- Email: [email protected]
Intestinal permeability changes in response to acetylsalicylic acid in relatives of patients with Crohn's disease
GASTROENTEROLOGY 1996;110:1395–1403
Intestinal Permeability Changes in Response to Acetylsalicylic Acid in Relatives of Patients With Crohn’s Disease ROBERT J. HILSDEN,*,‡ JON B. MEDDINGS,*,§ and LLOYD R. SUTHERLAND,*,‡ *Gastrointestinal Research Group, ‡Department of Community Health Sciences, and §Department of Medicine, University of Calgary, Calgary, Alberta, Canada
Background & Aims: Presence of a familial intestinal permeability defect in Crohn’s disease remains controversial despite numerous studies. The purpose of this study was to determine whether detection of a permeability defect in first-degree relatives of patients with Crohn’s disease can be enhanced using an acetylsalicylic acid provocation test. Methods: Lactulose-mannitol ratio, a measure of intestinal permeability, and total sucrose excretion, a measure of gastroduodenal permeability, were determined before and after ingestion of acetylsalicylic acid in healthy controls, in patients with Crohn’s disease, and in the first-degree relatives of patients with Crohn’s disease. Subjects were classified as hyperresponders if their results were above the mean / 2SD of the controls. Results: First-degree relatives had a 110% increase in intestinal permeability after acetylsalicylic acid compared with an increase of 57% in controls (P Å 0.001). Thirty-five percent of relatives were classified as hyperresponders. There was no significant difference in the change in sucrose excretion between relatives and controls (259% vs. 198%; P ú 0.05). Conclusions: First-degree relatives of patients with Crohn’s disease have an exaggerated increase in intestinal but not gastroduodenal permeability in response to acetylsalicylic acid. This study supports a familial permeability defect in Crohn’s disease, which may not be present in all families.
A
lthough the cause of Crohn’s disease remains unknown, evidence shows that both genetic and environmental factors are important. The evidence for a genetic predisposition includes striking ethnic differences in the frequency of Crohn’s disease, familial aggregation of Crohn’s disease, and higher concordance rate for monozygotic than for dizygotic twins.1 Ethnic differences include higher rates among whites than other racial groups; among white subjects, a higher incidence in Jewish subjects has been detected. A positive family history of Crohn’s disease is recognized as the most important risk factor for the disease. Siblings of affected individuals have / 5e0d$$0021
04-18-96 16:15:15
gasas
been estimated to have as high as a 30-fold increased risk of developing the disease compared with those without an affected sibling. The important role of genetics is further emphasized by twin studies. Tysk et al.,2 using a Swedish twin registry, found 44% of monozygotic twins concordant for Crohn’s disease compared with only 4% of dizygotic twins. Thompson et al.3 also found higher concordance rates for identical twins in a British population. The concordance rate for nonidentical twins was 5% compared with 22% for identical twins. However, despite the ample evidence supporting the role of genetics in the etiology of Crohn’s disease, the pathogenic mechanism by which a genetic predisposition might act remains unclear. Hollander et al.4 first suggested that an inherited defect in intestinal permeability could play a role in the pathogenesis of Crohn’s disease. Using the probe polyethylene glycol 400, they showed a twofold increase in intestinal permeability in healthy relatives of patients with Crohn’s disease compared with a healthy control group. However, whether relatives of patients with Crohn’s disease have abnormal intestinal permeability has been a contentious issue. Since the initial report by Hollander et al., several studies have provided conflicting results, although most failed to show evidence of a permeability defect.5 – 12 These studies used a variety of permeability probes and were generally designed to show a difference in the mean intestinal permeability between a group of controls and a group of healthy relatives. Our group has previously indicated that there could be two explanations for these conflicting reports. First, it would be unreasonable to assume that all relatives are affected with a familial permeability defect. After defining a normal range based on a healthy control group, we were able to show that approximately 10% of relatives Abbreviation used in this paper: 95% CI, 95% confidence interval. 䉷 1996 by the American Gastroenterological Association 0016-5085/96/$3.00
WBS-Gastro
1396 HILSDEN ET AL.
GASTROENTEROLOGY Vol. 110, No. 5
have abnormal intestinal permeability.9 However, others have argued that this proportion is within the variability of the test method.13 Second, the negative results of previous studies could be explained if differences in intestinal permeability between normals and relatives were relatively small. This would occur if only a subset of relatives had the defect or if the defect resulted in only a modest or intermittent increase in permeability. Very large sample sizes, larger than in any of the reported studies, would be required to detect small differences. In fact, when our data were combined with the data of Katz et al.,5 a statistically significant difference was found in the mean permeability between relatives and controls.9 All previous studies have used a single permeability test to characterize subjects. In 1992, Pironi et al.14 used lactulose-mannitol permeability tests before and after aspirin and found an exaggerated increase in intestinal permeability in first-degree relatives of patients with Crohn’s disease compared with healthy controls. In their study, the mean lactulose-mannitol ratio in healthy controls increased from 0.049 to 0.054 after aspirin, whereas in healthy relatives it increased from 0.032 to 0.090. The study was published only in abstract form and included only 5 subjects in each group. The issue has not been subsequently addressed. We hypothesized that a provocative intestinal permeability test using acetylsalicylic acid would allow enhanced discrimination between relatives with a familial permeability defect and controls. Our primary objective was to determine if first-degree relatives of patients with Crohn’s disease have a greater increase in intestinal permeability after acetylsalicylic acid administration than healthy controls. Our secondary objective was to determine if the permeability defect was generalized by examining gastroduodenal permeability before and after provocation with acetylsalicylic acid. Because Crohn’s disease predominantly affects the small and large intestine and uncommonly the stomach, a permeability defect might be limited to the intestines. We have previously developed and validated a sucrose permeability test to assess gastroduodenal permeability.15,16 Gastroduodenal permeability has not been tested previously in patients with Crohn’s disease or their relatives.
Materials and Methods Study Subjects Thirty patients with Crohn’s disease and 37 first-degree relatives of patients with Crohn’s disease were recruited through the outpatient gastroenterology service at the University of Calgary. All patients were in clinical remission (Crohn’s
/ 5e0d$$0021
04-18-96 16:15:15
gasas
Disease Activity Index of õ150).17 None of the relatives had gastrointestinal symptoms. The control group consisted of 26 healthy individuals with no gastrointestinal symptoms. Individuals were excluded if they regularly used acetylsalicylic acid or nonsteroidal anti-inflammatory drugs (more than once a week); had a history of intolerance, allergy, or other contraindication to acetylsalicylic acid; or had a history of diabetes mellitus, renal insufficiency, or peptic ulcer disease. The study protocol was approved by the University of Calgary Ethics Committee.
Study Protocol Before testing, all participants completed a questionnaire about past and current medical problems, gastrointestinal symptoms, current medications, and tobacco and alcohol use. Subjects were instructed not to consume alcohol or nonsteroidal anti-inflammatory drugs for at least 24 hours before the permeability tests. For baseline permeability tests, a test solution was consumed at bedtime after at least a 3-hour fast. The test solution contained 5 g lactulose (Technilab, Montreal, Quebec, Canada), 2 g mannitol (BDH Chemicals Inc., Toronto, Ontario, Canada), and 100 g sucrose in 350 mL of water. Subjects voided before drinking the test solution and then collected all urine, including the first morning void, in a preweighed container containing 5 mL of 10% thymol in methanol. All subjects underwent two permeability tests, a baseline test and a post–acetylsalicylic acid test. Five patients with Crohn’s disease with an increased baseline lactulose-mannitol ratio were excluded from undergoing a post–acetylsalicylic acid test, and their results are excluded from the analysis. Post–acetylsalicylic acid permeability tests were performed after two 1.3-g doses of acetylsalicylic acid (Apotex, Toronto, Ontario, Canada), the first at noon and the second 1 hour before drinking the test solution. The post–acetylsalicylic acid test was performed within 10 days of the baseline test, and urine was collected during a similar period as in the baseline test.
Analytical Methods Urine samples (10 mL) were deionized by adding 1 g of a 1:1.5 (wt/wt) mixture of Amberlite IR-120 and IRA-400 resin (BDH Chemicals Inc.). After centrifugation at 3000 rpm for 10 minutes, the supernatant was filtered through a 40mm/L Millipore filter (Millipore, Bedford, MA). Samples were separated on a Dionex MA-1 anion exchange column by highperformance liquid chromatography (Dionex, Oakville, Ontario, Canada) at room temperature using 500 mmol/L NaOH as the isocratic mobile phase. Peak identification was accomplished with the use of authentic standards and detected using pulsed amperometric electrochemical detection on a gold electrode. Samples were diluted as necessary after addition of cellobiose as an internal standard. Quantitation was performed using known standards at multiple concentrations with linear interpolation between concentrations. All samples were diluted
WBS-Gastro
May 1996
INTESTINAL PERMEABILITY IN CROHN’S DISEASE 1397
so that concentrations of interest fell within the range of the standards. Two permeability measures were calculated from the urine sugar concentrations: the lactulose-mannitol ratio as a measure of intestinal permeability and total sucrose excretion as a measure of gastroduodenal permeability. The lactulose-mannitol ratio is expressed as the ratio of the fractional excretion of lactulose to the fractional excretion of mannitol. As a ratio of two different permeability probes, it has the advantage of controlling for effects of premucosal and postmucosal factors, such as gastric emptying and intestinal transit time, that can affect permeability measures using only single probes.13 Although there are some differences in the rates of urinary excretion of different probes, most of the difference is observed in the early part of the collection period; by 7 hours, the minimum collection period in the present study, this difference is minimal.18 Total sucrose excretion is expressed as total milligrams of sucrose excreted in the urine. Data for each group are expressed as the mean and SEM.
Statistical Analysis Initial sample size estimates were based on the results of Pironi et al.14 The mean intestinal permeability for the control group increased by about 10%, and that of the relatives increased by 180%. Using a conservative estimate of an expected 20% increase in healthy controls and a 90% increase in relatives, a sample size of 32 subjects in each group was required to allow the detection of this difference with a power of 90% and an a of 0.05. Data analysis was performed using SPSS for Windows (SPSS Inc., Chicago, IL). All outcomes were first examined with descriptive statistics and box plots. All statistical tests were two sided using an a of 0.05. Hypothesis testing was not performed to compare the patient group with the other groups because the patient group was specially selected (i.e., all in remission and those with abnormal baseline lactulose-mannitol ratio excluded) and was therefore not considered to be representative of the general Crohn’s disease population. Descriptive statistics and graphical representation of the patient group are provided only to allow comparison. The primary outcome of interest was whether there was a difference in the change in intestinal permeability after acetylsalicylic acid between the group of relatives and the control group. This was examined in two ways: as the increase above the baseline lactulose-mannitol ratio after aspirin and as the percent change in the lactulose-mannitol ratio after aspirin. Both methods were used because it is unknown which is the most meaningful measure for assessing changes in permeability: i.e., is a change in a lactulose-mannitol ratio from 0.010 to 0.020 the same as a change from 0.040 to 0.050? They both represent an increase of 0.010, but in one case is an increase of 100%, whereas in the other only an increase of 25%. Differences between the two groups were tested using
/ 5e0d$$0021
04-18-96 16:15:15
gasas
Table 1. Demographic Characteristics of the Subject Groups
Mean age ( yr ) Women/men Current smokers
Patients
Relatives
Controls
32 15/10 4 (16%)
44 19/18 7 (21%)
40 13/13 1 (4%)
analysis of covariance with the baseline lactulose-mannitol ratio as a covariate. Using the data for the control group, two upper reference limits for the increase and the percent change in the lactulosemannitol ratio after acetylsalicylic acid were calculated to define hyperresponders. A standard upper reference limit was calculated as the mean / 2SD, and a conservative upper reference limit was calculated as the mean / 3SD. Each subject was classified as being above or below these upper limits. The proportion and its 95% confidence interval (95% CI) of relatives with results above these reference limits (hyperresponders) were then calculated. Secondary analysis examined gastroduodenal permeability using baseline and post–acetylsalicylic acid total sucrose excretion. This analysis was performed using the same methods described above. The difference between the post–acetylsalicylic acid change in sucrose permeability in the group of relatives and the control group was assessed by analysis of covariance using the baseline total sucrose excretion as a covariate. The concordance of responsiveness to acetylsalicylic acid among family members was assessed. Each pair of a patient with Crohn’s disease and the relative participating in the study was categorized as being either concordant or discordant based on having an exaggerated increase in intestinal permeability in response to acetylsalicylic acid, defined by a percent increase of greater than the mean / 2SDs of the controls. A post hoc decision was made to assess concordance, and only a limited number of patient-relative pairs were available for analysis; therefore, these particular findings should be interpreted cautiously.
Results Table 1 shows the characteristics of the three subject groups. Mean baseline lactulose-mannitol ratios were 0.0177 (SEM, 0.0010) for the relatives and 0.0174 (SEM, 0.0011) for controls. Results for subjects with Crohn’s disease are not given because subjects with a high baseline permeability were excluded from analysis. The difference between the mean lactulose-mannitol ratios of the group of relatives and control group was not significantly different (P Å 0.9). Two controls and 3 relatives had values greater than the standard upper reference limit of 0.0281 (mean / 2SD). Figure 1 shows individual paired data for baseline and post–acetylsalicylic acid lactulose-mannitol ratios for relWBS-Gastro
1398 HILSDEN ET AL.
GASTROENTEROLOGY Vol. 110, No. 5
Figure 1. Paired results for baseline and post–acetylsalicylic acid (ASA) lactulose-mannitol ratios for relatives of patients with Crohn’s disease and for healthy controls; no significant difference in baseline ratio was observed for the two groups. After acetylsalicylic acid treatment, the controls had a 57% increase, whereas the relatives had a 110% increase (P Å 0.001). In contrast to controls, some relatives were much more responsive to the effects of aspirin than others.
atives and controls. Individuals in the control group had a slight increase in their lactulose-mannitol permeability in response to acetylsalicylic acid, but several relatives of patients with Crohn’s disease had increases of a much greater degree (hyperresponders). Figure 2 shows the increase above the baseline lactulose-mannitol ratio after ingestion of acetylsalicylic acid. Mean increases in the lactulose-mannitol ratios were 0.0136 (SEM, 0.0021), 0.0184 (0.0026), and 0.0093 (0.0013) for patients, relatives, and controls, respectively. There was a statistically significant difference in the mean change in lactulose-mannitol ratio between relatives and controls (95% CI for the difference, 0.003–0.015; P Å 0.003). In contrast to the baseline test results, which were similar between the relatives and controls, the two groups differed dramatically after exposure to acetylsalicylic acid. The control group responded relatively uniformly to the acetylsalicylic acid. In contrast, the relatives had a much wider distribution; some were much more responsive to the effects of aspirin than others. Also, the pattern observed in the relatives and patients was quite similar and was different from that observed in the controls. The proportion of relatives who showed hyperresponsiveness to aspirin was estimated after calculating the standard (mean / 2SD) and conservative (mean / 3SD) upper reference limits for the post–acetylsalicylic acid increase in the lactulose-mannitol ratio. Twenty-seven / 5e0d$$0021
04-18-96 16:15:15
gasas
percent (95% CI, 14%–44%) of relatives had post–acetylsalicylic acid increases greater than the standard upper reference limit of 0.0230; 14% (95% CI, 5%–29%) of relatives had increases greater than the conservative upper reference limit of 0.030. Figure 3 shows the percent change in the lactulosemannitol ratio after acetylsalicylic acid. Post–acetylsalicylic acid changes in intestinal permeability, measured as the percent change, gave similar results when the change in permeability was measured as the absolute increase. Mean percent changes in the lactulose-mannitol ratios were 133% (SEM, 25%), 110% (13%), and 57% (9%) for patients, relatives, and controls, respectively. There was also a statistically significant difference in the mean percent change between relatives and controls (95% CI for the difference, 21%–85%; P Å 0.001). The standard upper reference limit for the post–acetylsalicylic acid percent change in the lactulose-mannitol ratio was 143%, and the conservative upper reference limit was 186%. Thirty-five percent of relatives (95% CI, 20%–53%) were classified as hyperresponders using the standard upper reference limit; 19% of relatives (95% CI, 8%–35%) were classified as hyperresponders using the conservative upper reference limit. There were more current smokers in the group of
Figure 2. One-way scatterplots of the increase in the lactulose-mannitol ratio after acetylsalicylic acid treatment in the three subject groups. The dotted line represents the standard upper reference limit (mean / 2SD of the controls), and the dashed line represents the conservative upper reference limit (mean / 3SD). The range of values is much wider for the relatives than for controls. Ten relatives had post–acetylsalicylic acid increases greater than the standard reference limit, and 5 had increases greater than the conservative reference limit. There is a significant difference in the mean change between the relatives and controls (P Å 0.003).
WBS-Gastro
May 1996
INTESTINAL PERMEABILITY IN CROHN’S DISEASE 1399
Figure 3. One-way scatterplots of percent change in the lactulosemannitol ratio after acetylsalicylic acid treatment for the three subject groups. The dotted and dashed lines represent standard and conservative upper reference limits, respectively. Thirteen relatives had post–acetylsalicylic acid percent increases greater than the standard reference limit, and 7 had increases greater than the conservative reference limit. The range of results is much wider for the relatives than for controls. There is a significant difference in the mean percent change between the relatives and controls (P Å 0.001).
relatives than in the control group. However, smoking status did not seem to be related to changes in permeability after acetylsalicylic acid treatment. When current smokers were excluded from the analysis, there was still a statistically significant difference between relatives and controls. Thus, for nonsmoking relatives the mean post– acetylsalicylic acid change in lactulose-mannitol ratio was 0.0180 compared with 0.0095 for nonsmoking controls (P Å 0.006). A limited number of patient-relative pairs participated in the study (8 patients with 12 relatives). For these 12 pairs, concordance of permeability changes in response to acetylsalicylic acid was assessed using the percent change data. Five patients had hyperresponsive changes in permeability after acetylsalicylic acid treatment, and 7 of their 8 relatives also had abnormal increases. Three patients did not have hyperresponsive post–acetylsalicylic acid changes, and all 4 of their relatives also had normal changes. Examples of concordance patterns for three families are shown in Figure 4. Sucrose excretion was used to assess gastroduodenal permeability. Mean baseline sucrose excretion was 105 mg (SEM, 11 mg) for the relatives and 88 mg (SEM, 10 mg) for controls. Although the relatives had a higher baseline sucrose excretion than controls, this difference was not statistically significant (P Å 0.26). / 5e0d$$0021
04-18-96 16:15:15
gasas
The effect of acetylsalicylic acid on sucrose excretion was assessed in the same way as the lactulose-mannitol ratio data. In contrast to the results for intestinal permeability, the distributions of the results for the relatives and controls are much more similar, suggesting a similar response to acetylsalicylic acid. Mean increases in total sucrose excretion were 169 mg (SEM, 26 mg), 234 mg (40 mg), and 127 mg (22 mg) for patients, relatives, and controls, respectively. The mean increase for the relatives is influenced by a single markedly high result. Mean percent changes in total sucrose excretion were 290% (SEM, 46%), 259% (40%), and 198% (39%) for patients, relatives, and controls, respectively (Figure 5). Although the results for post–acetylsalicylic acid sucrose excretion are higher for the relatives, when considering baseline sucrose excretion as a covariate, there was no significant difference between relatives and controls in the post– acetylsalicylic acid increase or percent change in sucrose excretion. When relatives classified as hyperresponders or normal responders, based on the post–acetylsalicylic acid increase in lactulose-mannitol ratio, were examined, there was no difference in the baseline lactulose-mannitol ratio between the two groups (95% CI for the difference, 00.007 to 0.03). The baseline total sucrose excretion was 90 mg for the normal responders and 145 mg for the hyperresponders. However, this difference did not reach statistical significance (95% CI, 0130 to 19). The change in sucrose excretion also did not differ between the two groups (P Å 0.15) when baseline sucrose excretion was considered as a covariate.
Figure 4. Three pedigrees showing the concordance pattern of baseline and post–acetylsalicylic acid lactulose-mannitol permeability results among family members with and without Crohn’s disease. Although these are small pedigrees, the degree of concordance is quite striking. 䊊, Normal female; 䊐, normal male; 䊊 / , Crohn’s disease; , abnormal baseline lactulose-mannitol ratio; , abnormal increase post–acetylsalicylic acid.
WBS-Gastro
1400 HILSDEN ET AL.
GASTROENTEROLOGY Vol. 110, No. 5
Figure 5. One-way scatterplots of percent change in total sucrose excretion after acetylsalicylic acid treatment for the three subject groups. The distribution of results is similar between the groups. There was no significant difference between relatives and controls.
Discussion Extensive evidence shows that Crohn’s disease has a familial predisposition. This evidence includes striking ethnic differences in the frequency of Crohn’s disease, familial aggregation of Crohn’s disease, and higher concordance rate for monozygotic than for dizygotic twins. Not only has familiality been linked to the risk for Crohn’s disease but also to disease characteristics, such as age at onset, location of disease, and perforating disease.19,20 The mechanism of action of a genetic predisposition remains unclear. Previous studies have examined the complement system, immunoglobulins, mucin, and the fecal flora. Because increased intestinal permeability has been found in patients with Crohn’s disease,21 – 23 a genetically determined permeability defect is an attractive explanation. Possibly, impairment of the barrier function of the intestine could lead to increased exposure of the mucosal immune system to luminal antigens, resulting in an inflammatory reaction.24 Hollander et al.4 showed increased intestinal permeability in healthy relatives compared with controls, but most subsequent studies have failed to confirm this. Standard lactulose-mannitol permeability tests may not be sufficiently sensitive to discriminate among individuals with and without a permeability defect. Comparing differences in standard permeability test results between relatives and controls may be prone to error because only a subset of relatives may have a permeability defect. The modest increase in permeability induced by / 5e0d$$0021
04-18-96 16:15:15
gasas
a defect may be lost when only the mean value for a group is examined. We previously suggested that approximately 10% of relatives have abnormal intestinal permeability.9 Furthermore, our study suggests that many patients with inactive Crohn’s disease have normal lactulose-mannitol permeability when it is tested in the standard way. These individuals would have a permeability defect if it exists. Although it is unknown what these patients’ intestinal permeability would be if they were not affected by Crohn’s disease, clearly it would not differ greatly from that of normal subjects. Therefore, standard lactulose-mannitol permeability tests appear to be an inadequate tool for evaluating relatives to detect a familial permeability defect. In this study, we show that healthy first-degree relatives of patients with Crohn’s disease have an exaggerated increase in intestinal permeability after provocation with acetylsalicylic acid. A provocation test provides enhanced discrimination if it increases the difference between normals and abnormals. All subjects had increases in intestinal permeability after acetylsalicylic acid administration, but relatives (and patients) had greater increases than controls. One would expect that only a proportion of relatives would have a permeability defect compared with none of the controls. Our findings support this because the relatives had a much greater range of values for the post–acetylsalicylic acid change in permeability than the controls (Figures 2 and 3). What proportion of relatives may have a permeability defect? Determining this depends on what is considered normal. Using the data for the control group, we have created upper reference limits for changes in the lactulose-mannitol ratio after acetylsalicylic acid administration. Because standard reference limits calculated as the mean / 2SD using sample sizes õ100 can be quite imprecise, with 95% CIs as wide as 1SD of the data, we calculated a conservative upper reference limit, which is the mean / 3SD.25 Values above this would be clearly abnormal. We found that 14% and 19% of the relatives had values greater than the conservative upper reference limit for post–acetylsalicylic acid increase and percent change in lactulose-mannitol ratio, respectively. We calculated both the change and percent change because it is unclear which is the more meaningful measure of changes in permeability. As discussed above, an increase in the lactulose-mannitol ratio of 0.01 may not mean the same thing in an individual with a baseline ratio of 0.01 as it does in an individual with a baseline ratio of 0.04. Overall, both measures gave similar results, which strengthens the validity of the findings. Confounding factors do not seem to be responsible for WBS-Gastro
May 1996
INTESTINAL PERMEABILITY IN CROHN’S DISEASE 1401
the study’s findings. As a group, the relatives are slightly older than the controls, but increasing age is not associated with increases in the lactulose-mannitol ratio.26 There was a higher proportion of smokers among the relatives, but the prevalence of permeability defects was equal among smokers and nonsmokers, and the relatives still had a statistically significant greater post–acetylsalicylic acid increase in lactulose-mannitol ratio when smokers were excluded. Finally, the effects of any environmental factors would be present in both the baseline and post–acetylsalicylic acid permeability tests and therefore should have been canceled when assessing permeability changes. Because of the specific selection of subjects with Crohn’s disease, we provided their results only to allow qualitative comparisons; however, some interesting observations can be made. As noted already, most patients with inactive Crohn’s disease have normal intestinal permeability. However, after provocation with acetylsalicylic acid, not all patients have an exaggerated increase in the lactulose-mannitol ratio. In fact, the distribution of the patients’ results is quite similar to that of the relatives (Figures 2 and 3), possibly suggesting that not all patients with Crohn’s disease have a familial permeability defect. This impression is strengthened by noting the concordance of normal or abnormal increases in permeability within families (Figure 5). Although only a limited number of patient-relative pairs was available, there was a striking degree of concordance among relatives. This has also been found in studies of other possible genetic markers.27 Recently, Peeters et al.28 challenged the concept that a permeability defect in relatives of patients with Crohn’s disease is genetic in origin. Using standard lactulosemannitol permeability tests in 17 complete Crohn’s families, they were unable to show a specific familial pattern to abnormal intestinal permeabilities. They also found that 4 of 13 spouses had abnormal permeabilities. They concluded that increased permeability could be better explained by an environmental factor. Our study was not designed to specifically address whether a permeability defect resulted from genetic or environmental factors. However, none of the 3 spouses participating in the study as controls had abnormal post– acetylsalicylic acid permeability changes. Furthermore, as mentioned above, there was a high concordance among family members in response patterns. This could result from a shared environmental factor; however, many of these individuals lived in different cities, making this less likely. It is possible that a permeability defect results from / 5e0d$$0021
04-18-96 16:15:15
gasas
an early life exposure that is not necessarily present when the subjects are studied. Previous studies have suggested that childhood factors, such as passive smoking and early weaning, may be important in the etiology of Crohn’s disease.29,30 Also, cow’s milk feeding has been associated with altered intestinal permeability in newborn infants. Weaver et al. found that compared with newborn infants who were breast-fed, those who received cow’s milk feeds had persistently higher intestinal permeability throughout the first week of life.31 However, it has not been shown that a permeability defect persists through later infancy or adulthood. Alternatively, those with a familial permeability defect may be more susceptible to luminal antigens or toxins that may lead to Crohn’s disease. We also examined gastric permeability using a sucrose permeability test. Sucrose permeability has not been examined previously in patients with Crohn’s disease or their relatives. Again, baseline gastroduodenal permeability seems to be similar in all groups. Acetylsalicylic acid caused a substantial increase in gastroduodenal permeability, but, in contrast to intestinal permeability, patients with Crohn’s disease and their relatives did not have an exaggerated increase compared with controls. This suggests that any permeability defect is restricted to the intestine. Alternatively, the dose of aspirin may have been too large, resulting in substantial increases regardless of an underlying defect, limiting the power of this study to detect a difference. However, this finding is consistent with the unusual occurrence of gastric or duodenal Crohn’s disease. We now have evidence that the first-degree relatives of patients with Crohn’s disease can be characterized by the following intestinal permeability patterns: (1) normal permeability, (2) abnormal baseline permeability, (3) exaggerated increase in permeability in response to acetylsalicylic acid, and (4) both an abnormal baseline permeability and an exaggerated increase in response to acetylsalicylic acid. One or more of these permeability patterns may characterize individuals who are predisposed to develop Crohn’s disease. However, there is no evidence that individuals with a permeability defect are at increased risk for developing Crohn’s disease. It might be concluded from this study, in which a large number of adults had a permeability defect but no evidence of Crohn’s disease, that it does not. However, cross-sectional studies will never be able to resolve this question. In studies such as this, subjects have already declared whether they will develop Crohn’s disease or not. A longitudinal study in which individuals are tested and then followed up for a prolonged period for the development of Crohn’s disease would be required. Because familiality WBS-Gastro
1402 HILSDEN ET AL.
GASTROENTEROLOGY Vol. 110, No. 5
has also been associated with young age at onset, such a study would have to recruit subjects as young children. Clearly, such a study is unlikely. This study suggests that there may be genetic heterogeneity among patients with Crohn’s disease: some patients are hyperresponders, some are not. This is consistent with other studies suggesting genetic heterogeneity in Crohn’s disease.27,32 Crohn’s disease is also characterized by clinical heterogeneity. A patient’s disease can be characterized by its anatomic location, anatomic extent, and disease behavior.33 For example, Greenstein et al. have suggested that perforating and nonperforating types are distinct patterns of disease.34 Possibly clinical disease patterns are associated with a particular permeability pattern. Crohn’s disease is a genetically and clinically complex disease. We should not be surprised that finding a genetic marker of the disease is difficult or that potential markers are not found in all patients or in all relatives. There is evidence that a permeability defect is present in some patients with Crohn’s disease and some of their relatives. Whether a familial permeability defect predisposes to Crohn’s disease is likely to remain unknown. Future studies should be directed at addressing unresolved issues: whether a permeability defect results from genetics or the environment and whether permeability defects are associated with clinical phenotypes.
9.
10.
11.
12.
13. 14.
15.
16.
17.
18.
References 1. Calkins BM, Mendeloff AI. The epidemiology of idiopathic inflammatory bowel disease. In: Kirsher JB, Shorter RS, eds. Inflammatory bowel disease. 4th ed. Baltimore, MD: Williams & Wilkins, 1995:31–68. 2. Tysk C, Lindberg E, Ja¨rnerot G, Flode´rus-Myrhed B. Ulcerative colitis and Crohn’s disease in an unselected population of monozygotic and dizygotic twins. A study of heritability and the influence of smoking. Gut 1988;29:990–996. 3. Thompson NP, Driscoll R, Pounder RE, Wakefield AJ. Genetics vs. environment in inflammatory bowel disease: concordance rates in 130 twin pairs (abstr). Gastroenterology 1995;108: A928. 4. Hollander D, Vadheim CM, Brettholz E, Petersen GM, Delahunty, T, Rotter JI. Increased intestinal permeability in patients with Crohn’s disease and their relatives. Ann Intern Med 1986;105: 883–885. 5. Katz KD, Hollander D, Vadheim CM, McElree C, Delahunty T, Dadufalza VD, Krugliak P, Rotter JI. Intestinal permeability in patients with Crohn’s disease and their healthy relatives. Gastroenterology 1989;97:927–931. 6. Teahon K, Smethurst P, Levi AJ, Menzies IS, Bjarnason I. Intestinal permeability in patients with Crohn’s disease and their first degree relatives. Gut 1992;33:320–323. 7. Ainsworth M, Erikson J, Rasmussen JW, Schaffalitzky de Muckadell OB. 51Cr-Labelled ethylenediaminetetraacetic acid in patients with Crohn’s disease and their healthy relatives. Scand J Gastroenterol 1985;24:993–998. 8. Ruttenberg D, Young GO, Wright JP, Isaacs S. Peg-400 excretion
/ 5e0d$$0021
04-18-96 16:15:15
gasas
19.
20.
21.
22.
23.
24.
25. 26.
27.
in patients with Crohn’s disease, their first-degree relatives, and healthy volunteers. Dig Dis Sci 1992;37:705–708. May GR, Sutherland LR, Meddings JB. Is small intestinal permeability really increased in relatives of patients with Crohn’s disease? Gastroenterology 1993;104:1627–1632. Munkholm P, Langholz E, Hollander D, Thornberg K, Orholm M, Katz KD, Binder V. Intestinal permeability in patients with Crohn’s disease and ulcerative colitis and their first degree relatives. Gut 1994;35:68–72. Howden CW, Gillanders I, Morris AJ, Duncan A, Danesh B, Russell RI. Intestinal permeability in patients with Crohn’s disease and their first-degree relatives. Am J Gastroenterol 1994;89:1175– 1176. Peeters M, Ghoos Y, Geypens B, Hiele M, Rutgeerts P. Lactulose/mannitol (L/M) permeability index is a suitable subclinical marker to identify subjects at risk in families of Crohn’s patients (abstr). Gastroenterology 1994;106:A750. Bjarnason I, MacPherson A, Hollander D. Intestinal permeability: an overview. Gastroenterology 1995;108:1566–1581. Pironi L, Miglioli M, Ruggeri E, Dallasta MA, Ornigotti L, Valpiani D, Barbara L. Effect of non-steroidal anti-inflammatory drugs (NSAID) on intestinal permeability of first degree relatives of patients with Crohn’s disease (abstr). Gastroenterology 1992; A679. Meddings JB, Sutherland LR, Byles NI, Wallace JL. Sucrose: a novel permeability marker for gastroduodenal disease. Gastroenterology 1993;104:1619–1626. Sutherland LR, Verhoef M, Wallace JL, Van Rosendaal G, Crutcher R, Meddings JB. A simple, non-invasive marker of gastric damage: sucrose permeability. Lancet 1994;343:998–1000. Best WR, Bechtel JM, Singleton JW. Rederived values of the eight coefficients of the Crohn’s disease activity index (CDAI). Gastroenterology 1979;77:843–846. Bjarnason I, Maxton D, Reynolds AP, Catt S, Peters TJ, Menzies IS. Comparison of four markers of intestinal permeability in control subjects and patients with coeliac disease. Scand J Gastroenterol 1994;29:630–639. Tokayer AZ, Reydel B, Bayless TM. Possible role of heredity in site and transmural aggressiveness of Crohn’s disease (abstr). Gastroenterology 1992;102:A705. Gower-Rousseau C, Plegat S, Evrard JP, Grandbastien B, Hugot JP, Dupas JL, Gendre JP, Modigliani R, Be´laı¨che J, Hostein J, Cortot A, Colombel JF. Familiality in clinical characteristics of Crohn’s disease in 69 French families (abstr). Gastroenterology 1995;108:A826. Ukabam SO, Clamp JR, Cooper BT. Abnormal small intestinal permeability to sugars in patients with Crohn’s disease of the terminal ileum and colon. Digestion 1983;27:70–74. Olaison G, Leandersson P, Sjo¨dahl R, Tagesson C. Intestinal permeability to polyethylene glycol 600 in Crohn’s disease. Preoperative determination in a defined segment of small intestine. Gut 1988;29:196–199. Lochs H, Wyatt J, Reinisch W, Gasche´ C, Maier T, Moser G, Gangl A. Clinical relevance of intestinal permeability in active Crohn’s disease (abstr). Gastroenterology 1994;106:A722. Hollander D. Permeability in Crohn’s disease: altered barrier functions in healthy relatives? Gastroenterology 1993;104:1848– 1851. Altman DG. Practical statistics for medical research. London: Chapman & Hall, 1991. Saltzman JR, Knowdley KV, Perrone G, Russell RM. Changes in small-intestine permeability with aging. J Am Geriatr Soc 1995; 43:160–164. Elmgreen J, Both H, Binder V. Familial occurrence of complement dysfunction in Crohn’s disease: correlation with intestinal symp-
WBS-Gastro
May 1996
28.
29.
30.
31.
32.
INTESTINAL PERMEABILITY IN CROHN’S DISEASE 1403
toms and hypercatabolism of complement. Gut 1985;26:151– 157. Peeters M, Geypens B, Ghoos Y, Nevens H, D’Haens G, Hiele M, Rutgeerts P. The pattern of increased small intestinal permeability abnormalities in families with Crohn’s disease: genetic or non-genetic (abstr)? Gastroenterology 1995;108:A892. Lashner BA, Shaheen NJ, Hanauer SB, Kirschner BS. Passive smoking is associated with an increased risk of developing inflammatory bowel disease in children. Am J Gastroenterol 1993; 88:356–359. Bergstrand O, Hellers G. Breast-feeding during infancy in patients who later develop Crohn’s disease. Scand J Gastroenterol 1983; 18:903–906. Weaver LT, Laker MF, Nelson R, Lucas A. Milk feeding and changes in intestinal permeability and morphology in the newborn. J Pediatr Gastroenterol Nutr 1987;6:351–358. Polito J, Tokayer A, Childs B, Harris M, Bayless T. Crohn’s dis-
/ 5e0d$$0021
04-18-96 16:15:15
gasas
ease: evidence for genetic heterogeneity and ’’anticipation‘‘ (abstr). Gastroenterology 1995;108:A895. 33. Sachar DB, Andrews H, Farmer RG, Pallone F, Pen ˜a AS, Prantera C, Rutgeerts P. Proposed classification of patient subgroups in Crohn’s disease. Working team report 4. Gastroenterol Int 1992; 3:141–154. 34. Greenstein AJ, Lachman P, Sachar DB, Springhorn J, Heimann T, Janowitz HD, Aufses AHJ. Perforating and non-perforating indications for repeated operations in Crohn’s disease: evidence for two clinical forms. Gut 1988;29:588–592. Received August 25, 1995. Accepted January 4, 1996. Address requests for reprints to: Lloyd Sutherland, M.D., Department of Community Health Sciences, 3330 Hospital Drive Northwest, Calgary, Alberta, Canada T2N 4N1. Fax: (403) 270-7307. Supported by a clinical fellowship from the Alberta Heritage Foundation for Medical Research (to R.J.H.).
WBS-Gastro
Intestinal Permeability Changes in Response to Acetylsalicylic Acid in Relatives of Patients With Crohn’s Disease ROBERT J. HILSDEN,*,‡ JON B. MEDDINGS,*,§ and LLOYD R. SUTHERLAND,*,‡ *Gastrointestinal Research Group, ‡Department of Community Health Sciences, and §Department of Medicine, University of Calgary, Calgary, Alberta, Canada
Background & Aims: Presence of a familial intestinal permeability defect in Crohn’s disease remains controversial despite numerous studies. The purpose of this study was to determine whether detection of a permeability defect in first-degree relatives of patients with Crohn’s disease can be enhanced using an acetylsalicylic acid provocation test. Methods: Lactulose-mannitol ratio, a measure of intestinal permeability, and total sucrose excretion, a measure of gastroduodenal permeability, were determined before and after ingestion of acetylsalicylic acid in healthy controls, in patients with Crohn’s disease, and in the first-degree relatives of patients with Crohn’s disease. Subjects were classified as hyperresponders if their results were above the mean / 2SD of the controls. Results: First-degree relatives had a 110% increase in intestinal permeability after acetylsalicylic acid compared with an increase of 57% in controls (P Å 0.001). Thirty-five percent of relatives were classified as hyperresponders. There was no significant difference in the change in sucrose excretion between relatives and controls (259% vs. 198%; P ú 0.05). Conclusions: First-degree relatives of patients with Crohn’s disease have an exaggerated increase in intestinal but not gastroduodenal permeability in response to acetylsalicylic acid. This study supports a familial permeability defect in Crohn’s disease, which may not be present in all families.
A
lthough the cause of Crohn’s disease remains unknown, evidence shows that both genetic and environmental factors are important. The evidence for a genetic predisposition includes striking ethnic differences in the frequency of Crohn’s disease, familial aggregation of Crohn’s disease, and higher concordance rate for monozygotic than for dizygotic twins.1 Ethnic differences include higher rates among whites than other racial groups; among white subjects, a higher incidence in Jewish subjects has been detected. A positive family history of Crohn’s disease is recognized as the most important risk factor for the disease. Siblings of affected individuals have / 5e0d$$0021
04-18-96 16:15:15
gasas
been estimated to have as high as a 30-fold increased risk of developing the disease compared with those without an affected sibling. The important role of genetics is further emphasized by twin studies. Tysk et al.,2 using a Swedish twin registry, found 44% of monozygotic twins concordant for Crohn’s disease compared with only 4% of dizygotic twins. Thompson et al.3 also found higher concordance rates for identical twins in a British population. The concordance rate for nonidentical twins was 5% compared with 22% for identical twins. However, despite the ample evidence supporting the role of genetics in the etiology of Crohn’s disease, the pathogenic mechanism by which a genetic predisposition might act remains unclear. Hollander et al.4 first suggested that an inherited defect in intestinal permeability could play a role in the pathogenesis of Crohn’s disease. Using the probe polyethylene glycol 400, they showed a twofold increase in intestinal permeability in healthy relatives of patients with Crohn’s disease compared with a healthy control group. However, whether relatives of patients with Crohn’s disease have abnormal intestinal permeability has been a contentious issue. Since the initial report by Hollander et al., several studies have provided conflicting results, although most failed to show evidence of a permeability defect.5 – 12 These studies used a variety of permeability probes and were generally designed to show a difference in the mean intestinal permeability between a group of controls and a group of healthy relatives. Our group has previously indicated that there could be two explanations for these conflicting reports. First, it would be unreasonable to assume that all relatives are affected with a familial permeability defect. After defining a normal range based on a healthy control group, we were able to show that approximately 10% of relatives Abbreviation used in this paper: 95% CI, 95% confidence interval. 䉷 1996 by the American Gastroenterological Association 0016-5085/96/$3.00
WBS-Gastro
1396 HILSDEN ET AL.
GASTROENTEROLOGY Vol. 110, No. 5
have abnormal intestinal permeability.9 However, others have argued that this proportion is within the variability of the test method.13 Second, the negative results of previous studies could be explained if differences in intestinal permeability between normals and relatives were relatively small. This would occur if only a subset of relatives had the defect or if the defect resulted in only a modest or intermittent increase in permeability. Very large sample sizes, larger than in any of the reported studies, would be required to detect small differences. In fact, when our data were combined with the data of Katz et al.,5 a statistically significant difference was found in the mean permeability between relatives and controls.9 All previous studies have used a single permeability test to characterize subjects. In 1992, Pironi et al.14 used lactulose-mannitol permeability tests before and after aspirin and found an exaggerated increase in intestinal permeability in first-degree relatives of patients with Crohn’s disease compared with healthy controls. In their study, the mean lactulose-mannitol ratio in healthy controls increased from 0.049 to 0.054 after aspirin, whereas in healthy relatives it increased from 0.032 to 0.090. The study was published only in abstract form and included only 5 subjects in each group. The issue has not been subsequently addressed. We hypothesized that a provocative intestinal permeability test using acetylsalicylic acid would allow enhanced discrimination between relatives with a familial permeability defect and controls. Our primary objective was to determine if first-degree relatives of patients with Crohn’s disease have a greater increase in intestinal permeability after acetylsalicylic acid administration than healthy controls. Our secondary objective was to determine if the permeability defect was generalized by examining gastroduodenal permeability before and after provocation with acetylsalicylic acid. Because Crohn’s disease predominantly affects the small and large intestine and uncommonly the stomach, a permeability defect might be limited to the intestines. We have previously developed and validated a sucrose permeability test to assess gastroduodenal permeability.15,16 Gastroduodenal permeability has not been tested previously in patients with Crohn’s disease or their relatives.
Materials and Methods Study Subjects Thirty patients with Crohn’s disease and 37 first-degree relatives of patients with Crohn’s disease were recruited through the outpatient gastroenterology service at the University of Calgary. All patients were in clinical remission (Crohn’s
/ 5e0d$$0021
04-18-96 16:15:15
gasas
Disease Activity Index of õ150).17 None of the relatives had gastrointestinal symptoms. The control group consisted of 26 healthy individuals with no gastrointestinal symptoms. Individuals were excluded if they regularly used acetylsalicylic acid or nonsteroidal anti-inflammatory drugs (more than once a week); had a history of intolerance, allergy, or other contraindication to acetylsalicylic acid; or had a history of diabetes mellitus, renal insufficiency, or peptic ulcer disease. The study protocol was approved by the University of Calgary Ethics Committee.
Study Protocol Before testing, all participants completed a questionnaire about past and current medical problems, gastrointestinal symptoms, current medications, and tobacco and alcohol use. Subjects were instructed not to consume alcohol or nonsteroidal anti-inflammatory drugs for at least 24 hours before the permeability tests. For baseline permeability tests, a test solution was consumed at bedtime after at least a 3-hour fast. The test solution contained 5 g lactulose (Technilab, Montreal, Quebec, Canada), 2 g mannitol (BDH Chemicals Inc., Toronto, Ontario, Canada), and 100 g sucrose in 350 mL of water. Subjects voided before drinking the test solution and then collected all urine, including the first morning void, in a preweighed container containing 5 mL of 10% thymol in methanol. All subjects underwent two permeability tests, a baseline test and a post–acetylsalicylic acid test. Five patients with Crohn’s disease with an increased baseline lactulose-mannitol ratio were excluded from undergoing a post–acetylsalicylic acid test, and their results are excluded from the analysis. Post–acetylsalicylic acid permeability tests were performed after two 1.3-g doses of acetylsalicylic acid (Apotex, Toronto, Ontario, Canada), the first at noon and the second 1 hour before drinking the test solution. The post–acetylsalicylic acid test was performed within 10 days of the baseline test, and urine was collected during a similar period as in the baseline test.
Analytical Methods Urine samples (10 mL) were deionized by adding 1 g of a 1:1.5 (wt/wt) mixture of Amberlite IR-120 and IRA-400 resin (BDH Chemicals Inc.). After centrifugation at 3000 rpm for 10 minutes, the supernatant was filtered through a 40mm/L Millipore filter (Millipore, Bedford, MA). Samples were separated on a Dionex MA-1 anion exchange column by highperformance liquid chromatography (Dionex, Oakville, Ontario, Canada) at room temperature using 500 mmol/L NaOH as the isocratic mobile phase. Peak identification was accomplished with the use of authentic standards and detected using pulsed amperometric electrochemical detection on a gold electrode. Samples were diluted as necessary after addition of cellobiose as an internal standard. Quantitation was performed using known standards at multiple concentrations with linear interpolation between concentrations. All samples were diluted
WBS-Gastro
May 1996
INTESTINAL PERMEABILITY IN CROHN’S DISEASE 1397
so that concentrations of interest fell within the range of the standards. Two permeability measures were calculated from the urine sugar concentrations: the lactulose-mannitol ratio as a measure of intestinal permeability and total sucrose excretion as a measure of gastroduodenal permeability. The lactulose-mannitol ratio is expressed as the ratio of the fractional excretion of lactulose to the fractional excretion of mannitol. As a ratio of two different permeability probes, it has the advantage of controlling for effects of premucosal and postmucosal factors, such as gastric emptying and intestinal transit time, that can affect permeability measures using only single probes.13 Although there are some differences in the rates of urinary excretion of different probes, most of the difference is observed in the early part of the collection period; by 7 hours, the minimum collection period in the present study, this difference is minimal.18 Total sucrose excretion is expressed as total milligrams of sucrose excreted in the urine. Data for each group are expressed as the mean and SEM.
Statistical Analysis Initial sample size estimates were based on the results of Pironi et al.14 The mean intestinal permeability for the control group increased by about 10%, and that of the relatives increased by 180%. Using a conservative estimate of an expected 20% increase in healthy controls and a 90% increase in relatives, a sample size of 32 subjects in each group was required to allow the detection of this difference with a power of 90% and an a of 0.05. Data analysis was performed using SPSS for Windows (SPSS Inc., Chicago, IL). All outcomes were first examined with descriptive statistics and box plots. All statistical tests were two sided using an a of 0.05. Hypothesis testing was not performed to compare the patient group with the other groups because the patient group was specially selected (i.e., all in remission and those with abnormal baseline lactulose-mannitol ratio excluded) and was therefore not considered to be representative of the general Crohn’s disease population. Descriptive statistics and graphical representation of the patient group are provided only to allow comparison. The primary outcome of interest was whether there was a difference in the change in intestinal permeability after acetylsalicylic acid between the group of relatives and the control group. This was examined in two ways: as the increase above the baseline lactulose-mannitol ratio after aspirin and as the percent change in the lactulose-mannitol ratio after aspirin. Both methods were used because it is unknown which is the most meaningful measure for assessing changes in permeability: i.e., is a change in a lactulose-mannitol ratio from 0.010 to 0.020 the same as a change from 0.040 to 0.050? They both represent an increase of 0.010, but in one case is an increase of 100%, whereas in the other only an increase of 25%. Differences between the two groups were tested using
/ 5e0d$$0021
04-18-96 16:15:15
gasas
Table 1. Demographic Characteristics of the Subject Groups
Mean age ( yr ) Women/men Current smokers
Patients
Relatives
Controls
32 15/10 4 (16%)
44 19/18 7 (21%)
40 13/13 1 (4%)
analysis of covariance with the baseline lactulose-mannitol ratio as a covariate. Using the data for the control group, two upper reference limits for the increase and the percent change in the lactulosemannitol ratio after acetylsalicylic acid were calculated to define hyperresponders. A standard upper reference limit was calculated as the mean / 2SD, and a conservative upper reference limit was calculated as the mean / 3SD. Each subject was classified as being above or below these upper limits. The proportion and its 95% confidence interval (95% CI) of relatives with results above these reference limits (hyperresponders) were then calculated. Secondary analysis examined gastroduodenal permeability using baseline and post–acetylsalicylic acid total sucrose excretion. This analysis was performed using the same methods described above. The difference between the post–acetylsalicylic acid change in sucrose permeability in the group of relatives and the control group was assessed by analysis of covariance using the baseline total sucrose excretion as a covariate. The concordance of responsiveness to acetylsalicylic acid among family members was assessed. Each pair of a patient with Crohn’s disease and the relative participating in the study was categorized as being either concordant or discordant based on having an exaggerated increase in intestinal permeability in response to acetylsalicylic acid, defined by a percent increase of greater than the mean / 2SDs of the controls. A post hoc decision was made to assess concordance, and only a limited number of patient-relative pairs were available for analysis; therefore, these particular findings should be interpreted cautiously.
Results Table 1 shows the characteristics of the three subject groups. Mean baseline lactulose-mannitol ratios were 0.0177 (SEM, 0.0010) for the relatives and 0.0174 (SEM, 0.0011) for controls. Results for subjects with Crohn’s disease are not given because subjects with a high baseline permeability were excluded from analysis. The difference between the mean lactulose-mannitol ratios of the group of relatives and control group was not significantly different (P Å 0.9). Two controls and 3 relatives had values greater than the standard upper reference limit of 0.0281 (mean / 2SD). Figure 1 shows individual paired data for baseline and post–acetylsalicylic acid lactulose-mannitol ratios for relWBS-Gastro
1398 HILSDEN ET AL.
GASTROENTEROLOGY Vol. 110, No. 5
Figure 1. Paired results for baseline and post–acetylsalicylic acid (ASA) lactulose-mannitol ratios for relatives of patients with Crohn’s disease and for healthy controls; no significant difference in baseline ratio was observed for the two groups. After acetylsalicylic acid treatment, the controls had a 57% increase, whereas the relatives had a 110% increase (P Å 0.001). In contrast to controls, some relatives were much more responsive to the effects of aspirin than others.
atives and controls. Individuals in the control group had a slight increase in their lactulose-mannitol permeability in response to acetylsalicylic acid, but several relatives of patients with Crohn’s disease had increases of a much greater degree (hyperresponders). Figure 2 shows the increase above the baseline lactulose-mannitol ratio after ingestion of acetylsalicylic acid. Mean increases in the lactulose-mannitol ratios were 0.0136 (SEM, 0.0021), 0.0184 (0.0026), and 0.0093 (0.0013) for patients, relatives, and controls, respectively. There was a statistically significant difference in the mean change in lactulose-mannitol ratio between relatives and controls (95% CI for the difference, 0.003–0.015; P Å 0.003). In contrast to the baseline test results, which were similar between the relatives and controls, the two groups differed dramatically after exposure to acetylsalicylic acid. The control group responded relatively uniformly to the acetylsalicylic acid. In contrast, the relatives had a much wider distribution; some were much more responsive to the effects of aspirin than others. Also, the pattern observed in the relatives and patients was quite similar and was different from that observed in the controls. The proportion of relatives who showed hyperresponsiveness to aspirin was estimated after calculating the standard (mean / 2SD) and conservative (mean / 3SD) upper reference limits for the post–acetylsalicylic acid increase in the lactulose-mannitol ratio. Twenty-seven / 5e0d$$0021
04-18-96 16:15:15
gasas
percent (95% CI, 14%–44%) of relatives had post–acetylsalicylic acid increases greater than the standard upper reference limit of 0.0230; 14% (95% CI, 5%–29%) of relatives had increases greater than the conservative upper reference limit of 0.030. Figure 3 shows the percent change in the lactulosemannitol ratio after acetylsalicylic acid. Post–acetylsalicylic acid changes in intestinal permeability, measured as the percent change, gave similar results when the change in permeability was measured as the absolute increase. Mean percent changes in the lactulose-mannitol ratios were 133% (SEM, 25%), 110% (13%), and 57% (9%) for patients, relatives, and controls, respectively. There was also a statistically significant difference in the mean percent change between relatives and controls (95% CI for the difference, 21%–85%; P Å 0.001). The standard upper reference limit for the post–acetylsalicylic acid percent change in the lactulose-mannitol ratio was 143%, and the conservative upper reference limit was 186%. Thirty-five percent of relatives (95% CI, 20%–53%) were classified as hyperresponders using the standard upper reference limit; 19% of relatives (95% CI, 8%–35%) were classified as hyperresponders using the conservative upper reference limit. There were more current smokers in the group of
Figure 2. One-way scatterplots of the increase in the lactulose-mannitol ratio after acetylsalicylic acid treatment in the three subject groups. The dotted line represents the standard upper reference limit (mean / 2SD of the controls), and the dashed line represents the conservative upper reference limit (mean / 3SD). The range of values is much wider for the relatives than for controls. Ten relatives had post–acetylsalicylic acid increases greater than the standard reference limit, and 5 had increases greater than the conservative reference limit. There is a significant difference in the mean change between the relatives and controls (P Å 0.003).
WBS-Gastro
May 1996
INTESTINAL PERMEABILITY IN CROHN’S DISEASE 1399
Figure 3. One-way scatterplots of percent change in the lactulosemannitol ratio after acetylsalicylic acid treatment for the three subject groups. The dotted and dashed lines represent standard and conservative upper reference limits, respectively. Thirteen relatives had post–acetylsalicylic acid percent increases greater than the standard reference limit, and 7 had increases greater than the conservative reference limit. The range of results is much wider for the relatives than for controls. There is a significant difference in the mean percent change between the relatives and controls (P Å 0.001).
relatives than in the control group. However, smoking status did not seem to be related to changes in permeability after acetylsalicylic acid treatment. When current smokers were excluded from the analysis, there was still a statistically significant difference between relatives and controls. Thus, for nonsmoking relatives the mean post– acetylsalicylic acid change in lactulose-mannitol ratio was 0.0180 compared with 0.0095 for nonsmoking controls (P Å 0.006). A limited number of patient-relative pairs participated in the study (8 patients with 12 relatives). For these 12 pairs, concordance of permeability changes in response to acetylsalicylic acid was assessed using the percent change data. Five patients had hyperresponsive changes in permeability after acetylsalicylic acid treatment, and 7 of their 8 relatives also had abnormal increases. Three patients did not have hyperresponsive post–acetylsalicylic acid changes, and all 4 of their relatives also had normal changes. Examples of concordance patterns for three families are shown in Figure 4. Sucrose excretion was used to assess gastroduodenal permeability. Mean baseline sucrose excretion was 105 mg (SEM, 11 mg) for the relatives and 88 mg (SEM, 10 mg) for controls. Although the relatives had a higher baseline sucrose excretion than controls, this difference was not statistically significant (P Å 0.26). / 5e0d$$0021
04-18-96 16:15:15
gasas
The effect of acetylsalicylic acid on sucrose excretion was assessed in the same way as the lactulose-mannitol ratio data. In contrast to the results for intestinal permeability, the distributions of the results for the relatives and controls are much more similar, suggesting a similar response to acetylsalicylic acid. Mean increases in total sucrose excretion were 169 mg (SEM, 26 mg), 234 mg (40 mg), and 127 mg (22 mg) for patients, relatives, and controls, respectively. The mean increase for the relatives is influenced by a single markedly high result. Mean percent changes in total sucrose excretion were 290% (SEM, 46%), 259% (40%), and 198% (39%) for patients, relatives, and controls, respectively (Figure 5). Although the results for post–acetylsalicylic acid sucrose excretion are higher for the relatives, when considering baseline sucrose excretion as a covariate, there was no significant difference between relatives and controls in the post– acetylsalicylic acid increase or percent change in sucrose excretion. When relatives classified as hyperresponders or normal responders, based on the post–acetylsalicylic acid increase in lactulose-mannitol ratio, were examined, there was no difference in the baseline lactulose-mannitol ratio between the two groups (95% CI for the difference, 00.007 to 0.03). The baseline total sucrose excretion was 90 mg for the normal responders and 145 mg for the hyperresponders. However, this difference did not reach statistical significance (95% CI, 0130 to 19). The change in sucrose excretion also did not differ between the two groups (P Å 0.15) when baseline sucrose excretion was considered as a covariate.
Figure 4. Three pedigrees showing the concordance pattern of baseline and post–acetylsalicylic acid lactulose-mannitol permeability results among family members with and without Crohn’s disease. Although these are small pedigrees, the degree of concordance is quite striking. 䊊, Normal female; 䊐, normal male; 䊊 / , Crohn’s disease; , abnormal baseline lactulose-mannitol ratio; , abnormal increase post–acetylsalicylic acid.
WBS-Gastro
1400 HILSDEN ET AL.
GASTROENTEROLOGY Vol. 110, No. 5
Figure 5. One-way scatterplots of percent change in total sucrose excretion after acetylsalicylic acid treatment for the three subject groups. The distribution of results is similar between the groups. There was no significant difference between relatives and controls.
Discussion Extensive evidence shows that Crohn’s disease has a familial predisposition. This evidence includes striking ethnic differences in the frequency of Crohn’s disease, familial aggregation of Crohn’s disease, and higher concordance rate for monozygotic than for dizygotic twins. Not only has familiality been linked to the risk for Crohn’s disease but also to disease characteristics, such as age at onset, location of disease, and perforating disease.19,20 The mechanism of action of a genetic predisposition remains unclear. Previous studies have examined the complement system, immunoglobulins, mucin, and the fecal flora. Because increased intestinal permeability has been found in patients with Crohn’s disease,21 – 23 a genetically determined permeability defect is an attractive explanation. Possibly, impairment of the barrier function of the intestine could lead to increased exposure of the mucosal immune system to luminal antigens, resulting in an inflammatory reaction.24 Hollander et al.4 showed increased intestinal permeability in healthy relatives compared with controls, but most subsequent studies have failed to confirm this. Standard lactulose-mannitol permeability tests may not be sufficiently sensitive to discriminate among individuals with and without a permeability defect. Comparing differences in standard permeability test results between relatives and controls may be prone to error because only a subset of relatives may have a permeability defect. The modest increase in permeability induced by / 5e0d$$0021
04-18-96 16:15:15
gasas
a defect may be lost when only the mean value for a group is examined. We previously suggested that approximately 10% of relatives have abnormal intestinal permeability.9 Furthermore, our study suggests that many patients with inactive Crohn’s disease have normal lactulose-mannitol permeability when it is tested in the standard way. These individuals would have a permeability defect if it exists. Although it is unknown what these patients’ intestinal permeability would be if they were not affected by Crohn’s disease, clearly it would not differ greatly from that of normal subjects. Therefore, standard lactulose-mannitol permeability tests appear to be an inadequate tool for evaluating relatives to detect a familial permeability defect. In this study, we show that healthy first-degree relatives of patients with Crohn’s disease have an exaggerated increase in intestinal permeability after provocation with acetylsalicylic acid. A provocation test provides enhanced discrimination if it increases the difference between normals and abnormals. All subjects had increases in intestinal permeability after acetylsalicylic acid administration, but relatives (and patients) had greater increases than controls. One would expect that only a proportion of relatives would have a permeability defect compared with none of the controls. Our findings support this because the relatives had a much greater range of values for the post–acetylsalicylic acid change in permeability than the controls (Figures 2 and 3). What proportion of relatives may have a permeability defect? Determining this depends on what is considered normal. Using the data for the control group, we have created upper reference limits for changes in the lactulose-mannitol ratio after acetylsalicylic acid administration. Because standard reference limits calculated as the mean / 2SD using sample sizes õ100 can be quite imprecise, with 95% CIs as wide as 1SD of the data, we calculated a conservative upper reference limit, which is the mean / 3SD.25 Values above this would be clearly abnormal. We found that 14% and 19% of the relatives had values greater than the conservative upper reference limit for post–acetylsalicylic acid increase and percent change in lactulose-mannitol ratio, respectively. We calculated both the change and percent change because it is unclear which is the more meaningful measure of changes in permeability. As discussed above, an increase in the lactulose-mannitol ratio of 0.01 may not mean the same thing in an individual with a baseline ratio of 0.01 as it does in an individual with a baseline ratio of 0.04. Overall, both measures gave similar results, which strengthens the validity of the findings. Confounding factors do not seem to be responsible for WBS-Gastro
May 1996
INTESTINAL PERMEABILITY IN CROHN’S DISEASE 1401
the study’s findings. As a group, the relatives are slightly older than the controls, but increasing age is not associated with increases in the lactulose-mannitol ratio.26 There was a higher proportion of smokers among the relatives, but the prevalence of permeability defects was equal among smokers and nonsmokers, and the relatives still had a statistically significant greater post–acetylsalicylic acid increase in lactulose-mannitol ratio when smokers were excluded. Finally, the effects of any environmental factors would be present in both the baseline and post–acetylsalicylic acid permeability tests and therefore should have been canceled when assessing permeability changes. Because of the specific selection of subjects with Crohn’s disease, we provided their results only to allow qualitative comparisons; however, some interesting observations can be made. As noted already, most patients with inactive Crohn’s disease have normal intestinal permeability. However, after provocation with acetylsalicylic acid, not all patients have an exaggerated increase in the lactulose-mannitol ratio. In fact, the distribution of the patients’ results is quite similar to that of the relatives (Figures 2 and 3), possibly suggesting that not all patients with Crohn’s disease have a familial permeability defect. This impression is strengthened by noting the concordance of normal or abnormal increases in permeability within families (Figure 5). Although only a limited number of patient-relative pairs was available, there was a striking degree of concordance among relatives. This has also been found in studies of other possible genetic markers.27 Recently, Peeters et al.28 challenged the concept that a permeability defect in relatives of patients with Crohn’s disease is genetic in origin. Using standard lactulosemannitol permeability tests in 17 complete Crohn’s families, they were unable to show a specific familial pattern to abnormal intestinal permeabilities. They also found that 4 of 13 spouses had abnormal permeabilities. They concluded that increased permeability could be better explained by an environmental factor. Our study was not designed to specifically address whether a permeability defect resulted from genetic or environmental factors. However, none of the 3 spouses participating in the study as controls had abnormal post– acetylsalicylic acid permeability changes. Furthermore, as mentioned above, there was a high concordance among family members in response patterns. This could result from a shared environmental factor; however, many of these individuals lived in different cities, making this less likely. It is possible that a permeability defect results from / 5e0d$$0021
04-18-96 16:15:15
gasas
an early life exposure that is not necessarily present when the subjects are studied. Previous studies have suggested that childhood factors, such as passive smoking and early weaning, may be important in the etiology of Crohn’s disease.29,30 Also, cow’s milk feeding has been associated with altered intestinal permeability in newborn infants. Weaver et al. found that compared with newborn infants who were breast-fed, those who received cow’s milk feeds had persistently higher intestinal permeability throughout the first week of life.31 However, it has not been shown that a permeability defect persists through later infancy or adulthood. Alternatively, those with a familial permeability defect may be more susceptible to luminal antigens or toxins that may lead to Crohn’s disease. We also examined gastric permeability using a sucrose permeability test. Sucrose permeability has not been examined previously in patients with Crohn’s disease or their relatives. Again, baseline gastroduodenal permeability seems to be similar in all groups. Acetylsalicylic acid caused a substantial increase in gastroduodenal permeability, but, in contrast to intestinal permeability, patients with Crohn’s disease and their relatives did not have an exaggerated increase compared with controls. This suggests that any permeability defect is restricted to the intestine. Alternatively, the dose of aspirin may have been too large, resulting in substantial increases regardless of an underlying defect, limiting the power of this study to detect a difference. However, this finding is consistent with the unusual occurrence of gastric or duodenal Crohn’s disease. We now have evidence that the first-degree relatives of patients with Crohn’s disease can be characterized by the following intestinal permeability patterns: (1) normal permeability, (2) abnormal baseline permeability, (3) exaggerated increase in permeability in response to acetylsalicylic acid, and (4) both an abnormal baseline permeability and an exaggerated increase in response to acetylsalicylic acid. One or more of these permeability patterns may characterize individuals who are predisposed to develop Crohn’s disease. However, there is no evidence that individuals with a permeability defect are at increased risk for developing Crohn’s disease. It might be concluded from this study, in which a large number of adults had a permeability defect but no evidence of Crohn’s disease, that it does not. However, cross-sectional studies will never be able to resolve this question. In studies such as this, subjects have already declared whether they will develop Crohn’s disease or not. A longitudinal study in which individuals are tested and then followed up for a prolonged period for the development of Crohn’s disease would be required. Because familiality WBS-Gastro
1402 HILSDEN ET AL.
GASTROENTEROLOGY Vol. 110, No. 5
has also been associated with young age at onset, such a study would have to recruit subjects as young children. Clearly, such a study is unlikely. This study suggests that there may be genetic heterogeneity among patients with Crohn’s disease: some patients are hyperresponders, some are not. This is consistent with other studies suggesting genetic heterogeneity in Crohn’s disease.27,32 Crohn’s disease is also characterized by clinical heterogeneity. A patient’s disease can be characterized by its anatomic location, anatomic extent, and disease behavior.33 For example, Greenstein et al. have suggested that perforating and nonperforating types are distinct patterns of disease.34 Possibly clinical disease patterns are associated with a particular permeability pattern. Crohn’s disease is a genetically and clinically complex disease. We should not be surprised that finding a genetic marker of the disease is difficult or that potential markers are not found in all patients or in all relatives. There is evidence that a permeability defect is present in some patients with Crohn’s disease and some of their relatives. Whether a familial permeability defect predisposes to Crohn’s disease is likely to remain unknown. Future studies should be directed at addressing unresolved issues: whether a permeability defect results from genetics or the environment and whether permeability defects are associated with clinical phenotypes.
9.
10.
11.
12.
13. 14.
15.
16.
17.
18.
References 1. Calkins BM, Mendeloff AI. The epidemiology of idiopathic inflammatory bowel disease. In: Kirsher JB, Shorter RS, eds. Inflammatory bowel disease. 4th ed. Baltimore, MD: Williams & Wilkins, 1995:31–68. 2. Tysk C, Lindberg E, Ja¨rnerot G, Flode´rus-Myrhed B. Ulcerative colitis and Crohn’s disease in an unselected population of monozygotic and dizygotic twins. A study of heritability and the influence of smoking. Gut 1988;29:990–996. 3. Thompson NP, Driscoll R, Pounder RE, Wakefield AJ. Genetics vs. environment in inflammatory bowel disease: concordance rates in 130 twin pairs (abstr). Gastroenterology 1995;108: A928. 4. Hollander D, Vadheim CM, Brettholz E, Petersen GM, Delahunty, T, Rotter JI. Increased intestinal permeability in patients with Crohn’s disease and their relatives. Ann Intern Med 1986;105: 883–885. 5. Katz KD, Hollander D, Vadheim CM, McElree C, Delahunty T, Dadufalza VD, Krugliak P, Rotter JI. Intestinal permeability in patients with Crohn’s disease and their healthy relatives. Gastroenterology 1989;97:927–931. 6. Teahon K, Smethurst P, Levi AJ, Menzies IS, Bjarnason I. Intestinal permeability in patients with Crohn’s disease and their first degree relatives. Gut 1992;33:320–323. 7. Ainsworth M, Erikson J, Rasmussen JW, Schaffalitzky de Muckadell OB. 51Cr-Labelled ethylenediaminetetraacetic acid in patients with Crohn’s disease and their healthy relatives. Scand J Gastroenterol 1985;24:993–998. 8. Ruttenberg D, Young GO, Wright JP, Isaacs S. Peg-400 excretion
/ 5e0d$$0021
04-18-96 16:15:15
gasas
19.
20.
21.
22.
23.
24.
25. 26.
27.
in patients with Crohn’s disease, their first-degree relatives, and healthy volunteers. Dig Dis Sci 1992;37:705–708. May GR, Sutherland LR, Meddings JB. Is small intestinal permeability really increased in relatives of patients with Crohn’s disease? Gastroenterology 1993;104:1627–1632. Munkholm P, Langholz E, Hollander D, Thornberg K, Orholm M, Katz KD, Binder V. Intestinal permeability in patients with Crohn’s disease and ulcerative colitis and their first degree relatives. Gut 1994;35:68–72. Howden CW, Gillanders I, Morris AJ, Duncan A, Danesh B, Russell RI. Intestinal permeability in patients with Crohn’s disease and their first-degree relatives. Am J Gastroenterol 1994;89:1175– 1176. Peeters M, Ghoos Y, Geypens B, Hiele M, Rutgeerts P. Lactulose/mannitol (L/M) permeability index is a suitable subclinical marker to identify subjects at risk in families of Crohn’s patients (abstr). Gastroenterology 1994;106:A750. Bjarnason I, MacPherson A, Hollander D. Intestinal permeability: an overview. Gastroenterology 1995;108:1566–1581. Pironi L, Miglioli M, Ruggeri E, Dallasta MA, Ornigotti L, Valpiani D, Barbara L. Effect of non-steroidal anti-inflammatory drugs (NSAID) on intestinal permeability of first degree relatives of patients with Crohn’s disease (abstr). Gastroenterology 1992; A679. Meddings JB, Sutherland LR, Byles NI, Wallace JL. Sucrose: a novel permeability marker for gastroduodenal disease. Gastroenterology 1993;104:1619–1626. Sutherland LR, Verhoef M, Wallace JL, Van Rosendaal G, Crutcher R, Meddings JB. A simple, non-invasive marker of gastric damage: sucrose permeability. Lancet 1994;343:998–1000. Best WR, Bechtel JM, Singleton JW. Rederived values of the eight coefficients of the Crohn’s disease activity index (CDAI). Gastroenterology 1979;77:843–846. Bjarnason I, Maxton D, Reynolds AP, Catt S, Peters TJ, Menzies IS. Comparison of four markers of intestinal permeability in control subjects and patients with coeliac disease. Scand J Gastroenterol 1994;29:630–639. Tokayer AZ, Reydel B, Bayless TM. Possible role of heredity in site and transmural aggressiveness of Crohn’s disease (abstr). Gastroenterology 1992;102:A705. Gower-Rousseau C, Plegat S, Evrard JP, Grandbastien B, Hugot JP, Dupas JL, Gendre JP, Modigliani R, Be´laı¨che J, Hostein J, Cortot A, Colombel JF. Familiality in clinical characteristics of Crohn’s disease in 69 French families (abstr). Gastroenterology 1995;108:A826. Ukabam SO, Clamp JR, Cooper BT. Abnormal small intestinal permeability to sugars in patients with Crohn’s disease of the terminal ileum and colon. Digestion 1983;27:70–74. Olaison G, Leandersson P, Sjo¨dahl R, Tagesson C. Intestinal permeability to polyethylene glycol 600 in Crohn’s disease. Preoperative determination in a defined segment of small intestine. Gut 1988;29:196–199. Lochs H, Wyatt J, Reinisch W, Gasche´ C, Maier T, Moser G, Gangl A. Clinical relevance of intestinal permeability in active Crohn’s disease (abstr). Gastroenterology 1994;106:A722. Hollander D. Permeability in Crohn’s disease: altered barrier functions in healthy relatives? Gastroenterology 1993;104:1848– 1851. Altman DG. Practical statistics for medical research. London: Chapman & Hall, 1991. Saltzman JR, Knowdley KV, Perrone G, Russell RM. Changes in small-intestine permeability with aging. J Am Geriatr Soc 1995; 43:160–164. Elmgreen J, Both H, Binder V. Familial occurrence of complement dysfunction in Crohn’s disease: correlation with intestinal symp-
WBS-Gastro
May 1996
28.
29.
30.
31.
32.
INTESTINAL PERMEABILITY IN CROHN’S DISEASE 1403
toms and hypercatabolism of complement. Gut 1985;26:151– 157. Peeters M, Geypens B, Ghoos Y, Nevens H, D’Haens G, Hiele M, Rutgeerts P. The pattern of increased small intestinal permeability abnormalities in families with Crohn’s disease: genetic or non-genetic (abstr)? Gastroenterology 1995;108:A892. Lashner BA, Shaheen NJ, Hanauer SB, Kirschner BS. Passive smoking is associated with an increased risk of developing inflammatory bowel disease in children. Am J Gastroenterol 1993; 88:356–359. Bergstrand O, Hellers G. Breast-feeding during infancy in patients who later develop Crohn’s disease. Scand J Gastroenterol 1983; 18:903–906. Weaver LT, Laker MF, Nelson R, Lucas A. Milk feeding and changes in intestinal permeability and morphology in the newborn. J Pediatr Gastroenterol Nutr 1987;6:351–358. Polito J, Tokayer A, Childs B, Harris M, Bayless T. Crohn’s dis-
/ 5e0d$$0021
04-18-96 16:15:15
gasas
ease: evidence for genetic heterogeneity and ’’anticipation‘‘ (abstr). Gastroenterology 1995;108:A895. 33. Sachar DB, Andrews H, Farmer RG, Pallone F, Pen ˜a AS, Prantera C, Rutgeerts P. Proposed classification of patient subgroups in Crohn’s disease. Working team report 4. Gastroenterol Int 1992; 3:141–154. 34. Greenstein AJ, Lachman P, Sachar DB, Springhorn J, Heimann T, Janowitz HD, Aufses AHJ. Perforating and non-perforating indications for repeated operations in Crohn’s disease: evidence for two clinical forms. Gut 1988;29:588–592. Received August 25, 1995. Accepted January 4, 1996. Address requests for reprints to: Lloyd Sutherland, M.D., Department of Community Health Sciences, 3330 Hospital Drive Northwest, Calgary, Alberta, Canada T2N 4N1. Fax: (403) 270-7307. Supported by a clinical fellowship from the Alberta Heritage Foundation for Medical Research (to R.J.H.).
WBS-Gastro