Lack of Absorption of Didanosine after Rectal Administration in Human Immunodeficiency Virus-Infected Patients

Abstract

The feasibility of rectal administration of didanosine (DDI) was studied in six human immunodeficiency virus-infected patients. After oral intake of a DDI solution (100 mg/m² of body surface area) combined with an antacid (Maalox), pharmacokinetic parametric values were in accordance with previously published data; the mean ± standard deviation for terminal half-life was 59.5 ± 15.0 min, that for peak concentration was 5.2 ± 3.9 μmol/liter, and that for the area under the time-concentration curve (AUC) was 494 ± 412 min · μmol/liter. After rectal administration of a similarly prepared DDI solution (100 mg/m² of body surface area), plasma DDI levels were below the detection limit (0.1 μmol/liter) at all time points in five of the six patients, and in the remaining patient the AUC after rectal application was only 5% of that after oral administration. We conclude that oral administration of DDI cannot be easily replaced by rectal application.

Didanosine (DDI) is an important component of antiretroviral therapy for human immunodeficiency virus (HIV)-infected children (7) and adults (6). At present only an oral formulation is available. DDI is rapidly degraded by gastric juice (8, 13). To ensure sufficient oral absorption it has to be administered together with an appropriate buffer to a patient with an empty stomach. Despite these arrangements, the average bioavailability is reported to be low (<40% [13]) and highly variable (2 to 89% [1]). In addition, drug intake is inconvenient and the neutralization of the gastric juice may interfere with the absorption of other drugs (14). Therefore, in order to avoid these disadvantages it is desirable to evaluate other routes of administration. In HIV-infected patients we have previously demonstrated a substantial absorption of zidovudine after rectal administration (16). In an animal model rectal application of DDI resulted in a mean bioavailability of about 15% (2). The aim of the present study was to evaluate the pharmacokinetics of DDI after rectal administration in HIV-infected patients.

Patients and methods.

Six HIV-infected male outpatients under treatment with DDI (100 to 120 mg/m² of body surface [BS] twice daily) were studied after their informed consent was obtained. The mean age of the patients was 30 years (range, 7 to 46 years), mean weight was 55 kg (range, 25 to 66 kg), and mean BS was 1.56 m² (range, 0.9 to 1.8 m²). The CD4 cell count ranged from 105 to 393 cells/μl (mean, 220 cells/μl). Disease stages according to the Centers for Disease Control classification (5) were as follows: for one patient, A1; for two patients, A2; and for three patients, C3. At the time of this study all patients had a hemoglobin level of >80 g/liter, neutrophil count of >1,000/μl, no evidence or prior history of pancreatitis or neuropathy, and no apparent opportunistic infections and did not manifest diarrhea or rectal ulcers.

Pharmacokinetic investigations were done in the morning. Patients fasted overnight and DDI treatment was withheld for the last 12 h. For drug administration DDI (Videx powder for children) was dissolved in distilled water and mixed with an equal volume of magnesium hydroxide–aluminum hydroxide (Maalox) to a final concentration of 10 mg/ml. This solution was administered at a dose of 100 mg/m² of BS for both parts of the study (after oral administration and rectal application). For rectal application DDI was given via a lubricated nasogastric tube inserted about 8 to 10 cm proximal to the external anus. Then, the tube was flushed with air. There was no evidence of leakage of the DDI solution during the study period of 4 h. Blood samples were collected by a peripheral indwelling catheter before and 30, 60, 90, 120, 150, 180, and 240 min after drug administration. The investigations after rectal application were done 1 week subsequent to those after oral administration.

Drug assay.

Concentrations of DDI in plasma were determined by high-performance liquid chromatography (10). In brief, 0.5 ml of plasma was extracted with methanol on a 500-mg Si-C₁₈ column (Bond Elut; ICT, Frankfurt, Germany). Samples were run isocratically on a 250 mm by 4.6 mm (height by outside diameter) Supersphere endcapped 5-μm RP-18 column (Merck, Darmstadt, Germany). The mobile phase consisted of a 50 mM potassium hydrogen phosphate solution, methanol, and triethanolamine (85/15/0.05 [vol/vol/vol]) adjusted to pH 4.0 with phosphoric acid. Concentrations of DDI were calculated by measuring peak height and referring to external standards and an internal standard (didehydrothymidine [D4T]). Retention times for DDI and D4T were about 5.35 and 6.40 min, respectively. The lower limit of detection (defined as threefold signal/noise ratio) was 0.1 μmol/liter. The method yielded linear results over the concentration range up to 50 μmol/liter (correlation coefficient [r²] was 0.9985, x-intercept was 0.15, and y-intercept −0.03). Intra- and interassay coefficients of variation were 4.9 and 6.7%, respectively.

All chemicals used were of analytical grade or better and purchased from Merck. Pure DDI was a kind gift of the Bristol-Myers Squibb Company, Syracuse, N.Y.

Pharmacokinetic analysis.

Maximum concentration (C_max), time to maximum concentration (T_max), terminal elimination half-life (t_1/2β), and area under the time-concentration curve (AUC) were calculated by a noncompartmental model. In detail, C_max and T_max were determined by visual inspection of the time concentration curve, and t_1/2β was determined by least-squares regression analysis. The AUC_0–∞ was calculated by using the linear-trapezoidal rule and extrapolated to infinity (C_last/β) with approximation of the last data point. For calculations the pharmacokinetic software package Topfit 2.0 was used (9).

Results.

After oral application, the mean ± standard deviation for C_max was 5.2 ± 3.9 μmol/liter, the mean T_max was 55.0 ± 23.0 min, and that for t_1/2β was 59.5 ± 15.0 min. The mean AUC was 494 ± 412 min · μmol/l (Table 1). Rectal application of the DDI solution was well tolerated by all patients, who showed no signs of bowel irritation or stimulation of defecation. After rectal application, in five of the six patients DDI concentrations were below the lower limit of detection at all time points. In one patient small amounts of DDI were detected (0.43, 0.16, and 0.14 μmol/liter after 30, 60, and 90 min, respectively). The corresponding AUC after rectal application in this patient was 5.3% of that after oral administration.

TABLE 1.

Pharmacokinetic parametric values for DDI after oral administration

Parameter (unit)	Mean ± SD (n = 6)	Range (n = 6)
Cmax (μmol/liter)	5.2 ± 3.9	2.6–12.0
Tmax (min)	55.0 ± 23.0	30.0–90.0
t1/2β (min)	59.5 ± 15.0	30.5–72.5
AUC (min · μmol/liter)	494 ± 412	200–1,313

Discussion.

In this study we attempted to compare the pharmacokinetics of DDI after oral and rectal administration. After oral application the pharmacokinetic parametric values in all patients were in good agreement with previously published data (1, 8, 11). Surprisingly, after rectal application no measurable absorption of DDI was detected in five of the six patients. In the sixth patient only very low levels of DDI were found in comparison to those observed after oral administration. Overall, this indicates a nearly complete presystemic elimination of DDI when it is applied rectally in humans.

In rats, however, rather similar bioavailabilities were demonstrated after oral (16% [15]) and rectal (15% [2]) administration. In the latter investigation DDI was infused in a manner such that the drug was deposited in the rectum and colon. The authors were able to show a 39-fold higher absorption rate in the rectum as compared to the colon. In our study, we applied DDI in a way that should lead to drug deposition exclusively in the rectum. Nevertheless, we did not detect any significant absorption of DDI.

In the present study, a single dose of 100 mg/m² (equivalent to approximately 2.5 to 5 mg/kg of body weight), which is within the range of a standard dose of a therapeutic regimen, was applied. In contrast, in rats Bramer et al. (2) administered a dose of 200 mg/kg of body weight. Even if the body surface rule is taken into consideration, this represents a strikingly higher dose. Assuming similar rectal bioavailabilities in humans and in rats (approximately 15%) we would have expected an average AUC after rectal application of about half of that observed after oral administration (approximately 30 to 40%) in humans. However, rectal absorption may not necessarily depend on the dose in a linear manner. In rats, a significant presystemic loss of DDI after rectal application was observed. This was mainly due to degradation by enzymes in the “intestinal contents” and to a smaller extent to first-pass metabolism in the intestinal epithelium or in the liver (3). This presystemic loss may occur in humans as well. Thus, a standard “oral” dose applied rectally may be rapidly degraded before significant absorption occurs. In liver homogenates, metabolism of DDI was shown to be saturable (3), and this may be true for the degradation capacity of the intestinal contents, too. In the study in rats the high dose of DDI may have exceeded the saturability of the degrading enzymes. Furthermore, Bramer et al. (2) supposedly reduced these intestinal contents by extensive enemas prior to drug application. In addition, these enemas might have caused mucosal damage that facilitated absorption of DDI in the rats.

Finally, DDI was rectally applied in a solution containing Maalox, which possibly may have impaired the absorption of the drug. DDI has a pKa of 9.12 (4, 12). The DDI-Maalox solution is supposed to be neutral if instilled into the rectum and neutral to slightly acidic in the stomach. Thus, it is not likely that the pH difference between the stomach and the rectum may lead to markedly different ionization states of DDI. Furthermore, DDI is quite soluble at neutral pH (12) and therefore probably does not precipitate out of the Maalox solution in the rectal environment. Taken together, these observations suggest that it is not likely that rectal application of DDI together with a buffered solution accounts for the lack of absorption in humans.

We conclude that oral administration of DDI cannot be easily replaced by rectal administration as suggested by the results of an animal study unless high-dose application or specific stabilizers are evaluated.