STUDY GROUPS

The study groups comprised 39 Crohn’s disease patients with a median (range) disease duration of 20 (1–34) years, 34 first degree relatives, 22 spouses, and 29 healthy volunteers (table 1). Five patients, two relatives, and one of the spouses did not participate in the ASA ingestion test. Table 2 gives data on the patients with Crohn’s disease. Diagnosis was based on standard radiographic and/or endoscopic findings and the Morson morphological criteria.^22 23 Disease activity was assessed by Best’s modification of his classical Crohn’s disease activity index (CDAI).24 The 39 Crohn’s disease patients had a total of 123 first degree relatives above the age of 18. Fifty one of these were available for the study (many relatives lived far away, some patients did not want to burden their relatives with the study, and some had little social contact with their relatives). Eight were excluded because of concurrent disease, three because of pregnancy, and six chose not to participate. The 34 first degree relatives included in the study (18 parents, nine siblings, and seven children from 21 families) had no signs or symptoms of Crohn’s disease. Eight of the relatives studied were living with the patient at both the time of diagnosis and the time of the test; 14 were living with the patient at the time of diagnosis but not at the time of the test; 12 were not living with the patient either at the time of diagnosis or at the time of the test. Eleven of the 12 relatives in the latter subgroup participated in both the permeability tests (baseline and response to ASA). They are hereafter called “solely relatives”; the period apart from the related patient ranged from 6 to 25 years. The 22 spouses had been living with the patients for a median (range) of 17 (1–41) years at the time of the test. Ten of the spouses were already living with the patient at the time of diagnosis, nine of which participated in both the permeability tests. These nine are designated “spouses at diagnosis”, and had shared their environment with their patient partner for a median (range) of 3 (0.5–30) years at the time of diagnosis. The apparently healthy volunteers showed no signs or symptoms of intestinal disorder, were not on medication, and had normal dietary habits and normal renal function; the female subjects were not pregnant.

As a separate group, nine of 14 previously identified monozygotic twin pairs8 were studied. Five pairs (aged 44–59; eight men, two women) were concordant for Crohn’s disease, and four pairs (aged 43–72; four men, four women) were discordant for the disease. The 14 twins with Crohn’s disease had a disease duration of 22 (9–30) years. At the time of the investigation, the median CDAI was 80 (0–200). Three of the twin patients had a CDAI above 150, and the disease was inactive in the other 11.

TEST PROCEDURE

Lactulose and mannitol were given orally as Permagnost (lactulose 10.0 g, mannitol 5.0 g, glucose 9.0 g, NaCl 0.1 g, aqua ad 100 ml, osmolarity 1304 mosm/l; Laevosan, Linz, Austria). The permeability test was performed twice, once under baseline conditions and again 5–10 days later after intake of ASA. The subjects were not allowed to take non-steroidal anti-inflammatory drugs (NSAIDs) less than one week before the test, nor to drink alcohol less than three days before. Permagnost (100 ml) was taken together with two glasses of tap water (300 ml) after an overnight fast, and all urine was then collected for the next five hours. During the first two hours of sampling, no oral intake was allowed, but water, coffee or tea (without milk or sugar) was permitted during the last three hours. The bladder was emptied an extra time at five hours. Five to ten days later, the ASA provocation was carried out: 1.25 g ASA in the form of two and a half tablets of Magnecyl (Pharmacia & Upjohn, Stockholm, Sweden) was given twice orally 8–10 hours and one hour before the oral administration of 100 ml Permagnost together with two glasses of water. Urine collection and oral intake were as above. To avoid bacterial growth, the urine was collected in plastic reservoirs containing 0.5 ml 20% chlorhexidine gluconate solution. Urine volume and creatinine were measured and 10 ml aliquots were stored at −20°C for subsequent analysis.

ANALYTICAL METHODS

Lactulose and mannitol were measured essentially as described by Kynaston et al,25except that cellobiose was used as internal standard instead of melibiose, because a compound that was eluted at the same time as melibiose was present in the urine samples. The amount of urine analysed was adjusted for all samples so that the measured concentrations were within the concentration range of the standard solutions. Typically, 0.25 ml urine and 0.75 ml water were mixed with 1.0 ml internal standard solution (cellobiose 0.75 mmol/l). The samples were desalted by adding 0.5 g of an ion-exchange mixture (Amberlite IR120 H and Amberlite IRA420 Cl in mass proportions 1:1.5). Thereafter, each sample was centrifuged at 1000g for 10 minutes, and the supernatant was filtered through a 0.45 μm disposable filter. Aliquots (25 μl) were then injected on to a 250 × 40 mm Carbopac PA-1 anion exchange column coupled to a quaternary gradient HPLC system with a pulsed amperometric detector (Dionex Corporation, Sunnyvale, California, USA). The carbohydrates were separated isocratically at a flow rate of 1.0 ml/min at 22°C, with an eluent consisting of 90 mmol/l NaOH and 0.375 mmol/l zinc acetate. Like Kynaston et al,25 we obtained a clear separation between lactose and lactulose. Five standard solutions containing lactulose (0.025–1.0 mmol/l) and mannitol (0.5–20 mmol/l) were used for quantification by internal standardisation with peak:area ratios. The coefficients of variation were 4.2 and 4.9% for mannitol and lactulose respectively.

CALCULATIONS AND STATISTICS

The lactulose:mannitol (L/M) ratio in the urinary output was calculated and used as a permeability index to minimise the influence of differences in gastrointestinal motility, absorptive capacity, urinary volume, and renal function.^3 4 The difference in L/M ratio between baseline and after ASA provocation was calculated in each individual as a measure of the permeability response to ASA. The 95th percentile in the controls was defined as the upper limit of normality with regard to baseline permeability and response to ASA. Data showed a skewed distribution and are presented as median (interquartile range), and scattergrams. The Kruskal-Wallis and Mann-Whitney tests were used for comparison between groups, and the paired sign test for within-group comparisons. The χ² and Fisher’s exact tests were applied to compare proportions. Analysis of covariance was used to assess the influence of various patient factors on permeability. Differences with p<0.05 were considered to be significant.

ETHICS

The study protocol was approved by the ethics committee of the Faculty of Health Sciences, Linköping University and was conducted according to the Declaration of Helsinki.

BASELINE PERMEABILITY

Table 3 gives the baseline L/M ratios. The 95th percentile for controls (0.0195) was used as the upper limit for normal permeability. Baseline permeability was increased in the patient group as compared with the controls (table 3; fig 1), with a baseline permeability above normal in 14/39 (36%) of the patients. Permeability was not related to the location or activity of the disease (analysis of covariance), but higher permeability was seen in patients on medication than in untreated patients (median 0.017 and 0.012 respectively; p<0.05). The baseline permeability of the relatives was equal to that of the controls (table 3; fig 1), although in a subgroup of the relatives (6/34; 18%) the permeability was increased. The spouses formed an intermediate group with increased permeability compared with controls (table 3; fig 1), and a raised baseline permeability in 5/22 (23%). In the group denoted “solely relatives”, a raised baseline permeability was found in 1/11 (9%), whereas the corresponding value for “spouses at diagnosis” was 3/9 (33%; table 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Paired scattergrams of lactulose:mannitol (L/M) ratio in five hour urine samples before (Base) and after oral ingestion of 2 × 1.25 g acetylsalicylic acid (ASA) in the various study groups. The subjects tested were healthy controls (n = 29), Crohn’s disease patients (n = 39), first degree relatives (n = 34), and spouses (n = 22). Solid lines indicate the median L/M ratios in the groups. The upper limit of normality (95th percentile in the control group) at baseline (0.0195) is indicated by the dotted line and the upper limit after ASA (0.032) is indicated by the dashed line. At baseline the median L/M ratio was increased in patients and spouses as compared with controls (p<0.05, Kruskal-Wallis). After intake of ASA the L/M ratio was higher in both patients and relatives than in controls (p<0.05, Kruskal-Wallis), but not in spouses. Note the marked increase in permeability in a large subgroup of the relatives after ingestion of ASA.

PERMEABILITY AFTER INGESTION OF ASA

ASA ingestion increased the L/M ratio significantly in all study groups (p<0.005, paired sign test; table 3), with an increase in the 95th percentile for controls to 0.032. After exposure to ASA, the ratio in the group of relatives was increased to that of the patient group (table 3; fig 1). L/M ratios for spouses did not show a similar rise, and did not differ from the controls after ingestion of ASA (table 3; fig 1).

RESPONSIVENESS TO ASA

The median permeability response to ASA—that is, the increase in L/M ratio from baseline—was 0.009 (0.004–0.011) in the controls, with the 95th percentile being 0.018. There were no differences in the median permeability responses between the various groups, being 0.009, 0.010, and 0.007 in the patients, relatives, and spouses respectively.

An abnormal permeability response to ASA—that is, a rise exceeding that of the 95th percentile for controls—was found in 11/34 (32%) of the patients and 13/32 (41%) of the relatives, whereas the proportion of spouses who were hyperresponders (3/21; 14%) did not differ from that for the controls (table 3). In the group of “solely relatives” the proportion of hyperresponders was 3/11 (27%), whereas no hyperresponders (0/9) were found among “spouses at diagnosis” (table 3).

MONOZYGOTIC TWINS

Baseline L/M ratios in the twin groups were 0.014 (0.010–0.016) and 0.015 (0.009–0.019) for twins with Crohn’s disease and healthy twins respectively. After ingestion of ASA the permeability increased to 0.022 (0.019–0.035) and 0.023 (0.018–0.023) respectively. None of the healthy twins (0/4) had an augmented permeability response to ASA, but an abnormal response was seen in four (28%) of the 14 twins with Crohn’s disease (p = 0.03 compared with controls; Fisher’s exact test).