BIOPHARMACEUTICS & DRUG DISPOSITION, VOL. 12, 233-246 (1991)

RELATIVE BIOAVAILABILITY OF OLSALAZINE FROM TABLETS AND CAPSULES: A DRUG TARGETED FOR LOCAL EFFECT IN THE COLON M. RYDE*$, B. HUITFELDT~AND R. PETTERSSON~ *Department of Human Pharmacology and 'Department of Biostatistics. Kabi Pharmacia Therapeutics AB, S75182 Uppsala, Sweden

ABSTRACT The aim of this investigation was to compare two formulations of the prodrug olsalazine (OLZ) with regard to local bioavailability of 5-aminosalicylic acid (5-ASA) in the colon. Since 5-ASA can not be measured directly in the colon, the bioavailability was evaluated by studying the plasma concentration and cumulative urinary excretion ( A e ) of its main metabolite N-acetyl-5-aminosalicylic acid (ac-5-ASA). The absorption of OLZ was also studied. A single dose of l g OLZ tablets and capsules was given to nine healthy fasting volunteers in two repeated two-period cross-over studies. Blood and urine samples were collected for 72 and 96 h, respectively. AUC, C,,, and Ae data from both studies were combined for statistical analysis. Ninety per cent confidence limits for differences in mean AUC for ac-5-ASA (tablet-capsule) compared to that of capsules were -0.31 per cent and 30.8 per cent. This indicates bioequivalence if a more relaxed criterion than the conventional f20 per cent is applied, which is justified in this situation. The 90 per cent confidence limits for C,,, were -10.5 per cent and 36.9 per cent while for Ae the values were -20.5 per cent and 23.7 per cent. Within and between subject variability estimates for AUC of ac-5-ASA were 24 per cent and 46 per cent, respectively. KEY WORDS

Olsalazine

Bioavailability

INTRODUCTION Sulfasalazine (salicylazosulfapyridine, SASP) has been used for treatment of ulcerative colitis for more than 40 years. SASP has a low systemic absorption and only 2-10 per cent is excreted in the urine. In the colon the azo bond in SASP will be reduced by bacterial azo reductase leading to a total conversion of SASP into 5-aminosalicylic acid (5-ASA) and sulfapyridine (SP)',*, the latter being recovered to about 80 per cent in the urine. In recent clinical studies local treatment with 5-ASA in comparison with SASP gave similar effects in proctitis and ulcerative colitiss5. The major action of SASP in the large intestine therefore appears to be associated with 5-ASA. Given orally, 5-ASA is rapidly and extensively absorbed from the stomach $ Address for correspondence.

0142-2782/91/030233-14$07.00 0 1991 by John Wiley & Sons, Ltd.

Received 29 October 1990 Accepted 19 December 1990

234

M . RYDE, B. HUITFELDT AND R . PETTERSSON

and the small intestine and most of the dose does not reach the large intestine. In order to make 5-ASA available in the large intestine without being prematurely absorbed, a new drug 5-5’-azodisalicylic acid (olsalazine; OLZ), was developed. In this compound two 5-ASA molecules are coupled with the same type of azo bond, which binds 5-ASA to sulfapyridine in SASP. OLZ was shown to be transported to the colon with minimal absorption6and to deliver 5-ASA efficiently to the proximal colon7. After formation in the colon, 5-ASA is assumed to have the same pharmacokinetic properties whether released from OLZ or SASP. In clinical studies OLZ has been shown to have a similar relapsepreventing effect in ulcerative colitis as SASPs. During the clinical evaluation of OLZ it was given as a powder in gelatine capsules. The recommended dose for maintenance therapy is 1 g (four capsules) twice a day. In order to reduce the number of units administered at each dosage occasion a tablet was developed containing twice the amount of OLZ compared to a capsule of equal size. The present investigation was performed in order to study the bioavailability of OLZ from the tablet in relation to the capsule. Since the two preparations are rapidly disintegrated (within 15 min in in vitro tests) no difference in availability was expected. Based on clinical experience the delivery of 5-ASA to the large intestine is essential for the therapeutic effect. However, there is no established method for quantitative assessment of 5-ASA in the large intestine. An indirect way to assess local bioequivalence would be to demonstrate the same systemic absorption of OLZ (bioavailability less than 2.3 per centlS) resulting from the same amount delivered to the colon. This procedure assumes that the systemic absorption and faecal excretion are the only two paths of elimination from the gastrointestinal tract. Another alternative would be to measure the plasma concentration and/or the urinary excretion of 5-ASA and its metabolites as a reflection of the amount of 5-ASA available in the large intestine. However, the absorption of 5-ASA in the large intestine is difficult to study since it is rapidly N-acetylated to acetyl-5-aminosalicylic acid (ac-5-ASA) at the following known sites: 1. in the colon by the bacterial flora9; 2. in the colonocytes lo; 3. in the liver as demonstrated after intravenous administration of 5-ASA”.

The systemic availability of ac-5-ASA from the large intestine is accordingly dependant on several processes which prolong and reduce the absorption12. This explains the relatively low plasma concentration of ac-5-ASA found in the studies where 5-ASA is delivered to the large intestinel-*J*J3-16. The fate of OLZ after an oral dose is schematically illustrated in Figure 1. In summary, we have a parent compound which is absorbed to a very low extent in the stomach and small intestine and is metabolized to 5-ASA in the colon, the site of therapeutic action. The amount of 5-ASA available in the

235

RELATIVE BIOAVAILABILITY OF OLSALAZINE

r

OLZ -0LZs

5-ASA

+ ac-5-ASA 1

I)

5-ASA

+ ac-5-ASA 1

*

5-ASA

+ I) ac-5-ASA

r OLZ ,OLZS

OLZ

5-ASA

5-ASA

ac-5-ASA

a c -5 -AS A

I

Figure 1. Schematic presentation of the fate of OLZ in the body

colon cannot be measured directly. This leaves us to assess the bioavailability indirectly by measuring the systemic concentration of 5-ASA and its main metabolite, ac-5-ASA. Consequently, the study of bioequivalence, when comparing two different preparations in this situation, is not well defined. For evaluation of the bioequivalence of the tablet and the capsule, we have chosen to study the plasma concentration of ac-5-ASA as the variable assumed to most closely reflect the situation at the site of action (Figure 1). As additional variables the urinary excretion of ac-5-ASA and OLZ and the plasma concentration of OLZ were evaluated. MATERIAL AND METHODS The first part of this investigation involved 10 healthy subjects in a randomized two-period crossover study (A) with a 1 month washout period between the treatments. No statistically significant differences were found between the two preparations but the results pointed to very large variability. Therefore, it was decided to repeat the study in the same subjects in order to get better evidence for the conclusion of bioequivalence (study B). In addition, this design now made it possible to investigate the within-individual variability. Study B was performed 6 months after study A using the same randomization scheme and drug batches, but with the treatment order reversed. One of the volunteers was unable to participate in study B due to his work situation and another volunteer by mistake received the preparations in the same order in the two studies. The volunteers were healthy according to medical history, physical examination, and clinical chemistry. The age of the volunteers ranged between 25 and

236

M. RYDE, B . HUITFELDT AND R. PETTERSSON

43 years, the weight from 68 to 79kg, and the height from 178 to 196cm. The volunteers were not taking any other medication. Products containing acetylsalicylic acid were not allowed during the 3 days before the start of the study. Single oral doses of 1 g (2.89 mmol) olsalazine sodium capsules (4 x 250 mg) and tablets (2 x 500 mg) were given with a standardized amount of water after a 10 h fast. Water, 100 ml, was allowed 1.5 and 3 h after the dosing. The subjects had lunch 4 h after the dose. Blood samples were drawn from a cubital vein at 0, 0.5, 1, 1-5, 2, 3, 6, 8, 12, 24, 48, and 72h after each dose into evacuated tubes containing heparin. Urine was collected in the intervals 0-24, 24-48, 48-72, and 72-96 h after the dose. The volume was recorded and an aliquot of lOml was kept frozen at -20" until analysis. The subjects were thoroughly informed about the importance of complete urine collection. The consistent appearance of OLZ and ac-5-ASA concentrations and the urine volumes were used as further verification. Adverse effects were elicited both by use of an open question and by individual reports on a special record form.

Analytical methods

Analysis of OLZ and ac-5-ASA in urine and plasma was based on the same LC-method as used in a previous studyI5. However, since it was not analyzed 5-ASA, no derivatization and no extraction procedures were needed, which improved the precision.

Pharmacokinetic evaluation

An indication of the extent of absorption of OLZ and ac-5-ASA from each of the two preparations was obtained from the areas under the plasma concentration-time curves (AUC), peak plasma concentrations (C,,,), and the cumulative urinary excretion data (Ae). AUC was calculated using the trapezoidal method.

Statistics

The relative bioavailability was evaluated by comparing the mean AUC, Cmax, and Ae for OLZ and ac-5-ASA from the two preparations, using the capsule as standard. AUC and Ae were found to follow a normal distribution approximately. The distribution of C,,, data was, however, skewed due to some extreme outliers. The statistical analysis of this parameter was therefore performed using the logarithmic transformation. Data for the subject who only participated in study A were not included in any of the statistical calculations.

RELATIVE BIOAVAILABILITY OF OLSALAZINE

237

The results were evaluated using analysis of variance for a two-period crossover design in which differences between subjects, periods, studies, and treatment sequences were accounted for in the modelI7. The treatment equality was evaluated in terms of 90 per cent confidence intervals for the mean difference between treatments based on the above method18. Treatment sequence was not found to be of statistical significance. This allowed us to include also the volunteer with the same treatment order in both studies in the statistical analysis for treatment differences. Accordingly, the statistical analysis was performed using data from nine subjects. The mean plasma concentration levels of OLZ for both capsules and tablets showed a clear difference between studies A and B, for which no obvious explanation was found. No statistically significant interaction between treatment and study was found. This justified the combined analysis of the two studies for the evaluation of treatment equivalence. The between-subject variability was calculated as the coefficient of variation (CV(%)) based on a combined estimate from the four experiments (study A: tablet, capsule; study B: tablet, capsule). CV (“A)between

=

J m i (EM):

C (SD)*= sum of squared standard deviations. CM = sum of mean values.

The within-subject variability was also expressed as CV (%) based on a combined estimate of the individual differences in the two studies. CV (YO) within =

[C(ID); + C(ID)&’6 (CM):

C(1D)X = sum of squared individual differences in series A. C(1D); = sum of squared individual differences in series B. The timing of blood sampling did not allow a meaningful estimate of the time for Cmx (Tm.J for the two compounds. Therefore T,,, was not evaluated in the statistical analysis. The statistical program SAS was used for the analysis of all data in this investigation. RESULTS The mean plasma concentration curves are shown in Figure 2. The observed C,,, of OLZ was reached within 30 min in some subjects but not later than 1 h in any of the subjects. C,,, values for OLZ in study A were 3-7 pM for the capsules and 4.7 pM for the tablets and the corresponding values in study B were 6.8 pM and 7-1 p. The mean urinary excretion of OLZ was very low, being 0.4 per cent for both capsules and tablets in study A and 0.7 per cent for the capsules and 0.5 per cent for the tablets in study B.

238

M . RYDE, B. HUITFELDT AND R. PETTERSSON

1

l6 14

6

I

I

9

12

15

I

I

I

18

21

24

Time (h) Figure 2. The mean and standard deviation (bars) of plasma concentrations of OLZ (-) and ac-5-ASA (----) in nine healthy male subjects. A 1 g oral dose of OLZ was given as capsules and tablets in two studies (A and B) with a two-period crossover design. (C, and C, capsules in study A and B, respectively; T, and T, tablets in study A and B, respectively)

The first appearance of ac-5-ASA in plasma was at 2-6 h, which coincides with the transit time through the small intestine found by Davis et ~ 7 1 . 'and ~ was an indication of the azo split in the proximal colon. The mean C,,, for ac-5-ASA in all experiments occurred at 8 h. The mean values were slightly higher in study A: 5.7 pM and 5.3 pM for the capsules and the tablets, respectively, compared to 4.4pM for both preparations in study B. The mean urinary excretion of ac-5-ASA was 17.0 per cent of the dose for the capsules and 21.0 per cent for the tablets in study A and 22.1 per cent and 18.7 per cent respectively in study B. These figures are in accordance with the results in other studies with OLZ.15 Individual and mean values, standard deviation and range for AUC, C,,, and Ae in studies A and B for both OLZ and ac-5-ASA are given in Tables

RELATIVE BIOAVAILABILITY OF OLSALAZINE

239

l(a) and l(b). In Figure 3(a) the AUC values for ac-5-ASA after intake of capsules and tablets in study A are plotted against those of study B. This figure illustrates that study A, on average, gave higher levels than study B, since almost all points are above the line of unity. The variability of the arrow directions illustrates the subject versus treatment interaction. The corresponding plot for the total urinary excretion is shown in Figure 3(b). Treatment differences (tablet-capsule in per cent of the capsule values) in studies A and B and the two studies combined are given as mean and standard error of the mean in Table 2. The 90 per cent confidence intervals from the combined studies are also included (Table 2 and Figure 4). The 90 per cent confidence intervals for the mean difference between treatments showed that the tablets did not exceed the capsules in availability with more than 30.8 per cent regarding AUC for ac-5-ASA, the primary bioequivalence variable. This variable has the shortest 90 per cent confidence interval but is skewed in relation to zero (Table 2 and Figure 4). A more central location of the confidence interval was found for the cumulative urinary excretion of ac-5-ASA. For this variable the bioavailability of the tablets did not differ from that of the capsules by more than -20.5 per cent to +23-7 per cent and the observed mean difference was only 1.6 per cent. The coefficients of variation of within-subject and between-subjectvariability of AUC, C,,,, and Ae of OLZ and ac-5-ASA are given in Table 3. The withinsubject variability of OLZ for these three variables lay between 40 and 50 per cent and for the between-subject variability between 58 and 85 per cent. The corresponding figures for ac-5-ASA were 24 to 40 per cent and 46 to 58 per cent, respectively. As expected the between-subject variability exceeded the within-subject variability. In a crossover study within-subject differences in plasma clearance can influence the comparison of AUC values. A correction for both renal and non-renal clearance was investigated as recommended by Upton et al.*O.However, none of these factors could diminish the large variability. Adverse effects

Three of the volunteers reported loose stools in all four experiments. One additional subject experienced loose stools only once when taking the capsules in study B. No other adverse drug effects occurred and the frequency of loose stools was thus similar with the two preparations. DISCUSSIONS AND CONCLUSIONS The plasma concentrations and urinary excretion of both the parent drug and its metabolites were of the same order of magnitude and pattern as found in earlier The mean serum concentration curves for both OLZ

9.1 12.4 15.0 28.0 19.8 12.9 7.3 10.7 11.1 11.1

13.1 6.0

1 2 3 4 t(5 6 7 8 9 10

Mean SD

Cmax

4.8 1.8

3.7 3.5 3.0 8.6 3.7 6.6 4.5 3.1 4.1 5.1

(pM)

Ae

12 7

6 7 7 28 6 10 9 7 13 18

(pmol)

t Not included in the mean value and SD.

AUC (pM*h)

Subject

Tablets

5.9 1.6 4.4 5.8 7.4 1.8 2.0 1.3 3.9 2.6 3.3 1.9

10.1 6.4

(pM)

Cma,

Capsules

17.2 9.5 13.6 22.4 33.3 6.0 4.8 3.1 8.0 6.1

AUC (pM*h)

Study A Ae

12 8

37) 13 25 4 6 8

7 15

23 3

(pmol)

5.2 15.2 3.0 3.9 16.5

17.5 34.6 8.3

18.3 9.7

30.0

7.1 5.2

-

-

15.6

3.6 2.7 1.4 6.8

9.2 6.9 19.3 23.1

AUC (pM*h)

Tablets C,,, (pM) Ae

14 13

15 46 7 10 4

-

9 6 12 14

(pmol)

16.4 11.4

11.5 35.9 13.9 7.0 14.0

-

11.1 6.3 36.0 11.8

AUC (pM*h)

Study B

6.8 6.6

5.4 22.5 2.2 3.7 5.4

-

19 17

z

g

F:

3 m

7r

9 21

>

s z w

r

3

2

X

11

19 57

-

13 2 31 9

m

3.9 1.9 12.0 4.5

"

Ae

(pmol)

Cmax