level of injury, duration of injury, body mass index, and injury severity were not associated with gallstones. Female gender and age are well-established risk factors for gallstones, and it is not surprising that patients with SCI would demonstrate these same associations. Symptom Complexes Associated With Gallstones Using binary logistic regression, various combinations of symptom locations and patterns were studied to determine their association with the presence of gallstones. Patients were categorized based on the location of pain, the ‘‘biliary’’ pattern of the pain, the presence or absence of bloating or nausea, and combinations of these attributes (table 2). The only pain pattern that was statistically associated with gallstones (95% confidence intervals do not include the value 1) was ‘‘biliary’’ RUQ pain. No other pain location was associated with gallstones, even when confined to patients whose pain occurred in a Table 2: Symptom Complexes as Risk Factors for Gallstones Frequency of Symptom Complex
Symptoms
Pain (any location) Biliary* pain (any location) Right upper quadrant pain Biliary* right upper quadrant pain Left upper quadrant pain Biliary* left upper quadrant pain Right lower quadrant pain Biliary* right lower quadrant pain Left lower quadrant pain or lower abdominal pain Biliary* left lower quadrant or lower abdominal pain Diffuse pain Biliary* diffuse pain Postprandial pain Nocturnal pain Bloating Nausea
Patients With Stones (n ⫽ 60)
Patients Without Bivariate 95% Stones Odds Confidence (n ⫽ 234) Ratio Intervals
51.7% 30.5% 10.0%
54.3% 24.2% 6.4%
0.9 1.4 1.6
.51-1.59 .73-2.58 .60-4.38
8.3% 10.0%
2.6% 6.0%
3.5 1.8
1.02-11.73 .64-4.75
6.7% 10.0%
2.6% 6.4%
2.7 1.6
0.7-9.9 0.6-4.4
5.1%
2.1%
2.5
0.6-10.6
8.3%
13.2%
0.6
0.2-1.6
3.3% 13.3% 6.7% 20.3% 13.6% 45.6% 16.9%
5.6% 22.2% 11.6% 12.6% 19.5% 46.9% 11.2%
0.6 0.5 0.5 1.8 0.7 1.6 1.0
0.1-2.7 0.2-1.2 0.2-1.6 0.9-3.7 0.3-1.5 0.7-3.6 0.5-1.7
* The adjective ‘‘biliary’’ denotes inclusion only of patients whose pain occurred after meals or at night.
GALLSTONE SYMPTOMS IN SCI PATIENTS, Moonka
‘‘biliary’’ fashion. ‘‘Biliary’’ characteristics, in which pain is postprandial or nocturnal, were not in and of themselves predictive of stones, nor were isolated nausea or bloating. Of the 294 SCI patients in the study, 11 had classic biliary colic, describing RUQ or epigastric pain that occurred following meals or at night. Five of these 11 patients had gallstones. The association between this classic biliary pattern of pain and the presence of gallstones might be expected to be perceived only by patients with SCIs below T10, in whom innervation of the gallbladder and the upper abdominal wall is essentially intact. However, this subgroup included 1 patient with a C7 level, 2 patients with T5 levels, and 2 patients with L2 levels, suggesting that this symptom complex is seen equally often in patients with ‘‘high’’ and ‘‘low’’ SCIs. Similarly, symptoms of biliary colic might be expect to be relatively less common in patients with complete injuries compared with those with incomplete injuries. However, 3 of these 5 patients with gallstones and biliary colic had a complete loss of sensory function below their level of injury. Multivariable Analysis To determine if RUQ pain in a biliary pattern is an independent risk factor for gallstones in patients with SCI, multivariable logistic regression was performed on a full model, including age and gender as potential risk factors in combination with this particular pain pattern. Age and gender were chosen because they were statistically significant demographic risk factors by bivariable analysis. In this analysis, ‘‘biliary’’ RUQ or epigastric pain remained a risk factor for gallstones, though this association was not as statistically significance as it had been in the bivariable analysis (table 3). DISCUSSION Our data suggest an association of gallstones with ‘‘biliary’’ RUQ pain in patients with SCI. This association was not confined to patients with low or incomplete SCIs. Prior reports suggest spinal cord patients with cholecystitis are often able to localize their pain to the RUQ, and this ability to localize pain is, to some extent, independent of injury level.15,16 Typically, sensory innervation from the gallbladder is thought to be carried by sympathetic afferent fibers entering the spinal cord at the T6 to T10 level.17 As has been previously suggested, in patients with lesions above these levels, the ability to experience biliary colic suggests that alternate pain pathways are in place or are recruited.18,19 The association between a biliary pattern of pain and gallstones may be underestimated as a result of the study design. The health care providers who staff the Veterans Affairs Puget Sound Health Care System Spinal Cord Injury Unit are aware of the propensity of patients with SCIs to develop gallstones and the usual symptoms associated with biliary disease. Consequently, these symptoms are elicited as part of the annual evaluation that all patients enrolled in the SCI unit regularly complete. Most patients who have a combination of gallstones and RUQ pain will be referred for surgery. Patients Table 3: Independent Risk Factors for Gallstones Factor
Odds Ratio
95% Confidence Intervals
p
Age* Male gender Biliary pain†
1.03 0.2 2.7
1.01-1.05 .04-0.8 0.8-9.9
.02 .02 .12
* Odds ratio per year of age difference. Right upper quadrant pain or epigastric pain that occurs after meals or at night.
†
1087
who have had their gallbladder removed are segregated into the ‘‘no stones’’ cohort, diminishing the number of patients with gallstones who suffer from this particular symptom complex, and weakening the association observed between biliary pain and gallstones. Similarly, placement of cholecystectomy patients in the ‘‘no stones’’ group also renders the study ill-suited to identifying demographic risk factors for gallstones, a goal that was been better taken up by other studies.6 The association of pain in a biliary pattern and gallstones in patients with SCI is not completely unexpected. Of greater practical concern to clinicians is the degree of association between gallstones and nonspecific symptoms such as generalized pain and dyspepsia in this population. A variety of survey studies have documented the wide prevalence of abdominal complaints among SCI patients. While many of these problems are the result of difficulties with elimination, poorly localized pain and dyspepsia are also prominent problems interfering with a high quality of life.3,20 If certain patterns of symptoms could be ascribed to gallstones, cholecystectomy could be offered to subpopulations of patients in hopes of providing symptomatic relief. The presence of such an association between nonspecific ‘‘dyspeptic’’ symptoms with gallstones in neurologically intact patients has been studied for the past 50 years. Unfortunately, the great variety of methodologies used to investigate this question does not easily lend itself to definitive statements. Even generally well accepted characteristics of biliary pain, such as its tendency to occur after meals and to be exacerbated by fatty meals, are not always clearly associated with gallstones or a satisfactory response to cholecystectomy.7,21,22 It appears that dyspeptic symptoms such as bloating and nausea will often respond to cholecystectomy, if they occur in conjunction with ‘‘biliary’’ pain in the RUQ and epigastrium.10,12,23 However, these symptoms do not dissipate as reliably as RUQ pain, and their presence often predicts a general dissatisfaction with the operative results on the part of the patient.12,13,24 Dyspepsia in isolation or diffuse or atypical pain generally is not accepted as an indication for surgery, and so the response of these symptoms to cholecystectomy has been only sporadically studied, and in small numbers of patients. Although some studies have suggested a tendency of these symptoms to resolve after surgery,25,26 their response rate is difficult to discriminate from a placebo effect of surgery in other reports.8,13 What has been definitively shown is that patients with and without gallstones demonstrate nonspecific symptoms with an equal frequency, suggesting that gallstones do not directly cause symptoms of dyspepsia.26-28 Similar to the findings observed in patients without SCI, we found no increase in the prevalence of dyspeptic symptoms in SCI patients with gallstones compared with those without gallstones, nor was pain in general or diffuse pain associated with the presence of gallstones. The lack of association between gallstones and atypical symptoms suggests that gallstones do not generally cause these symptoms in SCI patients. The study design does create several potential biases. Because the study was not blinded, patients who were aware that they had gallstones might have been more likely to complain of pain compared with their acalculus cohort. However, the absolute percentage of patients describing pain was the same in patients with and without stones, and patients were not explicitly told on the survey whether or not they had gallstones. Similarly, the tendency of the physicians who administered the survey to elicit symptoms of biliary colic from patients known to have gallstones was minimized through the use of a standard survey instrument. Finally, because only 71% of patients we Arch Phys Med Rehabil Vol 81, August 2000
1088
GALLSTONE SYMPTOMS IN SCI PATIENTS, Moonka
had hoped to question were actually surveyed, it raises the possibility that the conclusions of the study may have differed if all patients had been contacted. However, it is unlikely that the exclusion of patients based on our inability to contact them would result in any systematic bias that could significantly change the results. Despite the absence of a general association, it remains possible that in some fraction of patients with SCI, these atypical symptoms are being caused by gallstones and will respond to cholecystectomy, much as they sometimes do in patients without SCI. Further defining various aspects of reported symptoms not included in our survey, such as the frequency and duration of attacks and the oscillatory nature of any particular episode, might further stratify patients into groups more or less likely to benefit from a cholecystectomy.7 Similarly, a survey is an inadequate substitute for the live interview and subsequent physical examination, which usually underlies any impression that a given patient will or will not benefit from surgery. Nonetheless, in patients without SCI, the likelihood that a patient will find relief of subjective symptoms after a cholecystectomy declines proportionately to the degree that these symptoms deviate from what is generally considered to be classic biliary colic. It appears unlikely, based on our results, that a different pattern will be observed in patients with SCI. References 1. Blumer CE, Quine S. Prevalence of spinal cord injury: an international comparison. Neuroepidemiology 1995;14:258-68. 2. Samsa GP, Patrick CH, Feussner JR. Long-term survival of veterans with traumatic spinal cord injury. Arch Neurol 1993;50: 909-14. 3. Stone JM, Nino-Murcia M, Wolfe VA, Perkash I. Chronic gastrointestinal problems in spinal cord injury patients: a prospective analysis. Am J Gastroenterol 1990;85:1114-9. 4. Stiens SA, Bergman SB, Goetz LL. Neurogenic bowel dysfunction after spinal cord injury: clinical evaluation and rehabilitative management. Arch Phys Med Rehabil 1997;78 Suppl:S86-S102. 5. Apstein MD, Dalecki-Chipperfield K. Spinal cord injury is a risk factor for gallstone disease. Gastroenterology 1987;92:966-8. 6. Moonka R, Steins SA, Resnick WJ, McDonald JM, Eubank WB, Dominitz JA, et al. The incidence and natural history of gallstones in the spinal cord injured population. J Am Coll Surg 1999;189:27481. 7. Diehl AK, Sugarek NJ, Todd KH. Clinical evaluation for gallstone disease: usefulness of symptoms and signs in diagnosis. Am J Med 1990;89:29-33. 8. Fenster LF, Lonborg R, Thirlby RC, Traverso LW. What symptoms does cholecystectomy cure? Insights from an outcomes measurement project and review of the literature. Am J Surg 1995;169: 533-8. 9. Gilliland TM, Traverso LW. Modern standards for comparison of cholecystectomy with alternatives treatments for symptomatic cholelithiasis with emphasis on long term relief of symptoms. Surg Gynecol Obstet 1990;170:39-44. 10. Gunn A, Keddie N. Some clinical observations on patients with gallstones. Lancet 1972;2:239-41. 11. Gui GPH, Cheruvu CVN, West N, Sivaniah K, Fiennes AGTW. Is cholecystectomy effective treatment for symptomatic gallstones? Clinical outcome after long-term follow up. Ann R Coll Surg Engl 1998;80:25-32. 12. Konsten J, Gouma DJ, von Meyenfeldt MF, Menheere P. Longterm follow-up after open cholecystectomy. Br J Surg 1993;80: 100-2. 13. Ros E, Zambon D. Postcholecystectomy symptoms. A prospective study of gallstone patients before and two years after surgery. Gut 1987;28:1500-4. 14. Traverso LW. Clinical manifestations and impact of gallstone disease. Am J Surg 1993;165:405-7. Arch Phys Med Rehabil Vol 81, August 2000
15. Charney KJ, Juler GL, Comarr E. General surgery problems in patients with spinal cord injuries. Arch Surg 1975;110:1083-8. 16. Moonka R, Steins SA, Eubank WB, Stelzner M. The presentation of gallstones and results of biliary surgery in a spinal cord injured population. Am J Surg 1999;178:246-50. 17. Banfield WJ. Physiology of the gallbladder. Gastroenterology 1975;69:770-7. 18. Ingberg HO, Prust FW. The diagnosis of abdominal emergencies in patients with spinal cord injuries. Arch Phys Med Rehabil 1968;49: 343-8. 19. Neumayer LA, Bull DA, Mohr JD, Putnam CW. The acutely affected abdomen in paraplegic spinal cord injury patients. Ann Surg 1990;212:561-6. 20. Han TR, Kim JH, Kwon BS. Chronic gastrointestinal problems and bowel dysfunction in patients with spinal cord injury. Spinal Cord 1998;36:485-90. 21. Koch JP, Donaldson RM. A survey of food intolerances in hospitalized patients. N Engl J Med 1964;271:657-60. 22. Talley NJ, McNeil D, Piper DW. Discriminant value of dyspeptic symptoms: a study of the clinical presentation of 221 patients with dyspepsia of unknown cause, peptic ulceration, and cholelithiasis. Gut 1987;28:40-6. 23. Kingston RD, Windsor CWO. Flatulent dyspepsia in patients with gallstones undergoing cholecystectomy. Br J Surg 1975;62:231-3. 24. Bates T, Ebbs SR, Harrison M, A’Hern RP. Influence of cholecystectomy on symptoms. Br J Surg 1991;78:964-7. 25. Rhind JA, Watson L. Gallstone dyspepsia. Br Med J 1968;1:32. 26. Hinkel CL, Moller GA. Correlation of symptoms, age, sex, and habitus with cholecystographic findings in 1,000 consecutive exams. Gastroenterology 1957;32:807-15. ¨ stberg H. 27. Janzon L, Aspelin S, Eriksson J, Hildell J, Trell E, O Ultrasonographic screening for gallstone disease in middle-aged women. Scand J Gastroenterol 1985;20:706-10. 28. Price WH. Gallbladder dyspepsia. Br Med J 1963;2:138-41. Supplier a. SPSS, 233 S Wacker Dr, Chicago, IL 60606.
APPENDIX Gallstone Survey Name: SSN: Mark the choice that best applies to you: 1. How often do you have abdominal pain? Never Once or twice a year Once or twice a month Once or twice a week Every day 2. Where in your abdomen is your pain? All over In the upper half, close to my rib cage, on the right In the upper half, close to my rib cage, on the left In the lower half, below my belly button, on the right In the lower half, below my belly button, on the left Location varies with each episode Hard to say 3. Do you find that your pain usually occurs after meals? Yes No 4. Does this pain ever wake you up at night? Yes No 5. How often do you have nausea or vomiting that you can’t explain? Never Once or twice a year Once or twice month
GALLSTONE SYMPTOMS IN SCI PATIENTS, Moonka
Once or twice a week Every day 6. How often do you have a feeling of your abdomen being overly full or bloated? Never Once or twice a year Once or twice a month Once or twice a week Every day 7. Have you ever had an x-ray suggesting you have gallstones? Yes No 8. If this x-ray was done at a facility besides the Seattle VA Medical Center, what was the name of the hospital, what city is it in, and as best as you can remember, when was it done?
1089
9. Have you had your gallbladder removed? Yes No If yes, any details of the operation you can recall would be helpful. (Which hospital, in which city, the approximate date of the surgery, whether it was emergency surgery, and how well you recovered.) 10. Have your parents, brothers and sisters, or children been diagnosed with gallstones? Yes No 11. How tall are you? 12. How much do you presently weigh? 13. Do you have diabetes? 14. If you have abdominal pain at least once a month, please mark on this diagram where it is. (Accompanied by anterior and posterior views of a homunculus).
Arch Phys Med Rehabil Vol 81, August 2000