Technical Field The present invention relates to a process (1) for eva¬ luating disease activity in patients with chronic in¬ flammatory bowel diseases (IBD), termed ulcerative coli¬ tis (UC) and Crohn's colitis (CC), respectively, and (2) for predicting (a) the outcome of medical treatment in severe, active disease, where acute or subacute surgery may be an alternative, in addition to (b) the risk of relapse in symptom-free patients, who wish to stop re¬ lapse preventing medication.

Furthermore, the invention relates to a kit to be used in the process described.

Background Art

The antiinflammatory action of corticosteroids is well established and treatment with pharmacological doses of corticosteroids is the most important medical treatment in a large number of active inflammatory and immune based diseases. Thus corticosteroids are considered the most effective type of drugs for treatment of acute relapsing UC, but long-term treatment is associated with a number of serious side effects.

For long-term treatment of UC and CC in low-active phase it is more convenient, therefore, to use drugs contain- ing 5-amino-salicylic acid (5-ASA) linked to sulphapyri- dine (sulphasalazine (SAZ), Salazopyrine , Pharma¬ cia AB, Sweden), another 'carrier' molecule (the recent¬ ly developed sulphasalazine-analogues) or 5-ASA in a 'slow release' preparation, because free 5-ASA is completely absorbed in the upper small intestine. Thus the well-known adverse effects of corticosteroids are avoided and the risk of relapse markedly reduced.

Since the 1940's the most important 5-ASA containing preparation has been sulphasalazine (SAZ) with the che¬ mical structure:

SAZ

SAZ, which passes the small intestine unabsorbed, is split in the colon by bacterial azo-reductases into its two components. Free 5-ASA acts directly on the colonic mucosa without prior absorption. Within the last few years a number of SAZ-analogous and 'slow-release' 5-ASA preparations have been developed for treatment of pa¬ tients with IBD. In controlled clinical trials it has been attempted to decide whether patients with UC, re¬ fractory to SAZ-treatment, would benefit from treatment with the new sulphor-free salicylates, which may be administered in higher doses than SAZ without side effects. Although SAZ is an effective drug for mainte¬ nance treatment, with the purpose of avoiding a relapse of the disease, a large proportion of patients does not obtain a complete remission and relapses occur frequently during

long-term maintenance treatment with SAZ. In addition, 10-20% of the patients are allergic to SAZ or in other ways intolerant of the drug. This is considered to rely on sulphapyridine, which is easily absorbed when split from the therapeutically active moiety, 5-ASA. The effi¬ cacy of SAZ in preventing a relapse in patients with UC is dose dependent, but the frequency of adverse ef- fects, which may be ascribed to sulphapyridine as men¬ tioned above, increases with the dosage. Although oral administration of SAZ to Datients with CC or mild to mo-

derately active UC is effective the drug is markedly less effective than corticosteroids.

By contrast, topical treatment with SAZ and 5-ASA ap¬ pears to be as effective as topical treatment with glucocorticoids i.e. rectal instillation in patients with non-extensive disease localized to the lower colon or the rectum solely.

Lately, disodium-azodisalicylate (ADS; olsalazine: Dipentum™, Pharmacia AB, Sweden) has been appointed a promising candidate for all treatment of IBD, because the substance consists of two molecules of 5-ASA linked by an azo-bond. This means that each of its two active components acts as a carrier for the other:

Jones, A. Connell, J.H. Baron, F.A. avery Jones, 'Con¬ trolled trial of sulphasalazine in maintenance therapy for ulcerative colitis', the Lancet 1965, 1:185-88, SAZ significantly reduces the relapse rate in patients with UC tolerant to the drug and is considered, therefore, the most important basic treatment for prevention of relapse in UC. However, it is still controversial whether SAZ treatment of symptom-free patients with UC should be continued more than one year and two early studies showed conflicting results. Although several later studies have shown a higher relapse rate in pa¬ tients discontinuing SAZ treatment, the decision as regard duration of treatment is not based on objective measures, but rests solely on 'personal, clinical experience', since no clinical findings or conventional laboratory measurements improve decision making.

Furthermore, it is a serious problem that patients ha¬ ving had treatment for acute relapsing UC and subse¬ quently maintained on SAZ to prevent a relapse often discontinue medication, thus increasing the risk of re- lapse and complications. Oppositely, it is inappro¬ priate to continue treatment which is not Justified for medical reasons.

Description of the Invention It is therefore an object of the invention to provide a process enabling assessment of disease activity in pa¬ tients with active UC and CC and prediction of the out¬ come of medical treatment in severe disease, where sur¬ gery may be the alternative, said process also enabling judgement of the risk of relapse in patients with clini¬ cally inactive UC and CC during maintenance therapy.

A further object of the invention is to provide a kit to be used in said process.

In order to satisfy the above objects and advantages provided by the present invention a process for (1) assessment of disease activity in patients with chronic inflammatory bowel disease characterized as ulcerative colitis (UC) and Crohn's colitis (CC) and for (2) pre¬ diction of (a) efficacy of medical treatment in case of active disease prior to start of therapy and (b) risk of relapsing disease in case of clinically inactive disease in association with or following discontinuation of re- lapse preventing medical treatment wherein equilibrium in vivo dialysis of faeces and/or equilibrium in vivo dialysis of rectum is carried out for determination of faecal and/or rectal concentrations of prostaglandin E ₂ (PGE^). The equilibrium dialysis ensures a preli- minary purification of luminal PGE-, which makes spe¬ cific and accurate analysis possible.

Thus in connection with the present invention it has been shown that the concentration of prostaglandin (PG) E~ in the rectal lumen is a highly sensitive marker of disease activity and that pretreatment PGE- con centrations are significantly higher in patients not re¬ sponding to therapy than in those responding to the rapy. Furthermore, it has been demonstrated herein that an increase in the luminal concentrations of PGΞ ₂ in patients with inactive UC precedes the development of symptoms, in addition to objective clinical findings, abnormal histology of the colonic mucosa, and changes in conventional laboratory indices in patients who later experience a relapse.

Using an advantageous embodiment of the process of the invention the determination of faecal concentrations of PGE^ is carried out on the pooled contents of orally administered dialysis bags to eliminate possible varia¬ tions due to differences in the transit time of bags ha- ving passed the gastrointestinal tract and determina¬ tion of rectal concentrations of PGE^ is carried out on the contents of a dialysis bag which has been placed for four hours in the patient's emptied rectum.

Using this embodiment of the process of the invention concentrations of PGE ₂ can be determined radioimmuno- logically by a method validated by gas chromatography- mass spectro etry and quantitative high pressure liquid

3 chromatography using 10-100 ul dialysate, H-labelled PGE_ as an internal standard, acidification, extrac¬ tion with ethylacetate/cyclohexane (1:1) and chromato¬ graphy on microcolumns of Sephadex LH-20 before perfor¬ ming the radioimmunoassay itself on the eluate fraction containing the purified PGE,,. Other methods of deter- mining the concentration of PGΞ-, could also be used, e . g. , an ΞLISA or other immunologic method.

The test kit to be used in the process of the invention is characterized by including dialysis bags for oral in¬ take and/or rectal insertion, in addition to essential and specific test substances for the performance of ra- dioimmunological determination of FGE^ in the contents of the dialysis bags.

According to the invention the dialysis bags are prefe- ra- bly manufactured from Visking seamless tubing 8/32 (with an average pore size of 24A) by tieing off 3 or 12 cm segments with a suture ("Poly" from Ethicon) and fil¬ ling the segments (i.e. the dialysis bags) with Rheoma- crode TM 100 mg/1 (i.e. lOOg dextran 40 and 9 g sa¬ line per liter.

According to the invention the test substances contained in the kit are lyophilized antibodies against PGΞ-, tritium labelled FGE ₂ , and non-radiolabelled VGE~ .

Brief Description of the Drawings

Fig. 1 illustrates how the concentrations of PGE~ in the rectal lumen of patients with ulcerative colitis are positively correlated to clinical, endoscopic, as well as histologic disease activity.

Fig. 2 illustrates how the concentrations of PGE- in the rectal lumen of 13 untreated patients with relapsing ulcerative colitis change after treatment with Pentasa (5-aminosalicylic acid) enemas (1 g/day) for 2-4 weeks and the individual outcome of treatment is classified as clinical and endoscopic remission, clinical but not en¬ doscopic remission, and unchanged or worse.

Fig. 3 illustrates how the concentrations of PGE _? in the rectal lumen of 11 untreated patients with relapsing ulcerative colitis change after treatment with predniso- lone enemas (25 g/day) for 2-4 weeks and the individual

outcome of treatment is classified as clinical and en¬ doscopic remission, clinical but not endoscopic remis¬ sion, and unchanged or worse.

Fig. 4 illustrates how pretreatment concentrations of PGE- in the rectal lumen of 24 patients with relapsing ulcerative colitis correlate to the outcome of medical treatment with pentasa (5-aminosalicylic acid) enemas (1 g/day) or prednisolone enemas (25 mg/day) for 2-4 weeks. The results show that pretreatment concentrations of PGE- may be used as a predictor of the outcome of me¬ dical treatment with the named drugs.

Best Mode for carrying out the Invention

Material and Methods (I)

PATIENTS AND CONTROLS

The study group comprised the first 24 patients from the Department of Medical Gastroenterology, Odense University Hospital, entering a randomized, double-blind Danish multicenter trial for comparison of the therapeutic efficacy of topical 5-ASA and prednisolone in ulcerative proctosigmoiditis. The inclusion criteria were as follows: (a) proven UC (V. Binder, H. Both, P.K. Hansen, C. Hendriksen, S. Kreiner, K. Torp-Pedersen, 'Incidence and prevalence of ulcerative colitis and Crohn's disease in the county of Copenhagen', Gastroenterology 1982,83:563-8) localized to the sigmoid colon or the rectum, or both; (b) symptoms and signs of mild CC-,) or moderate (C-) disease activity (V. Binder, 'A comparison between clinical state, macro¬ scopic and microscopic appearances of rectal mucosa, and cytologic picture of ucosal exudate in ulcerative coli- tis', Scandinavian Journal of Gastroenterology 1970,5:627-32); (c) no drug treatment for UC the preceding month apart from maintenance treatment with sulphasalazine (SAZ); (d) outpatient capable of adminis¬ tering enemas following instruction; (e) normal renal and hepatic function. For patients maintained on SAZ (1 g b.i.d. ) this treatment was kept unchanged throughout the trial. No other drugs were given.

In 8 healthy volunteers (3 men, 5 women; aged 18-69 yr) serving as controls, equilibrium in vivo dialysis of rectum was carried out on a single occation.

The study was done in accordance with the Helsinki Declaration II and approved by the Ethical Committee of Funen and Vejle Counties. Informed consent was given by all participants.

EXPERIMENTAL DESIGN

The study was carried out as a block-randomized, double- blind trial. Before entry, clinical disease activity was assessed, and equilibrium in vivo dialysis of rectum was performed before the assessment of endoscopic and histo¬ logic disease activity. The patients were randomly allo¬ cated to one of the two treatments, i.e., 100 ml enemas containing either 1000 mg 5-ASA in acidic buffer suspen¬ sion, pH 4.8 (Pentasa; Ferring A/S, Copenhagen, Denmark) or 25 mg prednisolone (Predniment, Ferring), which were given once daily at bedtime for 14 days. Primary strati¬ fication was done for patients already on maintenance treatment with SAZ. At day 15 the equilibrium in vivo dialysis of rectum was repeated, and clinical, endosco- pic, and histologic activities were reassessed. Patients in complete clinical and endoscopic remission discon¬ tinued treatment with enemas, whereas the remainder con¬ tinued therapy for another 14 days. In all cases equi¬ librium in vivo dialysis of rectum and assessment of di- sease activity were repeated at day 29.

Withdrawal criteria included intolerance to treatment or deterioration (severe disease, C_ ) requiring systemic glucocorticoids.

Laboratory screening was performed on all study days and included blood hemoglobin, reticulocyte count, erythro- cyte sedimentation rate, leukocyte count, leukocyte dif¬ ferential count, platelet count, and serum concentra¬ tions of creatinine and orosomucoid, in addition to urine analysis for protein and microscopy for erythro- cytes, leukocytes, and casts.

DISEASE ACTIVITY

The disease activity was assessed at each study day ac- cording to a four-scaled semiquantitative grading: (a) clinical activity : inactive ( ^c ₀ ) _r mild (C.), mode¬ rate ( ^C ₉ ^ _' severe (C ₃ ); (b) endoscopic activity

inactive (E ), mild (E. ), moderate (E~), severe (E_,); (c) histologic activity : inactive (H ), mild (H ₁ ), moderate (H ₂ ), severe (H ₃ ).

The clinical activity was based on a diary in which the number of bowel movements and the presence or absence of blood were registered by the patient and the objective findings were registered by the investigator. To mini¬ mize variation among observers each patient was examined on all study days by the same investigator, who was una¬ ware of the results of the rectal dialyses.

EQUILIBRIUM IN VIVO DIALYSIS OF RECTUM

In vivo dialysis of rectum was performed after insertion of a 12-cm-long dialysis bag (volume 4 ml; Visking seam¬ less cellulose tubing 8/32 filled with Rheomacrodex; Pharmacia, Uppsala, Sweden) containing 10% dextran (mean ol wt 40,000) in saline into the emptied rectum. To ob¬ tain equilibrium with the fluid covering the surface of rectal mucosa, the dialysis bag was left for 4 h before its contents were emptied into a test tube, which was stored immediately at -20 C (rectal dialysate) for analysis within 2 wk.

To ensure that PGΞ_ had reached an equilibrium, intact dialysis bags were incubated with ( 3H)PGE ₂ (50,000 cpm/ l) in phosphate-buffered saline (pH 7.4; 37°C).

After 1, 3, and 4 h, respectively, the radioactivities within the bags were 50%, 80%, and 85% of the measured in the external solution. Binding of PGΞ_ to the dia¬ lysis membrane was negligible (K. Lauritsen, J. Hansen, P. Bytzer, K. Bukhave, J. Rask-Madsen, 'Effects of sul- phasalazine and disodium azodisalicylate on colonic PGE- concentrations determined by equilibrium in vivo dialysis of faeces in patients with ulcerative colitis and healthy controls'. Gut 1984,25:1271-8).

ANALYTIC PROCEDURES

Prostaglandin E ₂ was measured as previously described in detail (K. Bukhave, J. Rask-Madsen J, 'Prostaglandin E- in jejunal fluids and its potential diagnostic va- lue for selecting patients with indomethacin-sensitive diarrhoea', 1981 European Journal of Clinical Investiga¬ tion, 11:191-7) by a radioimmunologic method validated by gas chromatography-mass spectrometry (K. Bukhave, K Green, J. Rask-Madsen, 'Comparison of radioimmunological determinations with gas chromatography mass spectrometry dosage', 1983 Biomedical Mass Spectrometry, 10:265-8). The method included purification by extraction with ethylacetate/cyclo- hexane (1:1) and chromatography on microcolumns of Sephadex LH-20 (Pharmacia AB, Sweden) before performing the radioimmunoassay itself on the eluate fraction containing PGE ₂ .

STATISTICAL ANALYSES

Nonparametric statistics were used. The results were gi- ven as medians with ranges and Q_ _Q interquartile ran¬ ges (defined by the lower and upper quartiles). The data were analyzed by the Chi square method, the Mann-Whitney test for unpaired variates, Wilcoxon's test for paired variates, or Kendall's rank correlation test as indi- cated either in the text or in footnotes in the tables. Values of p<0.05 were considered significant.

Results (I)

DESCRIPTION AND COMPARISON OF TREATMENT GROUPS The patients consecutively entered the trial and equili¬ brium in vivo dialysis of rectum was done three times in all cases but one. The condition of the latter patient, who was on 5-ASA, deteriorated and this patient was con¬ sequently withdrawn after the second week of treatment to receive systemic glucocorticoids. Table 1 describes patient characteristics and the outcome of treatment in the two groups. There was no statistically significant

difference with regard to age, sex, duration of disease, present disease activity, number of patients treated with SAZ, or outcome of the enema treatments with or without ad uvant SAZ.

In retrospect, the pretreatment assessment of clinical disease activity was a poor predictor of the outcome of treatment (p>0.05) in contrast to endoscopic and histo¬ logic indices (Table 2).

One patient in each treatment group complained of nau¬ sea, and 2 patients on 5-ASA reported difficulties in retaining the enemas. The laboratory screening dis¬ closed no clinically significant abnormalities, except for a slight increase in platelet counts and serum con¬ centrations of orosomucoid in a few cases.

PGE ₂ CONCENTRATIONS IN THE RECTAL LUMEN OF HEALTHY VO LUNTEERS AND PATIENTS WITH UC Healthy volunteers . Concentrations of PGE ₂ in rectal lumen of healthy controls (n=8) ranged from 0.15 to 0.59 ng/ml (median 0.32 ng/ml) .

Patients with active ulcerative colitis. The luminal concentrations of PGE_ were markedly elevated compared to healthy controls (p<0.05, Mann-Whitney test). A sig¬ nificant positive correlation (p<0.05, Kendall's rank correlation test) was found between concentrations of PGE- and disease activity judged by clinical, endosco- pic, or histologic gradings (fig. 1). Also, some pa¬ tients with clinically or endoscopically inactive di¬ sease (C or E ) had increased PGE- levels com¬ pared to healthy controls (Table 3). In the rectal dia- lysates obtained from patients assessed to have the same grade of disease activity, no statistically significant differences (p>0.05) were observed (a) between levels of PGΞ ₂ in patients allocated to 5-ASA and prednisolone

enemas, respectively, (b) between levels in patients untreated or off treatment and levels in patients trea¬ ted with enemas, and (c) between levels in patients with or without adjuvant SAZ treatment (Table 3). However, the numbers of observations within some of these sub¬ groups were small.

EFFECTS OF 5-AMINOSALICYLIC ACID AND PREDNISOLONE ON RECTAL PROSTAGLANDIN E ₂ CONCENTRATIONS IN ULCERATIVE COLITIS

Fig. 2 shows the effect of treatment with 5-ASA enemas on PGE ₂ concentrations in the rectal lumen of patients with active UC. No significant effects on the named va¬ riables could be traced in the whole group (p>0.5, Wil- coxon's test for paired variates), but PGE ₂ levels de¬ creased toward control levels in those patients respond¬ ing to treatment.

Similarly, fig. 3 illustrates the results of treatment with prednisolone enemas, which inhibited PGE ₂ concen¬ trations significantly (p<0.05).

RELATIONSHIP BETWEEN PRETREATMENT PROSTAGLANDIN Ξ ₂ CONCENTRATIONS IN ULCERATIVE COLITIS AND OUTCOME OF TREATMENT

Fig. 4 illustrates that the pretreatment PGΞ_ levels were signifcantly higher (p<0.05, Mann-Whitney test) in the patients not responding to therapy than in those who improved. The laboratory screening was normal apart from marginally elevated serum concentrations of orosomucoid in some of the former patients, all of whom required systemic glucocorticoids to control disease activity.

Discussion (I) Since S.R. Gould, 'Prostaglandins, ulcerative colitis, and sulphasalazine' 1975 The Lancet, 2:988, first repor¬ ted raised concentrations of PG-like material in stools

from patients with an acute attack of UC, many in vitro studies have shown that the amount of PGE ₂ and the ac¬ tivities of PG synthetase measured in fresh biopsy spe¬ cimens of rectal mucosa (as well as the generation of GE ₂ and the stable breakdown products of PGI ₂ and thromboxane A- in cultures of rectal mucosa from pa¬ tients with relapsing UC) were significantly higher than those obtained in material from healthy controls. These observations have been confirmed in vivo by means of the method of equilibrium dialysis of faeces (K. Lauritsen, J. Hanseh, P. Bytzer, K. Bukhave, J. Rask-Madsen, 'Ef¬ fects of sulphasalazine and disodium azodisalicylate on colonic PGE- concentrations determined by in vivo dia¬ lysis of faeces in patients with ulcerative colitis and healthy controls', 1984 Gut, 25:1271-8). In spite of these publications, it has not been previously apprecia¬ ted that the level of GE ₂ in a patient could be a va¬ lid marker of disease activity as well as a useful pre¬ dictor of the outcome of medical treatment and of the risk of relapse. The method of equilibrium in vivo dia¬ lysis of rectum used in the present study permitted the demonstration of a close positive correlation between PGE- _j and disease activity. This observation supports the notion that arachidonic acid metabolism plays a pa- thophysiologic role in UC. The PGs enhance vasodilation and edema formation, acting synergistically with other mediators of inflammation (e.g., histamine, bradyki- nin), whereas the leukotrienes promote leukocyte migra¬ tion and stimulate aggregation and degranulation of neu- trophils, in addition to release of lysozomal enzymes and superoxide production. A major argument against the hypothesis that the role of FGΞ- is primary rather than secondary is the observation that treatment with potent cyclooxygenase inhibitors, such as indomethacin and flurbiprofen, has no beneficial clinical effect, and may even provoke a relapse, in the face of reduced le¬ vels of PGs in the urine and in the rectum. However,

the named compounds do not affect the production of 5- lipoxygenase metabolites. On the other hand, drugs with proven efficacy in UC, such as the glucocorticoids and SAZ, seem to share such an effect, at least in vi- tro: The glucocorticoids are considered to inhibit phospholipase activity, necessary for the liberation of arachidonic acid from membrane bound phospholipids, thereby leading to decreased metabolism along both cy- clooxygenase and lipoxygenase pathways. Similarly, 5- ASA, the active moiety of SAZ, inhibits the lipoxygenase system in human neutrophils. The results of the present study, which demonstrate that rectal concentrations of PGE ₂ decrease toward control levels in patients who respond to therapy (i.e., 5-ASA or prednisolone), may, however, be interpreted to reflect the mode of drug ac¬ tion. Furthermore, the reduction in rectal PGE ₂ levels seems to occur promptly (i.e., within 12-72 h), at least after high-dose systemic treatment with glucocorticoids (K. Lauritsen, L. S. Laursen, K. Bukhave, J. Rask-Mad sen, 'In vivo effects of orally administered predniso¬ lone on prostaglandin and leucotriene production in ul¬ cerative colitis', 1987 Gut, 28:1095-99).

The patients who entered this study were only part of a larger multicenter clinical trial for comparison of the therapeutic efficacy of the two drug regimens. In this connection, it should be noted that most patients with high pretreatment PGE ₂ concentrations in the rectum were treatment failures and required high-dose systemic glucocorticoids to control the disease. Approximately 15% of all patients with UC confined to the rectum and sigmoid colon are resistant to most forms of therapy, and the management of so-called refractory pro'ctosigmoi- ditis is an unrewarding task. If pretreatment levels of PGΞ reflect local disease activity, they may prove more useful predictors of the outcome of medical treat¬ ment in patients with UC than clinical indices of di-

sease activity.

Material and Methods (II)

PATIENTS Patients fulfilling the following entry criteria were candidates for the study: (a) proven UC or proven CC; (b) relapsing disease (i.e., a relapse within the past six months) or chronic continous disease activity (i.e. absence of remission (definition: see below) for three months or more); (c) refractory to SAZ (i.e., relapsing disease despite maintenance treatment with SAZ 2 g/day or chronic continous disease despite treatment with glu¬ cocorticoids and SAZ) or intolerant of SAZ (i.e., aller¬ gic to or otherwise intolerant of SAZ) .

Patients below 18 or above 80 years of age were exclu¬ ded, as were women who were pregnant or lactating and those with childbearing potential not using oral contra¬ ception or an intrauterine device. Furthermore, patients with CC, who had also radiologic verified small bowel disease were excluded.

Each patient gave informed consent and the study was ap¬ proved by the Ethical Committee of Funen and Vejle Counties.

EXPERIMENTAL DESIGN

The study design was open with an initial dose of ADS 1 g/day (0.5 g b . i. d . ) , in SAZ intolerant patients. In SAZ refractory patients the initial dose was 2 g/day (1 g Tb. i . d. ) . The dose was gradually increased at the sche¬ duled visits until the endpoint was reached (i.e., su¬ stained remission - as defined below), or drug intole¬ rance occurred, or a maximum of 4 g/day (1 g q. i. d . ) equivalent to approximately 10 g/day of SAZ as regards 5-ASA on a molar basis. In patients on concomitant cor¬ ticosteroids the dose was gradually tapered and, if pos-

sible, the administration was stopped.

A one year treatment period was planned including scheduled visits at two weeks and at two, four, six, eight, ten, and 12 months, in addition to in-between vi¬ sits, if needed. Before entry and at each visit the cli¬ nical disease activity was assessed (see below) and a laboratory screening was done. This included blood hae¬ moglobin, reticulocyte count, erythrocyte sedimentation rate, leucocyte count, leucocyte differential count, platelet count, serum concentrations of Na, K, Ca, albu¬ min, creatinine, urea, alanine aminotransferase, lactate dehydrogenase, alkaline phosphatase, bilirubin, and plasma prothrombin, in addition to urine analysis for protein and glucose and microscopy for erythrocytes, le¬ ukocytes, and casts. Additional serum was obtained for determination of ADS, 5-aminosalicylic acid (5-ASA), and acetyl-5-aminosalicylic acid (Ac-5-ASA).

Before entry, at the two week's visit, and at the six and 12 months' visits equilibrium in vivo dialysis of faeces as well as equilibrium in vivo dialysis of rectum were done prior to endoscopy of rectum and sigmoid colon with biopsies. ADS was administered in gelatin capsules, each containing 250 mg of the powdered drug without ad¬ ditives.

Compliance to prescription was estimated at each study visit by counting the number of returned ADS capsules.

DISEASE ACTIVITY

The disease activity was graded according to the four- scaled semiquantitative grading: (a) clinical activity: inactive (C _n ) , mild (C.), moderate (C ₂ ), severe (C«); (b) endoscopic activity: inactive (E«), mild (E ₁ ), moderate (E ₂ ), severe (E ₃ ) ; and (c) histolo¬ ic activity: inactive (H„) , mild (H- _j ), moderate

(H.-), severe (H._). The clinical activity was based on a diary and the objective findings registered by the investigator. To minimize variation among observers each patient was examined on all study days by the same in- vestigator. All days the week preceeding a scheduled vi¬ sit the patient recorded the number of bowel movements in the diary, in addition to the presence or absence of blood, and graded abdominal pain and general well- being according to W. R. Best, J. M. Becktel, J. W. Singleton, 'Rederived values of the eight coefficients of the Crohn's Disease Activity Index (CDAI)', 1979 Gastroenterology, 77:843-6) for calculation of the Crohn's disease activity index (CDAI) . Remission in pa¬ tients with UC was defined as clinically inactive di- sease C~; i.e., below three bowel movements a day without blood present) and no evidence of inflammation (E ₀ ) on sigmoidoscopy. Remission in patients with CC was defined as a CDAI <150 and no evidence of inflam¬ mation on sigmoidoscopy. Patients in whom remission did not occur without concomitant use of corticosteroids were classified as treatment failures. Due to the open design of the study no further clinical or endoscopic estimations of efficacy were done.

IN VIVO DIALYSIS METHODS

Equilibrium in vivo dialysis of faeces was done as pre¬ viously described in detail by pooling the contents of five swallowed dialysis bags following their intestinal transit (faecal dialysate). These dialysates were ana- lyzed for concentrations of ADS, 5-ASA, and Ac-5-ASA, in addition to PGΞ ₂ (faecal PGΞ ₂ ) .

Equilibrium in vivo dialysis of rectum was performed af¬ ter insertion of a 12-cm-long dialysis bag into the emp- tied rectum as previously described (K. Lauritsen, L.S. Laursen, K. Bukhave, J. Rask-Madsen, "Effects of topical 5-aminosalicylic acid and prednisolone on prostaglandin

E ₂ and leukotriene B, levels determined by equili¬ brium in vivo dialysis of rectum in relapsing ulcerative colitis, Gastroenterology 1986;91:837-844). The bag was left for four hours and its contents (rectal dialysate) were analyzed for concentrations of PGE ₂ (rectal PGE ₂ ).

ANALYTICAL PROCEDURES

Determination of ADS, 5-ASA, and Ac-5-ASA in serum and faecal dialysates were done by high performance liquid chromatography (HPLC) as described above. The lower de¬ tection limits in serum and in faecal dialysates were 1.0 umol/1 and 6.0 umol/1, respectively. Prostaglandin E ₂ was measured by the radioimmunologic method, re- ferred to above, validated by gas chromatography- ass spectrometry. These measurements were done blindly and independently of the assessment of disease activity and clinicians were unaware of the results until the da¬ ta had been completed.

REFLECTIONS ON SAMPLE SIZE

This trial was a pilot series with the purpose of pro¬ viding information for the planning of controlled trials of long term ADS treatment in UC and CC. Among all it was speculated that in such trials any fixed-dose regi¬ men might be associated with changed pharmacokinetical characteristics (e.g., incomplete splitting of the azo bond and subsequently changed lu inal concentrations of 5-ASA in active* disease) depending on the extension and the severity of disease, which might influence the tole¬ rance and the efficacy of the drug. It was considered relevant to compare colonic olsalazine metabolism and tolerance, (a) in patients with extensively located co¬ litis and in patients with 'left-sided' colitis, (b) in SAZ refractory patients and in SAZ intolerant patients, (c) in patients with chronic continuously active di¬ sease and in patients with intermittently active di-

sease, (d) in patients with severe disease activity (C«) and in patients with moderate and mild disease activity (C..-C-), and (e) in patients with UC and CC. For each such comparison a group size of eight to 12 was considered to provide relevant pharmacologic infor¬ mation. It was estimated, therefore, that inclusion of a total of 40 to 50 patients in the study would be de¬ sirable.

STATISTICAL ANALYSES

The changes, if any, in safety variables after entry were tested for trends using the paired t-test and the Wilcoxon matched pairs signed ranks test, as appro¬ priate. Values of p<0.05 were considered significant.

Results (II) DESCRIPTION OF PATIENTS

Forty patients (31 with UC and nine with CC) consecu¬ tively entered the study within a six month period. Four more patients with UC and one with CC fulfilled the entry criteria within the period of accrual but refused participation. Table 4 describes patient characteri¬ stics. The duration of treatment and the mean dose per patient and total doses used are shown in Table 5, whereas Table 6 gives the duration of the individual do¬ sages administered throughout the treatment period.

WITHDRAWALS, TOLERANCE, AND SAFETY

Eight patients did not complete the planned one year treatment period: two patients with UC and one with CC had a colectomy performed due to lack of efficacy after four, seven, and eight months, respectively. One patient with a 12 year history of UC had a colectomy after six months' treatment because biopsies revealed a rectal mu- cosal adenocarcinoma. Three patients wished to stop treatment prematurely, in two cases because of lack of efficacy (one with UC after eight months, one with CC

after ten months), and a third with CC after three months because she was in complete remission. Finally, one patient with UC in remission was excluded after eight months because he had not taken the study medi- cine.

Tolerance problems after ADS prescription such as loose or watery stools and/or increased stool frequency were reported in 11 cases (seven with UC, four with CC). In- tolerance occurred most often (nine cases) in patients with extensively located disease and active disease (nine cases), but was transient in eight cases (i.e., resolved within a few weeks during continuous admini¬ stration of the drug). No patients were withdrawn due to drug intolerance. Two patients reported transient pain in the bones and joints, respectively. The Wilcoxon matched pairs signed ranks test revealed no significant differences in laboratory screening values before and after the trial. In six cases slightly increased levels of alanine aminotransferase or alkaline phosphatase were found, but the values normalized during continuous treatment.

CONCENTRATIONS OF OLSALAZINE, 5-AMINOSALICYLIC ACID, AND ACETYL-5-AMINOSALICYLIC ACID IN FaecAL DIALYSATES

The average concentrations of ADS, 5-ASA, Ac-5-ASA, and total 5-ASA (5-ASA + Ac-5-ASA) measured in faecal dialy¬ sates from each patient were calculated for each dosage of ADS. The results obtained in patients with UC are gi- ven in Table 7. The results obtained in the nine pa¬ tients with CC were within the range of those obtained in UC treated with equal dosages of ADS, but the number of observations in the CC group precluded statistical comparison with the UC group. Concentrations of Ac-5-ASA in faecal dialysates were rather constant within the applied ADS dose range (from 1 g/day to 4 g/day), but concentrations of free 5-ASA, and thus also concentra-

tions of total 5-ASA increased dose dependently (Table 7). ADS levels in faecal dialysates were low (Table 7) compared to those of the split product (5-ASA) and its metabolite (Ac-5-ASA). The efficacy of azoreduction was

5 estimated by calculation of the 'split ratio', i.e., the ratio between molar concentrations in -faecal dialysates of total 5-ASA and total 5-ASA + 5-ASA contents in unsplit ADS, respectively. The median split ratio was 99.4% ( Q- _Q interquartile range 93.8-100%, range 40.1-

10 100%) in all dialysates obtained from patients with UC. In 28% of all patients the split ratio was below 90%, at least on one occasion, and more than half (six of 11) of these cases reported transient intolerance to the pre¬ scribed ADS dose. Resolution of drug intolerance was as-

15 sociated with an increased split ratio.

During active disease the mean concentrations of 5-ASA and Ac-5-ASA in faecal dialysates tended to be lower than in dialysates from patients in remission (Table 8),

20 but the number of observations on each dose of ADS was too small to permit statistical analysis. No differences between 5-ASA levels in patients with extensively loca¬ ted colitis and patients with 'left-sided' disease were found.

25

CONCENTRATIONS OF OLSALAZINΞ, 5-AMINOSALICYLIC ACID, AND ACETYL-5-AMINOSALICYLIC ACID IN SERUM

The average serum concentrations of ADS, 5-ASA, and Ac- 5-ASA in each patient were calculated for each dosage of

30 ADS. The results are given in Table 9 for patients with UC. Dose dependent increases in the levels of olsalazine and 5-ASA were observed within the whole dose range, but no further increase in serum concentrations of Ac-5-ASA occurred beyond a dose of ADS equal to 3 g/day (Table

35 9) . No statistically significant differences (p>0.05) were found between serum concentrations of ADS, 5-ASA, and Ac-5-ASA in patients with UC and CC, respectively,

when treated with an equal dosage of ADS. Similarly, no statistically significant differences (p>0.05) were found between levels of ADS in patients with active di¬ sease and patients in remission (Table 8) or in patients with pancolitis and patients with 'left-sided' disease. In individual patients on a stable dose of ADS no in¬ crease in serum concentrations of the unsplit drug or its metabolites were observed in the long term.

EFFECTS ON PROSTAGLANDIN E ₂ IN FaecAL AND RECTAL DIA¬ LYSATES

The levels of PGE ₂ in faecal and rectal dialysates de¬ creased significantly over time in patients with UC (Table 10). Concentrations of PGE ₂ were signifi- cantly raised in patients with active UC (C.-C ) compared to those obtained in patients in remission (C _n ). In 13 UC patients, in whom remission never oc¬ curred (treatment failures), pre-trial rectal PGE- concentrations were significantly raised (p<0.05) com- pared to those observed in patients not classified as treatment failures. The rectal PGE ₂ concentrations in treatment failures persisted markedly increased during the medication period, whereas those observed in the re¬ maining cases decreased toward the normal range. Thus also post-trial rectal PGE ₂ concentrations were in¬ creased in treatment failures (p<0.05). All but one of seven patients with pre-treatment rectal PGE- concen¬ trations >30 ng/ml were classified as treatment fai¬ lures.

By contrast, no statistically significant trends were observed in PGE ₂ levels in patients with CC (Table 10).

ESTIMATES ON EFFICACY

Thirteen patients with UC were regarded as treatment failures (i.e., clinical and sigmoidoscopic remission

was never obtained). This outcome was observed in one of eight patients intolerant of SAZ and in 12 of 23 pa¬ tients refractory to SAZ (p=0.12, Fisher's exact test). In six cases the disease was 'left-sided', whereas seven cases had a pancolitis. In the remaining patients a re¬ lapse occurred on at least one occasion: four experi¬ enced a relapse on 1 g/day, two on 2 g/day, eight on 3 g/day, and two on 4 g/day.

Five of the nine patients with CC were classified as treatment failures (i.e., a CDAI <150, without sigmoido- scopic evidence of inflammation, was never obtained).

Discussion (II) Previous studies have shown that small intestinal ab¬ sorption and metabolism of ADS is minimal and that a single oral dose is completely recovered from ileostomy fluid (H. Sandberg-Gertzen, M. Ryde, G. Jarnerot, 'Ab¬ sorption and excretion of a single 1-g dose of Azodisal sodium in subjects with ileostomy' 1983 Scandinavian Journal of Gastroenterology, 18:107-11). In patients with inactive UC complete azoreduction of olsalazine oc¬ curs on 1 g h . i. d. and concentrations of 5-ASA in faecal dialysates, in fact, double when SAZ is replaced by the same dose of ADS (K. Lauritsen, J. Hansen, M. Ryde, J. Rask-Madsen, 'Colonic azodisalicylate metabolism deter¬ mined by in vivo dialysis in healthy volunteers and pa¬ tients with ulcerative colitis' , 1984 Gastroenterology, 86:1496-1500). Oral ADS 0.5 g h . i . d. appears more effec- tive than placebo in preventing relapse in SAZ intole¬ rant patients (H. Sandberg-Gertzen, G. Jarnerot, W. Kraaz, 'Azodisal sodium in the treatment of ulcerative colitis. A study of tolerance and relapse-prevention properties', 1986 Gastroenterology, 90:1024-30) and 1 g h . i. d. orally for two weeks is more effective than pla¬ cebo in the treatment of mildly active UC (W. S. Selby, G. D. Barr, A. Ireland, C. H. Mason, D. P. Jewell, '01-

salazine in active ulcerative colitis', 1985 Br Med J, 291:1373-5). The major adverse effect in these prelimi¬ nary clinical trials was loose or watery stools, which were observed only in 10% of patients. This complaint usually occurred in patients with extensively located colitis and might be explained by the action of the unsplit ADS molecule that, unlike SAZ, markedly increa¬ ses ileostomy output

(H. Sandberg-Gertzen, G. Jarnerot, K. Bukhave, K. Lau¬ ritsen, J. Rask-Madsen, 'Effect' of azodisal sodium and sulphasalazine on ileostomy output of fluid, PGE ₂ , and PGF^^. in subjects with a permanent ileostomy', 1986 Gut,27:1306-11) . The efficacy of azoreduction in active disease is unknown, but indirect evidence points to a reduced azoreduction of SAZ. Among all it might be spe¬ culated that the intestinal transit time as well as the character of intestinal microflora depend on the exten¬ sion and the severity of the disease and thus the di- stribution of ADS and its metabolites. This concept is supported by the present results demonstrating that in¬ tolerance is associated with insufficient azoreduction. The increased incidence of diarrhea observed in the pre¬ sent series most likely reflects the use of a higher dose of ADS than in the above mentioned previous study. Symptoms of this type of intolerance call, therefore, for a temporary reduction of the dose and the problem is more readily handled in clinical practice than in con¬ trolled trials with a fixed-dose regimen. Also the oc- currence of diarrhea, which often subsided sponta¬ neously with continued therapy, suggests that the thera¬ peutic effect of 5-ASA may influence colonic function to permit a more effective azoreduction.

The present results also demonstrate that high doses of ADS can be administered safely in the long term and that azoreduction of ADS is almost complete in the major pro-

portion of patients, even during administration of the highest dose (4 g/day). In this pharmaceutical respect, the ADS capsules fulfils the anticipation of providing a highly effective means of delivery of 5-ASA to the colo- nic mucosa, even in active disease. The results accord with our previous experience in patients with inactive UC and in healthy controls. No other 'second generation' candidate reported on and developed to replace SAZ - whether azobound or a pharmaceutically modified free 5- ASA preparation - has hitherto been shown to provide co- Ionic concentrations of 5-ASA in the same order of mag¬ nitude as those demonstrated in the present study.

The findings of rather constant concentrations of Ac-5- ASA in faecal dialysates and dose dependently increasing concentrations of 5-ASA, indicate that saturation of the intraluminal acetylation capacity occurs at a low do- sage. This may be clinically important, because a thera¬ peutic effect of Ac-5-ASA has not been unequivocally established (C. P. Willoughby, J. Piris, S. C. Truelove, 'The effect of topical N-acetyl-5-aminosalicylic acid in ulcerative colitis', 1980 Scand J Gastroenterol,15:715-9 and V. Binder, S. Halskov, E. Hvidberg, et al., 'A con¬ trolled study of 5-acet-aminosalicylic acid (5-Ac-ASA) as enema in ulcerative colitis' (Abstract) 1981 Scand J Gastroenterol,16:1122).

The serum concentrations observed on ADS 1 g/day and 2 g/day are comparable to those reported in healthy con¬ trols. The results (Table 9) indicate a dose related in- crease in absorption of ADS and its metabolites, but do not suggest that cumulation occurs in the long term.

The demonstration of extremely high intraluminal levels of PGΞ ₂ in treatment failures accords with our pre- vious findings and suggests that these quantifications of local disease activity are useful predictors of the outcome. We propose such determinations should be used

prospectively in clinical trials on drug efficacy in UC.

The design of the present study did not permit any esti¬ mation of a dose response effect or a correlation be- tween faecal 5-ASA levels and response to treatment. In planning the study we chose for safety reasons the prag¬ matic viewpoint to include only 'problematic' patients, who were not satisfactorily managed by conventional treatment (i.e., SAZ and a short course of gluαocorti- coid), although ADS may prove to be a SAZ alternative. More treatment failures occurred among patients refrac¬ tory to SAZ than among patients intolerant of SAZ. The complete splitting of olsalazine and the high con¬ centrations of free 5-ASA in faecal dialysates from SAZ refractory patients, who did not experience a remission, and from patients who relapsed, although maintained on a 4-g dose of ADS, indicate that SAZ resistance is not merely a question of dosage, at least in a range up to 10 g/day (i.e., equal to 4 g/day of ADS).

In conclusion, prolonged administration of high doses of ADS to patients with UC and CC appears acceptably well tolerated. The drug is a . very effective means of deli¬ very of 5-ASA to the colonic mucosa and, despite some cases of therapeutic resistance to even high intralumi¬ nal concentrations of 5-ASA, ADS seems to be a drug particularly appropriate for testing the efficacy of 5- ASA through controlled trials in patients with UC and CC in the long term.

Material and Methods (III) PATIENTS

Patients fulfilling the following entry criteria were candidates for the study: (a) proven UC. The diagnosis had been established on the basis of symptoms, endosco¬ pic findings, histology of the rectal mucosa, and radio- logic appearance, in addition to an exclusion of infec-

tious diarrhea by stool cultures and microscopy; (b) in¬ active disease (remission) for at least 12 mo during maintenance treatment with sulphasalazine (SAZ; Pharma¬ cia AB, Uppsala, Sweden) 2 g/day. Patients were consi- dered in remission when symptom-free (i.e., stool fre¬ quency 2 x a day, without discharge of blood, pus, or mucus from the rectum), with normal sigmoidoscopic ap¬ pearance, and with no significant inflammation on rectal biopsy.

The study was done in accordance with the Helsinki De¬ claration II and approved by the Ethical Committee of Funen and Vejle Counties. Informed consent was given by all partici ants.

EXPERIMENTAL DESIGN

Equilibrium in vivo dialysis of faeces as well as equilibrium in vivo dialysis of rectum were done in con¬ senting patients fulfilling the entry criteria. Subse- quently, SAZ treatment was stopped and the patients were observed free of medication. Evaluation of clinical di¬ sease activity, in addition to sigmoidoscopy with rectal biopsies for assessing endoscopic and histologic disease activity, were done at scheduled visits after two weeks, after two, six, and 12 months, and in-between, if indi¬ cated by recurrence of colitic symptoms. A relapse was defined as a recurrence of colitic symptoms accompanied by endoscopic signs of inflammation. If a relapse oc¬ curred the patient was withdrawn. At the scheduled vi- sits also in vivo dialysis measurements were carried out. If faecal CΞ ₂ concentrations and/or rectal PGΞ- concentrations exceeded control levels (i.e., faecal PGE ₂ >0.5 ng/ml; rectal GΞ ₂ >1.0 ng/ml) at any study day the patient was allocated at random to double blind maintenance treatment with SAZ (tablets 0.5 g; 2 tablets h. i . d . ) or placebo tablets (2 tablets h. i . d . ) of identical appearance for six months. The SAZ

and placebo tablets were provided by Pharmacia AB, Upp¬ sala, Sweden. The randomisation was done according to a computer generated list in blocks of four, stratified for visit (two weeks, two months, six months, 12 months) . In randomized patients a final visit was sche¬ duled after six months of randomisation for in vivo dia¬ lysis measurements and for assessment of clinical, endo¬ scopic, and histologic disease activity in case a re¬ lapse had not occurred earlier.

To minimize the variation between observers each patient was examined on all study days by the same investigator, who was unaware of the results of the dialysis measure¬ ments.

A laboratory screening was performed on all study days and included blood haemoglobin, sedimentation rate, leu¬ cocyte count, leucocyte differential count, platelet count, and serum concentrations of creatinine and oroso- mucoid.

Thus patients completed the trial either (a) after 12 months of passive observation (in cases with normal PGE ₂ levels throughout); or (b) six months after ran- domisation (in case PGE ₂ levels exceeded control le¬ vels); or (c) in-between, if a relapse occurred.

Results obtained in the cohort of randomized patients served to evaluate the primary purpose of the study, that is, to test the predictive value of 'an increase in PGE ₂ levels' for experiencing a relapse in patients with inactive UC off SAZ treatment. Furthermore, this cohort also permitted a secondary purpose of evaluating the efficacy of SAZ in maintaining remission in the long term.

IN VIVO DIALYSIS METHODS

Equilibrium in vivo dialysis of faeces was carried out as previously described, by pooling the contents of five swallowed dialysis bags following their intestinal tran- sit (faecal dialysate). These dialysates were analyzed for concentrations of GE ₂ (faecal PGE ₂ ).

Equilibrium in vivo dialysis of rectum was performed af¬ ter insertion of a 12-cm-long dialysis bag into the emp- tied rectum as described above. The bag was left for four hours when it was removed and its contents (rectal dialysate) were analyzed for concentrations of PGE ₂ (rectal PGE ₂ ).

ANALYTICAL PROCEDURES

Prostaglandin E ₂ was measured as previously described by a radioimmunological method validated by gas chroma¬ tography mass spectrometry. The method, using 10-100 ul

3 samples, included addition of H-labelled ^GE (Amersham International, Buckinghamshire, U.K. ) as an internal standard, acidification, extraction with ethyl- acetate/cyclohexane (1:1), and chromatography on micro- columns of Sephadex LH-20 (Pharmacia AB, Uppsala, Swe¬ den) before performing the radioimmunoassay itself on the eluate fraction. The detection limit, defined as the procedure blank value +2 SD, was 0.03 ng/ml.

Further evaluations of the radioimmunological procedures for determination of PGE_ in 'rectal dialysates' were performed by comparing data obtained by simultaneous analysis of aliquots of the same samples (obtained from patients with active UC and controls) by radioimmunoas¬ say and quantitative high pressure liquid chromatography for PGE ₂ - All organic solvents were from Merck, Darm- stadt, F.R.G. The results were expressed as ng/ml of co- Ionic fluid. In ten healthy controls the following re¬ sults have previously been obtained: Faecal PGE_ <0.5

ng/ml and rectal PGE ₂ <0.6 ng/ml.

STATISTICAL ANALYSIS

Fisher's exact test was used for analysis of fourfold tables. A value of p <0.05 was considered significant.

Results

DESCRIPTION OF PATIENTS

Twenty-four patients consecutively entered the trial within an 18-month period. One patient was withdrawn be¬ cause revision of her records showed that she had CC. Three more patients fulfilled the entry criteria, but refused participation. Table 11 describes characteri¬ stics of 23 patients eligible for the study.

PROSTAGLANDIN E ₂ LEVELS IN FaecAL AND RECTAL DIALY¬ SATES

Table 12 lists the faecal and rectal PGE ₂ levels in individual patients throughout the trial. In 11 patients faecal PGE ₂ levels exceeded 0.5 ng/ml (six cases) or rectal PGE ₂ levels exceeded 1.0 ng/ml (all 11 cases) after withdrawal of SAZ: in six cases at two weeks; in one case at two months; in two cases at six months; and in two cases at 12 months (Table 12). Six of these pa- tients were randomized to SAZ and five were randomized to placebo. The outcome of the disease within the six- month treatment period is also shown in Table 12: In the SAZ group none of the six patients experienced a relapse and PGE„ concentrations decreased toward control le- vels. By contrast, four of five patients in the placebo group experienced a relapse (p <0.05, Fisher's exact test) .

Twelve patients remained non-randomized and ten of these experienced a relapse. Retrospectively: In two cases (both relapsing within two weeks after cessation of SAZ treatment) PGΞ- levels were increased at entry (Table

12: no. 1 and 3); in four cases (Table 12: no. 6, 9, 11, and 13) the PGE ₂ levels increased more than two fold, but did not reach the randomisation levels. The PGE ₂ concentrations persisted within control levels in the patients not randomized and not experiencing a relapse (Table 12: no. 7 and 21).

CLINICAL OUTCOME

The clinical outcome is listed in Table 12. All relapses were proven by histology of rectal biopsies. As de¬ scribed above, a relapse occurred in none of six and in four of five patients, respectively, randomized to SAZ and placebo (p=0.03) and in ten of 12 non-randomized pa¬ tients. The cumulative relapse rate among patients not randomized to SAZ (i.e., cumulative numbers with a re¬ lapse/numbers at risk off SAZ) was 53% (10/19) and 78% (14/18) after six and 12 months, respectively (Table 12).

In all patients relapsing, and thus withdrawn from the trial, remission was reinstated: in seven cases by resu¬ ming SAZ treatment 2-3 g/day orally; in six cases by re¬ suming SAZ 2-3 g/day orally, in addition to topical treatment with corticosteroids, for two to eight weeks; and in a single case after prolonged treatment with sy¬ stemic corticosteroid, in addition to SAZ.

Discussion (III)

The present results suggest that increased PGΞ ₂ con- centrations in rectal dialysis fluid determined by equi¬ librium in vivo dialysis of rectum are predictive of a relapse in patients with inactive UC. First, in the co¬ hort randomized (because such an increase in PGE ₂ con¬ centrations had occurred) four of five patients alloca- ted to placebo relapsed within six months, whereas the other six patients, allocated to SAZ, all remained in remission and showed a subsequent decrease in PGE_

concentrations toward control levels. Second, in both patients, who relapsed within 14 days after cessation of SAZ therapy, the GE ₂ concentrations proved to have exceeded control levels at entry. Third, in none of the ten remaining patients the demonstration of a normal PGE ₂ value was followed by an immediate (i.e., within two weeks) relapse. However, the observation, that only two of these patients completed the observation period of 12 mo in remission, demonstrates that the 'lack of an increase in PGE ₂ concentration' is predictive for not experiencing a relapse in the short term only. Thus, to improve decision-making and to limit the risk of a re¬ lapse in patients off SAZ treatment would require more frequent dialysis collections than performed in the pre- sent study.

The faecal and rectal PGE ₂ concentrations determined by the methods of equilibrium in vivo dialysis of faeces and equilibrium in vivo dialysis of rectum, respective- ly, are both sensitive markers of disease activity. In patients with inactive UC an increased PGE ₂ concentra¬ tion in luminal dialysis fluid, whether faecal or. rec¬ tal, seems indicative of a mucosal abnormality, even in the absence of conventional signs of inflammation. The data indicate that the rectal dialysis method is more sensitive than the in vivo dialysis of faeces: in all but one case (Table 12) the detection of an increased faecal PGE- concentration (>0.5 ng/ml) was associated with a concomitant increase in the rectal CE ₂ con- centration (>1.0 ng/ml) and in no case a patient was randomized because of an increased faecal (but a normal rectal) PGE._ concentration. Oppositely, an increased rectal PGΞ~ concentration was often accompanied by a normal faecal PGE_ concentration (Table 12). This may be of clinical relevance, because outpatients experience the collection of stools more tedious than the per¬ formance of equilibrium in vivo dialysis of rectum.

The cumulative relapse rates observed among patients off SAZ therapy in the present study are in keeping with previous results, which have shown a relapse rate of 55% and 69% over six months, and 75% over 12 months. Fur- thermore, the present data are in agreement with the findings of A. S. Dissanayake, S. C. Truelove, 'A con¬ trolled therapeutic trial of long-term maintenance treatment of ulcerative colitis with sulphasalazine (Sa- lazopyrin) ' ,1973 Gut,14:923-6 and later confirmed else- where, that SAZ is superior to placebo or other re¬ medies, such as a high-fiber diet or mast-cell stabili¬ zers, in maintaining remission in UC in the long term.

Based on the present data it might be argued that all patients with inactive disease should stay on SAZ in¬ definitely, because the relapse rate was high. However, the recurrences experienced in the present series were easily handled in most cases and the courses were un¬ complicated. To decide on cessation or continuation of SAZ in patients with inactive UC, one needs, therefore, to address other aspects of this relapse preventing in¬ tervention, for example, potential prognostic gains, side effects, costs, in addition to patients' own expec¬ tations and wishes. The latter is sometimes expressed as a keen desire to discontinue the longstanding drug in¬ take. The present study describes a method for improving decision-making in individual patients demanding to go off SAZ. The data demonstrate that an increased PGE ₂ concentration in rectal dialysis fluid identify patients with a substantial risk of relapsing. However, only a few experience prolonged remission, which is the reason why frequent measurements are warranted.

abl e 1 Description of t/roups and ou tcome of treatment ___— 5-Aminosalicylic acid Prednisolone

(n=13) (π=H)

Age, mean year (range)

Sex, male/ emale

Duration of disease, median yr (range)

Clinical activity, mild (C /moderate (C„)

Endoscopic activity, mild (E )/moderate (E„/ uevei'u (B ₃ ) c

Histologic activity mild (II )/moderate (H ₂ yuvure (ll ₃ )

Additional maintenance treatment, wi th/wlthout sul asalazine (2g/day) . Duration of treatment, 2 wlc/4 v_k

Outcome of treatment

Clinical and endoscopic remission (on SAZ) 3 (1) 0 (3)

Clinical but not endoscopic remission (on SA7.) 4 (3) 1 (1) llnαbanged or worse (on sulfasala&ine) 6 (3) 2 (2)

a Hot significant (p>0.05) by Mann-Whi ney test. ^l) Not significant (p>0.05) by Chi square unit..ad. Biopsy specimens were insuf icient for histologic grading in 5 patients.

89/05452

36

Table 2 Pretreatment gradings of clinical , endoscopic and histologic disease activity as predictors of outcome of treatment with 5-aminosalicylic acid or prednisolone enemas for 2 - 4 weeks.

Pretreatment Outcome of treatment semiquantitative Improvement Unchanged or Probability grading (n=16) worse (n=8) value

Clinical p > 0.05

6 2 10 6

Endoscopic p < 0.01

Histologic p > 0.05

^H l 3 0 ^H 2 7 1 2 6 a Chi square method. bBiopsy specimens were insufficient for histologic grading in 5 patients.

Table 3 Concentrations of PGE„ (ng/ml) in rectal lumen of patients with ulcerative colitis at different grades of activity, untreated or treated with 5-aminosa¬ licylic acid or prednisolone enemas and/or sulfasalazinβ

Enema SAZ ^α Clinical activity Endoscopic activity (+/-) 0 1 2-3 0 1 2

5-ΛSΛ 0.71(n=7) 24(n=5) 33(n=0) 0 ,75(n=3) 5.5(n=4) 24(n=0) 31(n=5)

(0.10-6.0) (2.5-50) (14-120) (0 ,10-3.7) (0.29-6.2) (2.5-58) (20-120)

Predni¬ ^' /- 1.6(n=9) 5.4(n=6) 9.5(n=3) 0 ,82(n=0) 2.B(n=7) 10.5(n=3) solone (0.16-5.0) (0.50-42) (5.9-12) (0.16-5.0) (0.57-3.9) (5.5-42) None ^d 0.62(n=6) 9.8(n=9) 9.0(n=10) 0.61(n=5) 3.2(n=5) 8.2(n=17) 21(n=6) (0.47-2.2) (0.35-36) (0.32-52) (0.47-2.2) (0.35-0.8) (0.32-36) (0.9-5.2)

0.61(n=3) 9.4(n=5) 0.5(n=9) 0.61(n=3) 3.2(n=3) 8.2(n=0) 36(n=3)

(0.47-2.2) (0.35-36) (0.32-52) (0.47-22) (0.35-5.6) (0.32-36) (8.9-52)

0.62(n=3) 9.3(n=4) 10(n=9) 0.86(n=2) 4.7(n=2) 0.2(n=9) 16(n=3)

(0.52-1.2) (3.4-11) (3.4-26) (0.52-1.2) (0.62-8.0) (3.4-26) (11-26)

5-ASΛ, 5-aminosalicylic acid; PGE ₂ , prostaglandin E _2> Values represent median (range). ^a SΛZ, sulfasalazine: whole groups (+/-) irrespective of SAZ therapy or not, or subgroups with ( or without (-) SAZ maintenance therapy. Clinical activity: 0 (inactive), 1 (mild), 2-3 (moderate-severe). ^c Endoscopic activity: 0 (inactive), 1 (mild), •_> (moderate), 3 (severe). ^d None, before enema treatment or off treatment.

Table 4 Description of patients.

Ulcerative colitis Crohn's colitis (n=31) (n=9)

Age, mean year (range)

Sex, male/female

Smokers/non-smokers CO

Duration of disease, mean year (range)

Course last year, continuously active/intermittently active Sulphasalazine, refractory/intolerant Extension of disease, extensively located ('total')/'left sided'

able 5 Duration of treatment and olsalazlne doses used in patients with ulcerative colitis and Crohn ^' s colitis

Ulcerative colitis Crohn's colitis (n=31) (n=9)

Duration of treatment days 341+69 319±95 (132-390) (93-383)

Mean daily dose' g/day 2.610.7 2.5 ±0.9 (1.0-3.6) (1.0-3.5)

Total dose in treatment period g 8721310 7941396 (369-1349) (217-1309)

Values represent mean±SD (range). Calculated as the ratio between total dose in treatment period and days of treatment.

Table 6 Duration of individual olsalazlne dosages administered in 31 patients with ulcerative colitis and 9 patients with Crohn 's colitis

Olsalazlne dose/day

i g 2 g 3 g 4 g

Ulcerative colitis Numbers on dose 12 29 21 22 Days on dose 165+147 116+107 99+50 133*68

(14-395) (14-381) (32-194) (49-244)

Crohn's colitis Numbers on dose 3 8 5 4

Days on dose 2181160 102+119 133+108 172±73

(39-349) (14-371) (31-295) (78-257)

Values of 'days on dose' represent mean ±SD (range).

Table 7 Concentra ions of olsalassine, 5-ASA. and Λc-5-ΛSA in faecal dialysates of patients with ulcerative colit is : averages in each patient for each dosage of olsalazine

Olsalazine dose/day

Olsalazlne ιmnol/1 0.10*0.14 0.3410.40 0.4010.50 0.27+0.41

(<0.006-0.72) (<0.006-1.7) (<0.006-2.0) (<0.006-1.4)

5-ΛSA mmol/1 5.714.9 0.016.1 13111 23U2

(<1.0-10) (<1.0-31) (<1.0-37) (6.7-55)

Ac-5-ASA mmol/1 lltθ.5 9.519.0 7.51 ^Q .2 1115.6

(<0.7-26) (<0.7-60) (<0.7-26) (2.0-26)

Total 5-ASA mmol/1 1719.0 10+14 20±18 34il7

(<1.7-40) (<1.7-72) (<l.7-63) (0.6-75)

5- AS - 5-aini i.osali.c_.yl lc ac Lrt; Ac-5-ASA - acetyl-5-aιulnosalicylic acid . Values represent meanstSO (range)

Table 0 Mean concentrations of olsaϊas-lr.e in serum and 5-ASΛ and Ac-5-ΛSΛ in fecal dialysates in patients with active (C _j -C^ ) versus inactive

(C _β ulcerative colitis on different doses of olsalazlne

Olsalazlne dose/day

Disease activity I 0 2 g 3 g α

tie rum i

Olsalazlne mmol/1 C-^Cg 0.010(n=5) Q.0l6(n=22) 0.021(n=l6) 0.026(n=l6)

^J 0 0.009(n=7) 0.017(n=6) 0.022(n 0) 0.024(n=10)

Fecal diaJysate:

5-ΛSA mmol/1 C- _^ -Cg 6.0 (n=4) 7.0 (n=10) 12 (n=12) 22 (n=12)

C ₍₎ 4.3 (n=5) 9.5 (n=5) 17 (n=3) 25 (n=9)

Λc-5-ASA mmol/1 C^-Cg ^6, ° (» ⁼⁴ > ^Q - ^~ (n ^* *l ^β ) 7.2 (n=12) 9.7 (n=12)

C _Q 12 (n=5) 16 (n=5) 11 (n=3) 12 (n=9)

5-ΛSA " 5-u inosalicylic acid; Ac-5-ΛSA = auetyl-5-aminosallcylic acid,

Table 9 Concentrations of olsalazlne, 5-ΛSΛ, and Λc-5-ΛSΛ in serum of patients with ulcerative colitis: average in each patient for each dosage of olsalazlne

Olsalazlne dose/day

Olsalazlne umol/1 9.213.4 16+6.0 22*t0.1 25111

(5.4-15) (3.0-32) (7.9-39) (4.5-40)

5-ΛSΛ umol/1 0.510.2 1.110.9 2.4H.6 3.212.4

(<0.3-1.0) (<0.3-3.0) (<0.3-6.0) (<0.3-0.0)

Λc-5-ΛSΛ umol/1 2.3H.3 3.912.4 5.7±2.9 5.7+3.2

(<0.3-4.0) (<0.3-9.6) (l.B-14) (<0.3-12)

5-ΛSΛ = 5-aminosalicyllc acid; Λσ-5-ΛSΛ = acetyl-5-aιnlnosalicylic acid. Values represent meanstSD (range).

Table 10 Concentrations of PGE„ in fecal and rectal dialysates

Fecal PGE, Hectal PGE, levels ⁸ levels

Patients Study day (ng/ml) (ng/ml)

Ulcerative colitis

(SAZ refractory, n=23)

(SAZ intolerant, n=0)

Crohn's colitis (n=9)

Values represent meaniSD (range). Normal range: <0.5 ng/ml. h. Normal range : <0.6 ng/ml . llStafcistically signi ficant ( ρ<0.05 ) .

Table II Selected characteristics of 23 patients with inactive ulcerative colitis

Recruited Randomized Non-randomized

(n=23) SAZ(n=-6) Placebo(r.»5) (n-12)

Age, median year (range) 4β(tβ-β0) 60(37-65) 48(26-72) 37(18-80)

Sex, male/female 11/12 4/2 2/3 5/7

Duration disease, median month (range) 72(10-400) 64(36-400) 108(24-192) 66(18-228)

! Duration remission, median month (range) 16(12-72) 15(12-24) 22(12-72) 12(12-72)

Duration of SAZ treatment, median month (range) 60(12-144) 48(20-120) 60(24-144) 67(12-132)

Previously treated with corticosteroids, no. 23 6 5 12

Smokerβ/non-smokerβ, no. 4/19 3/3 1/4 0/12 b Extension of disease , ^■ total'/left sided/ rectal, no. 6/10/7 0/4/2 3/1/1 3/5/4

Randomized because of increased fecal and/or rectal concentrations of PGE ⁽ Table 12 ⁾ . Based on previous barium «>noιna. See text for details.

Table 12 Fecal and rectal PGE levels, randomization, and outcome in 23 patients with inactive ulcerative colitis in whom sulfasulazlne Mas stopped at entry

Facal PGE /Rectal PGE levels (ng/ml) Outcome

DuratIon remission 6 months after (months) Entry

2 weeks 2 months 6 mo ths 12 months * code Remission Relapse (day) b e

12 0.34/0.41 (ND /13.3) ♦ (14) 12 0.1 /0.99 0.31/0.33 0.14/1.521. 0.22/0.26 SAZ 1 0.91/26.1 (0.70/7.32)" + (14) 72 0.73/3.26 I.24/3. I6R PL + (RH2 ⁾ 16 0.08/0.49 0.20/0.86 0.12/0.31 0.17/0.82 0.25/I.66R (0.30/2.08)' PL + (RH80) 17 0.08/0.20 0.14/0,15 0,10/0.46 0.42/0.71 + (295) 54 0.20/0.41 0.14/0.49 0.03/0.19 0.28/0.67 0.46/0.19 22 0.26/0.30 0.16/0.23 0.47/0.76 3.32/I.54R (8.94/10.3) PL + (R+IS9) 48 0.12/0.56 0.10/0.24 0.37/0.88 t (162) 24 0.08/0.13 0.12/0.58 0.27/0.84 0.45/0.30 O.I4/2.09R 0.20/0,64 SAZ 72 0.09/0,12 0.05/0.10 0.24/0.29 0.12/0,30 (0.19/9.89)" + (358) 12 0.24/1.26 0.24/0.46 (0.19/23.1)' ♦ (56) 12 HO/0.47 0.23/0.31 0.30/0.9S + (81) 18 O.Oβ/2.1 0.32/2.40H 0.38/0.59 SAZ 12 0.50/1.39 0.58/2.56R 0.57/0.23 PL 12 0.20/0.55 0.32/0.24 (9.6I/N0) + (30) 24 1.00/3.59 0.99/12. IR 0.43/2.11 SAZ 12 0.37/0.50 0.42/0.22 0.34/0.31 ♦ (156) 12 0.33/0.60 0.40/0.32 0.35/0.40 0.68/I.2IR 0. I4/HO SAZ 12 0.36_ |.β9 0.38/2.49R 0.32/0.30 SAZ 24 0.22/0.46 0.18/0.34 0.16/0.35 0.30/0.27 0.27/0.55 24 0.60/3.47 2.36/I.42R (5.38/1.24) PL + (RH80) 12 0.65/2,21 . 0.30/0.99 0.31/0.99 + (ISO)

_.. . randomization (I.e.. In case of focal PGE >0.5 ng/ml and/or rectal PGE >I.O ng/ml). NO, not done. SAZ. sulfasalazlne. PL, placebo, "Values in squared brackets indicate that endoscopic relapse was present at the study day. If a relapse occurred the patient was withdrawn.