Introduction

Top Abstract Introduction Methods Results Discussion References

A^1,2 report from the Food and Drug Administration by Baum et al. (1986) noted that patients 60 years of age received 66% more prescriptions for sedative-hypnotic benzodiazepines than did patients 40 to 59 years of age. Unfortunately, there is also evidence that the elderly experience more adverse effects from these drugs (Thompson et al., 1983; Ray et al., 1989). Several groups of investigators have reported that the increased magnitude of pharmacological effects such as sedation and memory and psychomotor impairment observed in the elderly is associated with higher benzodiazepine plasma concentrations (Greenblatt et al., 1991; Nikaido et al., 1987; Pomara et al., 1984). This is supported by studies that demonstrate lower benzodiazepine clearance in the elderly, as reviewed by Greenblatt et al. (1989) and reported by Dehlin et al. (1991).

Numerous studies have evaluated the effect of aging on benzodiazepine response; most have defined elderly as 60 years. Castleden et al. (1977) reported that psychomotor effects of 10 mg of nitrazepam were greater than placebo in an elderly population but not in the young despite similar concentrations measured 12 and 36 hr after a single oral dose, suggesting an increase in sensitivity with age. Reidenberg et al. (1978) and Cook et al. (1984) reported that plasma concentrations of intravenous diazepam required for sedation in patients undergoing elective cardioversion and endoscopic procedures, respectively, were as much as 2- to 3-fold lower in the elderly than in young patients, with a significant inverse relationship between age and dose of diazepam. Other groups demonstrated that single-dose diazepam 2.5 mg (Pomara et al., 1985) or 10 mg (Swift et al., 1985) produced greater memory and psychomotor performance impairment and significantly higher concentrations of diazepam and desmethyldiazepam in the elderly than the young. Using logistic regression of the presence or absence of response to a verbal command, it was recently demonstrated that the elderly were more sensitive to the hypnotic effects of an intravenous dose of midazolam (Jacobs et al., 1995). Nikaido et al. (1990) assessed the effect of single oral doses of alprazolam and triazolam and found a more prolonged duration of psychomotor effects in the elderly than in the young. Triazolam was recently evaluated in young and elderly subjects using psychomotor performance and sedation measures after a single doses of 0.125 and 0.25 mg; results indicated that the greater impairment in the elderly than young subjects could be attributed to higher plasma concentrations rather than to sensitivity differences (Greenblatt et al., 1991).

From published data it is not possible to determine whether a receptor-based or other pharmacodynamic change occurs with age separate from the observed change in pharmacokinetics. Some studies did not include a young population for comparison (Pomara et al., 1984). Others did not control for potential confounding factors such as chronic diseases or drug interactions (Castleden et al., 1977; Cook et al., 1984; Reidenberg et al., 1978). In some cases, no concentration data (Bell et al., 1987; Nikaido et al., 1990) or very minimal concentration and assessment data were obtained (Castleden et al., 1977; Reidenberg et al., 1978; Cook et al., 1984). Controlling for variability due to gender (Ellinwood et al., 1984; Kroboth et al., 1985; McAuley et al., 1995) and race (Kalow et al., 1986) may also be important when determining the effects of age on drug sensitivity to limit variability due to factors other than age. None of the previous comparative studies assessed response relative to concentrations of drug to quantify sensitivity in the young and elderly.

Acute tolerance to the psychomotor effects of benzodiazepines such as diazepam (Ellinwood et al., 1985), triazolam (Kroboth et al., 1993), midazolam (Fleishaker et al., 1996) and alprazolam (Ellinwood et al., 1985; Kroboth et al., 1988) is known to occur in humans. Acute or rapid tolerance is defined as a shortened duration and decreased intensity of drug effects that occurs within hours after administration (Crabbe et al., 1979; Frey et al., 1986). The rate of development of acute tolerance to the psychomotor effects of triazolam (Kroboth et al., 1993) and alprazolam (Kroboth et al., 1988) has been quantified in young adult men. Tolerance is important to the assessment of sensitivity because acute tolerance shifts the effect-concentration curve to the right, causing an apparent decrease in sensitivity. Thus, a difference in rate of development of tolerance between young and elderly could account for a difference in apparent sensitivity.

This study was designed with two major objectives. The first was to determine whether age influences sensitivity to alprazolam (i.e., to determine whether response is greater in the elderly after taking into account concentration differences). The second objective was to determine whether age influences the effect offset rate (rate of acute tolerance development) of a benzodiazepine. Alprazolam was chosen because it is a widely prescribed intermediate-acting triazolobenzodiazepine, ranking ninth among all drugs in total prescriptions dispensed for 1993 (Simonsen, 1994). In addition, its availability in an intravenous formulation for experimental use and metabolic profile made it appropriate for use in this design. Alprazolam is oxidized to less active metabolites that are rapidly conjugated and appear to have an insignificant role in the pharmacological activity (Greenblatt and Wright, 1993; Smith et al., 1984).

	Methods

Top Abstract Introduction Methods Results Discussion References

Twenty-five young and 13 elderly nonsmoking, healthy white men gave written informed consent to participate in this study, which was approved by the University of Pittsburgh Biomedical Institutional Review Board. Women were excluded because the effects of progesterone on response to benzodiazepines (McAuley et al., 1995) may have confounded the effects of aging with changes in hormone concentrations. All men were screened by history, physical examination, laboratory, urine drug screen and blood alcohol concentration before participation. Subjects were excluded if they were taking any chronic medications (other than a multiple vitamin), had laboratory values that were abnormal (>10% out of normal range), had physical examination findings indicating the presence of a chronic disease, had a positive urine drug screen or blood alcohol level or had a history of psychiatric illness, drug or alcohol abuse or dependency. Subjects were also excluded if they had participated in a clinical drug study using central nervous system drugs within the previous 6 months.

Study design. This three-way crossover study was conducted in two parts. In part 1, all subjects received alprazolam as a rapid (2-min) intravenous infusion in an open-label, single-dose design. Four weeks later, subjects began part 2, which was a randomized two-way double-blind crossover of placebo or an individualized infusion of alprazolam. All treatment days were separated by 28 days. For part 2, the infusion regimen was individualized for each subject to target either a concentration that would produce a 30% psychomotor performance decrement (EC₃₀) predicted from the sigmoid E_max model in the bolus treatment or the maximum maintainable concentration with a dosage limit of 2 mg of alprazolam, whichever was lower.

Pharmacokinetic evaluation. Blood samples of 7 ml each were obtained in heparinized collection tubes from the opposite forearm catheter. In the bolus treatment, samples were obtained at time 0 (just before drug administration) and 5, 10, 20 and 40 min and 1, 1.5, 2, 3.5, 5, 7, 9, 12, 16 and 24 hr after the dose. Subjects returned to provide two additional samples at ~36 hr (range, 29.5-38.2 hr) and 48 hr (range, 45.6-55.6 hr); actual times were used in the pharmacokinetic analysis. An additional 5-ml sample was collected before alprazolam administration for serum protein binding determination. In part 2, blood samples were obtained at time 0 (just before initiation of the infusion), at 0.5 hr (end of the loading infusion) and at 1, 2.5, 4, 5.5, 7 and 9 hr. Samples were centrifuged, and plasma was harvested and frozen at 20°C until analysis.

Psychomotor performance, sedation and memory. Subjects practiced the battery of psychomotor tests on four occasions, including the evening before each of the three treatment days, to a plateau of performance defined as no improvement in score on two consecutive trials. Base-line performance was assessed after one practice session on the morning of each treatment. After the alprazolam bolus dose, testing sessions were administered nine times from 20 min to 12 hr. The RMT was administered six times: at 0 and 40 min and 2, 3.5, 5 and 9 hr. Sedation scores were obtained before each blood sample through 12 hr. In the continuous-infusion treatments, the battery of tests was administered at base line plus six times from 1 to 9 hr after initiation of the infusion.

Assays. Alprazolam plasma concentrations were determined by a validated capillary gas chromatographic method using electron capture detection with triazolam as the internal standard. This is a modification of a previously reported assay (Derry et al., 1995; Greenblatt et al., 1981). Intra-assay and interassay variability was 10% for concentrations of 0.25 to 16 ng/ml. Samples from 5 min to 12 hr were diluted with blank plasma to attain concentrations within the detectable range.

Pharmacokinetic analysis. Alprazolam plasma concentration-time data from the bolus treatment was analyzed by compartmental and noncompartmental methods using PCNONLIN Version 4.2. (1992). For the noncompartmental analysis,  (linear regression of terminal points), t_[1/2] (0.693/), AUC_0- (trapezoidal rule + last concentration/) and clearance (dose/AUC_0-) were determined for each subject. Vd was calculated as dose/*AUC_0-. AUMC_0- was determined by calculating the area of the concentration*time vs. time plot (linear trapezoidal rule + last concentration/²), and mean residence time was determined by [AUMC_0-/AUC_0-  (infusion time/2)]. Vd_ss was calculated by clearance*mean residence time. AUC_0-5hr, AUC_0-9hr and AUC_0-12hr were also determined using the linear trapezoidal rule (Yeh and Kwan, 1978). The 5-min concentration point for each subject was excluded from all analyses because it often was lower than the 10-min sample or was so high that it was physiologically unexplainable. A two-compartment model with micro-rate constants was also used to describe this data; results were used to individualize the continuous-infusion treatment.

Pharmacodynamic analysis. The CS score is the number of cards sorted per second; DSST score is the number of symbols correctly drawn in 90 sec. CS and DSST percent decrements were calculated by dividing the difference between the base-line and actual score by the base-line score and multiplying by 100, where the maximum possible decrement is 100%. CPT score is the average latency to press a control button (n = 40 Ss; maximum latency, 1500 msec). Percentage decrement was calculated with the following formula: In this equation, 1500 represents the maximum latency and 100% is the maximum attainable decrement. The RMT score is the correctly recognized items (maximum, 7); delayed recall score represents the pictures correctly recalled (maximum, 5).

Bolus treatment. Plots of effect-time and effect-concentration data were made for each subject. The MaxOE and t_MaxOE were determined for each test for each subject. AUEC values were calculated for each subject using percent decrement vs. time data from 0 to 9 hr for psychomotor tests and scores from 0 to 12 hr for NRSS. For RMT data, AUEC (from 0 to 5 hr) was calculated by using the difference between a perfect score area and the actual score area, which provides a measure of recognition decrement (Kroboth et al., 1995). The intervals for AUEC evaluation were chosen to include all subjects in the analysis. Effect ratios (AUEC/AUC) for each subject were also calculated to correct for interindividual variation in concentration during the time of pharmacodynamic evaluation.

Continuous-infusion treatment. The target psychomotor decrement for each subject was 30% decrease from baseline based on data from the bolus treatment. To determine the accuracy of prediction and quantify the rate of tolerance development to each psychomotor performance test, percent decrement was calculated for each time. The data for each subject for DSST, CS and CPT in the alprazolam treatment were plotted on a semilogarithmic scale. A tolerance rate constant or effect offset rate constant (k_t) was determined for individual subjects using natural log-linear regression of the data from 1 through 9 hr or through the time that the score returned to base line, whichever occurred first (Kroboth et al., 1993). The k_t is the slope of the resulting regression analysis; the corresponding half-life for offset of effect (t_[1/2]^t) is 0.693/k_t. Scores were considered to be at base line when they were within 10% of base line; these decrements are a conservative estimate of the observed variability in these tests. For any score 0% decrement from base line, a natural log value of 0 (equivalent to a 1% change from base line) was assigned to allow inclusion of that point in the regression. Regression analysis was performed on subject data that included a performance decrement of 25%, provided visual evidence of a decline in performance decrement over time and included at least three data points for the regression. A quadratic term was added to the regression to assess whether there was an improvement in model fit.

Statistical analysis. Pharmacokinetic parameter estimates were assessed for differences between groups using one-way analysis of variance. Repeated measures analysis of variance was used to assess differences in repeated assessments. Duncan's post hoc test was used to assess specific differences when indicated. Parameter estimates from the inhibitory sigmoid E_max model were compared using the nonparametric Mann-Whitney U test with the two-group t test approximation because of unequal variances. All analyses were performed using SAS Version 6.03 (1985). Differences were considered significant if P  .05.

	Results

Top Abstract Introduction Methods Results Discussion References

Twenty-five young and 13 elderly healthy men completed the alprazolam bolus treatment. Unless otherwise specified, results are based on these 38 subjects. The mean age was 27.5 years (range, 22-35 years) in the young and 68.2 years (range, 65-75 years) in the elderly; mean weight was 77.9 kg (range, 61.2-98.6 kg) for young and 80.2 kg (range, 62.2-106.3 kg) for elderly subjects. In the continuous-infusion treatments, there were 13 young and 13 elderly men. All had participated in the bolus treatment 28 days earlier.

Adverse effects were minor and required no intervention; effects included burning on injection (two), dizziness (one), nausea (one) and hiccups (seven). Drowsiness and sleep also occurred and is reported in the NRSS data. Pulse, respirations and blood pressure were monitored throughout the study day and remained stable in all subjects, with little deviation from prealprazolam values.

Pharmacokinetics. Figure 1 demonstrates mean concentration-time plots for the alprazolam bolus (fig. 1A) and infusion (fig. 1B) treatments for each age group. Table 1 summarizes mean pharmacokinetic parameter estimates for the young and elderly in the bolus treatment. AUC_0-12hr and C_max values illustrate that mean concentrations did not differ between groups over the 12 hr of effect assessments. However, data from the entire 48 hr of sampling (fig. 1A, insert) indicate that the elderly have a slower clearance and longer t_[1/2], which would result in increased steady-state concentrations and longer time to steady state during chronic treatment.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   A, Mean alprazolam concentration-time data for young () and elderly () men after a 2 mg/2 min i.v. dose through the time of last pharmacodynamic assessment at 12 hr and through 48 hr (insert). Error bars represent S.D. B, Concentrations at the end of the loading infusion (0.5 hr) and during the time of performance testing from 1 to 9 hr in the alprazolam continuous-infusion treatment in young () and elderly () subjects.

Pharmacodynamics in the bolus treatment. Figure 2A is a plot of DSST score vs. time in the bolus treatment. CS, CPT, NRSS and RMT data from the bolus treatment are summarized in table 2. Predose base-line (E_o) and MaxOE values for all psychomotor performance tests are presented in table 3 for young and elderly. Impairment is indicated by a lower score on CS (cards/sec), DSST (symbols) and RMT (pictures recognized); a higher score on CPT (latency in msec) indicates impairment. Nearly all (12 of 13) elderly but only 5 of 25 young had a performance decrement of 75% on at least one psychomotor test.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Mean DSST scores vs. time for the young () and elderly () in the bolus treatment (A), continuous-infusion treatment (B) and placebo treatment (C). Error bars represent S.D. *, For the one time point, n = 22 young (instead of 25) and n = 10 elderly (instead of 13).

View larger version (40K): [in this window] [in a new window]   Fig. 3.   Mean AUEC0-9 hrr/AUC-0-9hr for each psychomotor performance test in the young (filled columns) and elderly (shaded columns) group in the bolus treatment. Error bars represent S.D. Data from young and elderly groups were compared using the Mann-Whitney U Test. ** P < .01. *** P < .001.

Pharmacodynamics in the continuous-infusion treatments. Figure 2, B and C, shows DSST score vs. time data for the continuous-infusion treatments of alprazolam and placebo, respectively. Tables 6 and 7 summarize response data from the young and elderly men after the continuous-infusion and placebo treatments, respectively. Figure 2C demonstrates the stability of DSST scores during the placebo treatment day in young and elderly subjects; placebo CPT and CS data show similar stability (table 7). Repeated-measures analysis of variance revealed that there was no effect due to time or to a group-by-time interaction for any psychomotor test (P  .26); thus, time did not influence performance scores in either group during the placebo treatment. On the RMT, no subject had a score 4 of 7 in either age group at any time on the placebo day.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Mean percent recovery over time in the young () and elderly () subjects for DSST (n = 10 elderly and 11 young) (continuous infusion treatment) (A) and CS (n = 13 elderly and 13 young) Continuous infusion treatment) (B). Error bars represent S.D.

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Proportion of young (hatched bars) and elderly (filled bars) with significant RMT (A) and sedation (B) during alprazolam continuous infusion. Lines for the elderly (dashed lines) and young (unbroken lines) represent the corresponding logistic regression fit.

	Discussion

Top Abstract Introduction Methods Results Discussion References

The data from this study demonstrate that age influences both the pharmacokinetics and pharmacodynamics of alprazolam. In addition to decreasing alprazolam clearance, age increases sensitivity to the psychomotor, sedative and memory effects of alprazolam through a mechanism other than increased concentrations. Age also decreases the rate of offset of effect of the psychomotor, memory and sedative effects of alprazolam. Additionally, the data provide insight about the intensity of initial effect as a determinant of rate of tolerance development.

An important but unsurprising observation is the lower performance scores in the elderly in the absence of drug (base line and placebo). Hinrichs and Ghoneim (1987) reported that lower performance scores are evidence for homeostatic changes with aging present in the absence of drug. In the presence of alprazolam, our data show that the elderly experience greater impairment than young men: the elderly have lower absolute performance scores, greater absolute decrements in scores and greater percentage decrements than do young subjects at equivalent concentrations.

Pharmacokinetics and sensitivity. Although there is a lower clearance and a longer t_[1/2] in the elderly men, these differences do not explain the increased psychomotor effects observed after bolus alprazolam administration. Mean plasma alprazolam concentrations are similar in the young and elderly during the first 12 hr; in contrast, memory and psychomotor performance impairment as well as sedation are greater in the elderly. Furthermore, when response is corrected for individual variability in alprazolam concentration and for differences in base-line scores between young and elderly using AUEC/AUC, the elderly have higher ratios than the young; effect-concentration modeling with either the linear or the sigmoid E_max model shows a greater response through the range of alprazolam concentrations.

Pharmacokinetics and evidence for tolerance. Despite the maintenance of plateau concentrations of alprazolam during the continuous infusion, subjects demonstrated a gradual improvement in psychomotor performance and memory and an offset of sedation after 1 hr. This pattern of response can be explained by the development of acute functional tolerance.

Tolerance and aging. The data from this study suggest that the elderly develop acute tolerance to alprazolam effects more slowly than the young. Again, alprazolam concentrations do not explain the observations because there were no significant differences in concentrations between the young and elderly at any time point during the continuous infusion. This study was designed so that the young and elderly would achieve the same level of initial psychomotor decrement (MaxOE) during the alprazolam continuous-infusion treatment. This was achieved with DSST and CPT. Despite the fact that MaxOE for DSST was similar in young and in elderly men, the MnE_1-9hr was greater in the elderly (table 8). In addition, the elderly have a shallower slope of the regression line for offset of effect (k_t) and a longer t_[1/2]^t (table 8), and they recovered function more slowly than the young (fig. 4A). Furthermore, the only two subjects who did not show any evidence of offset of effect during the alprazolam continuous infusion were elderly. As stated earlier, the offset of CPT effect occurred too rapidly to allow mathematical evaluation (MnE_1-9hr <8% for both young and elderly men). CS results are discussed in the section on intensity of effect. The elderly also had a slower offset of effect of sedation and memory impairment (fig. 5).

Tolerance and intensity of effect. Two separate observations from the bolus and continuous-infusion treatments lead to the generation of a hypothesis that the greater the intensity of initial effect, the more rapidly is tolerance developed. First, in the bolus treatment, the elderly had a greater psychomotor decrement and a significantly steeper slope of the effect vs. concentration curve (slope and s, depending on model).

Conclusions and implications. Based on the data from this study, the elderly would be more impaired than the young during treatment with alprazolam for two reasons: (1) higher alprazolam concentrations during chronic treatment and (2) greater sensitivity coupled with slower offset of effect. To estimate the clinical impact of these differences during a treatment regimen of 0.5 mg of alprazolam orally every 6 hr, steady-state concentrations and corresponding psychomotor impairment were calculated for young and elderly. Predicted average steady-state concentrations (using mean clearance data and assuming 90% bioavailability) are 16.7 and 20.3 ng/ml in the young and elderly, respectively. The corresponding decrements at these concentrations (from the sigmoid E_max model and mean CS data) are 11.8% in the young and 28.7% in the elderly. In other words, although mean steady-state concentrations for the elderly would be ~1.2-fold higher than in the young, psychomotor performance decrement would be 2.4-fold higher. The estimate of psychomotor decrement does not take into account the tolerance development and therefore is more accurate for potential age-related differences at the initiation of therapy. The effect of aging on the therapeutic effect is uncertain, as is the presence of anxiety on psychomotor impairment. However, low doses and caution should be used while titrating response in the elderly.