Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

There is accumulating evidence that cocaine abuse is associated with disruptions of menstrual cycle regularity as well as abnormalities of prolactin regulation (Cocores et al., 1986; Dackis and Gold, 1985; Gawin and Ellinwood, 1988; Mello, 1997; Mello and Mendelson, 1997; Mendelson et al., 1988, 1989; Teoh et al., 1994). Clinical studies indicate that women who abuse cocaine may have a variety of menstrual cycle disorders, including amenorrhea and luteal phase dysfunction, which compromise fertility (Mello, in press; Mendelson and Mello, in press; Smith and Smith, 1990; Smith et al., 1984; Teoh et al., 1994). However, it is difficult to attribute these disorders to cocaine abuse alone, because many cocaine abusers also abuse alcohol, opiates and marijuana, and each of these drugs has been shown to disrupt reproductive function (Braude and Ludford, 1984; Cicero, 1980; Mello and Mendelson, 1997; Mello et al., 1992b; Mendelson and Mello, in press; Smith and Smith, 1990; Teoh et al., 1992; 1994).

One advantage of animal models is that the effects of chronic cocaine exposure on the menstrual cycle can be studied under controlled conditions, without the confounding influence of polydrug abuse or related medical disorders. In rodent models, it has been consistently observed that chronic cocaine exposure disrupts the estrous cycle (King et al., 1990; King et al., 1993; Roberts et al., 1989). In a study designed to evaluate the effects of estrous cycle phase on cocaine self-administration, female rats developed irregular estrous cycles after 18 days of cocaine exposure (Roberts et al., 1989). Administration of cocaine (10 mg/kg/day s.c.) for 3 to 6 weeks also resulted in irregular estrous cycles characterized by repetitive days of estrus, absence of proestrus and prolonged periods of diestrus (King et al., 1990). Subsequently, this pattern of estrous cycle disruption was found to be cocaine dose-dependent (King et al., 1993). In a carefully designed study, the effects of 4 weeks of no treatment or saline control treatment were compared with those of four doses of cocaine. A group that had restricted access to food (matched with the high-dose cocaine group) was also included to control for cocaine's anorectic effects. Estrous cyclicity was significantly disrupted after 10 and 20 mg/kg/day of cocaine, but not after 1 or 5 mg/kg/day of cocaine, saline control treatment or food restriction (King et al., 1993). Ovulation was also significantly reduced in the high-dose cocaine groups, as inferred from oocyte retrieval after sacrifice (King et al., 1993). Over half the rats given 10 mg/kg/day of cocaine resumed normal estrous cycles once cocaine treatment was discontinued, but few rats that received 20 mg/kg/day of cocaine returned to normal estrous cycles over 5 to 6 weeks of observation (King et al., 1993).

Although the estrous cycle in rodents is a valuable model of reproductive function, in part because a complete estrous cycle occurs every 4 days (Freeman, 1988), there appear to be many differences between the neuroendocrine regulation of the estrous cycle in rodents and that of the menstrual cycle in women and higher primates (Knobil, 1974, 1980; Knobil and Hotchkiss, 1988; Yen, 1991). The rhesus monkey (Macaca mulatta) has long been a model of choice in reproductive biology because neuroendocrine control of the menstrual cycle is similar to that in women (Ferin et al., 1984; Goodman and Hodgen, 1983; Knobil, 1974, 1980). Moreover, the rhesus monkey model permits long-term evaluation of the effects of chronic cocaine exposure on menstrual cycle duration and hormonal indices of menstrual cycle adequacy. The reproductive lifespan of female rhesus monkeys lasts about 15 years, and 2 to 3 years of cocaine self-administration represents about 20 percent of the reproductive life and corresponds roughly to a period of 5 to 7 1/2 years in the human reproductive cycle.

In addition, rhesus monkeys can be trained to self-administer cocaine using operant behavioral procedures (for review see Mello and Negus, 1996). Cocaine self-administration procedures can be used to simulate naturalistic patterns of cocaine abuse reported by humans. Under these conditions, monkeys self-regulate the daily dose of cocaine and maintain stable cocaine self-administration patterns for months or years. Furthermore, there is now compelling evidence that drug self-administration procedures, in which animals control the frequency and amount of cocaine injected, have an important advantage over investigator-determined drug administration procedures (Dworkin et al., 1995). Response-independent cocaine administration resulted in higher rates of lethality in rats than self-administration of the same doses of cocaine (Dworkin et al., 1995). In previous studies, we used operant drug self-administration procedures to evaluate the effects of chronic alcohol exposure on the menstrual cycle in rhesus monkeys (Mello et al., 1983). We found that during chronic alcohol self-administration, rhesus females developed amenorrhea, anovulation and luteal phase dysfunction, conditions comparable to the menstrual cycle disorders reported in alcoholic women (Mello et al., 1983, 1989b, 1992b). To the best of our knowledge, there have not been any comparable studies of the effects of chronic cocaine exposure on the menstrual cycle in rhesus monkeys (see Mello, 1997; Mello and Mendelson, 1997 for review).

In the present study, we examined the effects of chronic cocaine self-administration on menstrual cycle duration and basal levels of gonadotropins and ovarian steroid hormones in formerly drug-naive rhesus monkeys. The effects of chronic cocaine exposure were compared with the effects of occasional cocaine exposure in an otherwise drug-free control group. We now report the effects of chronic cocaine self-administration on menstrual cycle duration and ovulation in eight female rhesus monkeys. In these prospective longitudinal studies, 48% of the menstrual cycles were of abnormal duration in the cocaine self-administration group, whereas only 6% of the menstrual cycles were abnormal in the control group. We observed recurrent episodes of amenorrhea, anovulation and luteal phase defects during cocaine self-administration and during cocaine withdrawal. These data indicate that chronic cocaine exposure can disrupt the menstrual cycle in rhesus monkeys and that menstrual cycle abnormalities often persist during cocaine withdrawal.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Subjects

Fourteen adult female rhesus monkeys (Macaca mulatta) (5.0-9.0 kg) lived in individual cages and were maintained at ad libitum weight. Eight females were subjects in studies of the effects of chronic cocaine self-administration on the menstrual cycle (Group 1). Six control females were drug-free except for exposure to single acute doses of cocaine (0.4 or 0.8 mg/kg i.v.) at intervals of 2 or more months (Group 2). The control monkeys were subjects in studies to evaluate the acute effects of cocaine on anterior pituitary hormones (Mello et al., 1990a and b, 1993a). Monkeys in Groups 1 and 2 lived in adjacent rooms and were maintained under identical conditions. A 12-hr light-dark cycle (lights on 7 A.M. to 7 P.M.) was in effect, and constant temperature and humidity levels were maintained. Monkeys were given multiple vitamins, fresh fruit and vegetables and Lab Diet Jumbo monkey biscuits (PMI Foods, Inc., St. Louis, MO.) to supplement a nutritionally fortified banana pellet diet. Water was continuously available.

Monkeys were obtained from a commercial supplier, and after quarantine, all monkeys were adapted to the laboratory for at least 6 months before these studies began. The onset and duration of menstrual bleeding was monitored daily with vaginal swabs. Monkeys were adapted to venipuncture procedures, and blood samples were collected for analysis of anterior pituitary and gonadal hormones on alternate days, starting at day 8 of the menstrual cycle and continuing until the onset of the next menstruation. Blood samples were collected at the same time each day between 1 and 2 P.M. Monkeys were periodically evaluated with laboratory tests to monitor the status of liver function, lipid and carbohydrate metabolism, electrolyte homeostasis and hematologic function.

Animal maintenance and research were conducted in accordance with the guidelines provided by the NIH Committee on Laboratory Animal Resources. The facility is licensed by the U.S. Department of Agriculture, and protocols were approved by the Institutional Animal Care and Use Committee. The health of the monkeys was monitored periodically by consultant veterinarians expert in primatology. Monkeys had visual, auditory and olfactory contact with other monkeys throughout the study. Environmental enrichment was provided by exposure to music, television and toys for at least 2 hr/day. Operant food and drug acquisition procedures described below provided an additional opportunity for environmental stimulation and enrichment (Line, 1987; Line et al., 1989).

Sequence of Conditions

Each monkey in Group 1 was studied as her own control before, during and after a period of chronic cocaine self-administration. During the drug-free base-line period, monkeys were trained to work at a simple operant task for food as described below, and the capacity for normal ovulatory menstrual cycles was evaluated. The basis for inferring normal ovulatory function was a mid-cycle LH surge and an elevation in progesterone levels during the mid-luteal phase. Subsequently, monkeys were surgically implanted with an i.v. catheter to permit cocaine self-administration. Details of the training and surgical procedures are described in the following sections.

Six of the 8 monkeys in Group 1 were exposed to two successive periods of cocaine self-administration and cocaine withdrawal. During the first period of cocaine availability, each monkey continued cocaine self-administration for as long as the i.v. catheter remained patent. Then the effects of cocaine abstinence on menstrual cycle duration were observed for an average of six menstrual cycles (range 2-10) or until menstrual cycle duration returned to normal length. Subsequently, another i.v. catheter was implanted, and monkeys were again given access to i.v. cocaine until the catheter was no longer patent. Then monkeys were observed during a second period of cocaine abstinence.

Rationale for Cocaine Self-Administration Procedures

Monkeys were allowed to control the frequency of cocaine injections and, consequently, the total dose of cocaine self-administered each day, within certain limits described below. This procedure was used instead of an investigator-determined cocaine administration regimen to simulate cocaine use patterns reported clinically and to minimize the possibility of toxic effects. Cocaine's potentially adverse effects on cardiovascular as well as cerebrovascular function in humans are well known (Cregler and Mark, 1986; Holman et al., 1991; Jacobs et al., 1989). The relative safety of drug self-administration procedures compared with non-response contingent methods of drug self-administration has recently been demonstrated (Dworkin et al., 1995). Cocaine was significantly more lethal in a rodent model when drug administration was response-independent than when it was controlled by the animal (Dworkin et al., 1995). This difference in lethality occurred despite the fact that all rats received identical amounts of cocaine at identical frequencies, but one group self-administered cocaine and a yoked control group received response-independent cocaine injections at the same time (Dworkin et al., 1995). Similarly, in rhesus monkeys, response-independent phencyclidine administration resulted in lethality, whereas self-administration of higher doses of phencyclidine did not (Johanson and Schuster, 1981). We have used the same cocaine and food self-administration procedures described here in a number of behavioral studies and have found that monkeys remained healthy and well nourished under these subject-controlled drug access conditions (Mello et al., 1989a, 1990c, 1992a, 1993b).

Operant Behavioral Procedures and Apparatus

Monkeys lived in a well-ventilated stainless steel chamber equipped with an operant panel, a banana pellet feeder and a water dispenser. Drug injections were delivered by a syringe pump in a single pulse that dispensed 0.1 ml of fluid over 0.9 sec. The operation of the syringe pump (Model 981210, Harvard Apparatus, Inc., South Natick, MA) was audible to the monkey. Schedules of reinforcement were programmed by custom-designed software and run on Apple II GS computers.

Monkeys worked at an operant task for food and for i.v. cocaine injections on a second-order schedule of reinforcement (FR2 [VR:16S]) that required an average of 32 responses for each food pellet or cocaine injection. Food availability and cocaine availability conditions were associated with different colored stimulus lights (S+) projected on a translucent Plexiglas response key (2-in. diameter) in the center of the operant panel. The key was dark during time-out periods when responses had no programmed consequences. When a food pellet or drug injection was delivered, the appropriate colored stimulus light (S+) (red or green) was illuminated for 1 sec on one of the three circles (3/4-in. diameter) located in a vertical column below the response key. Flashes of the 1-sec colored stimulus lights (S+) also signaled the completion of each successive VR component of a second-order schedule response requirement. When cocaine was not available because of catheter loss, the response key was dark except during food sessions.

Each experimental day consisted of four food availability and four drug availability sessions. Food sessions began at 11 A.M., 3 P.M., 7 P.M. and 7 A.M. each day, and drug sessions began 1 hr later at 12 noon, 4 P.M., 8 P.M. and 8 A.M. Consecutive food and drug sessions were separated by time-out periods 2 hr (1-3 P.M., 5-7 P.M. and 9-11 A.M.) or 10 hr in duration (9 P.M.-7 A.M.). The response key was dark during time-out periods, and responses had no programmed consequences. Each food or drug session lasted 1 hr or until 100 banana pellets (1 gm) or 20 cocaine injections (0.10 mg/kg/injection) were delivered. Cocaine injections were limited to 80 per day (8 mg/kg/day) to minimize the possibility of adverse drug effects.

Surgical Procedures

After operant performance for food was stable on the final schedule of reinforcement and there was evidence of normal ovulatory base-line menstrual cycles, each monkey was surgically implanted with an i.v. double lumen silicon catheter (I.D. 0.028 in., O.D. 0.080 in.) under aseptic conditions. Monkeys were sedated with ketamine (5 mg/kg s.c.), and anesthesia was induced with sodium thiopental (10 mg/kg i.v.). Atropine (0.05 mg/kg) was given to reduce salivation. After insertion of an intratracheal tube, a surgical level of anesthesia was maintained with halothane (1-1.5% in oxygen). Catheters were implanted in the jugular or femoral vein and exited in the mid-scapular region. After surgery, monkeys were given 200,000 units of Combiotic Dihydrostreptomycin and Penicillin G i.m. on alternate days for a total of 5 injections. The i.v. catheter was protected by a tether system consisting of a custom-fitted nylon vest connected to a flexible stainless steel cable and fluid swivel (Spaulding Medical Products, Birmingham, AL), which permitted monkeys to move freely. Catheter patency was maintained by i.v. cocaine administration and a saline flush. Fluid swivel and catheter patency were checked manually each day. A short-acting barbiturate, methohexital sodium (3 mg/kg i.v.), was used to evaluate catheter patency, if necessary. The catheter was considered to be patent if i.v. administration of methohexital produced a loss of righting within 10 sec of its administration.

Preparation of Drug Solution

Cocaine hydrochloride was obtained in crystalline form from the National Institute on Drug Abuse (NIDA). The purity was certified by Research Triangle to be greater than 98%. Cocaine was dissolved in Sterile Saline U.S.P. for injection, to make a stock solution at a concentration of 50 mg/ml. The solution was then filter-sterilized using a 0.22-micron Millipore filter and stored in sterile, pyrogen-free vials. Doses for cocaine self-administration were calculated on the basis of each monkey's weight so that a final dilution of the stock solution (with Sterile Saline for U.S.P. injection) resulted in a unit dose of 0.10 mg/kg/injection in a volume of 0.1 ml/injection.

Radioimmunoassay Procedures

LH. Plasma LH concentrations were determined in duplicate by a double-antibody radioimmunoassay procedure similar to that described by Midgley (1966), using materials prepared by Dr. W. Peckham and following his suggestions. Purified ceropithecus pituitary LH for radioiodination (WP-XV-117-3239), rabbit antiserum (WP-R13, pool D) to human choriogonadotropin and rhesus pituitary LH reference preparation (NICHD-rhLH, also known as WP-XV-20) were provided by the National Hormone and Pituitary Program, supported by the National Institute of Child Health and Human Development and the National Institute of Arthritis, Diabetes and Digestive and Kidney Diseases. Radioiodination was performed using the chloramine-T (Greenwood et al., 1963) with sodium iodide-125 purchased from DuPont New England Nuclear Products (Billerica, MA). Goat antirabbit gammaglobulin was obtained from Behring Diagnostics (San Diego, CA). Results are expressed in nanograms per milliliter in terms of the reference preparation. The assay sensitivity was 5.6 ng/ml. Intra- and interassay CVs were 4.6% and 13.9%, respectively.

Progesterone. Plasma progesterone concentrations were measured in duplicate by a direct double-antibody radioimmunoassay method using a kit purchased from ICN Biomedicals, Inc. (Costa Mesa, CA). Results are expressed in nanograms per milliliter in terms of the reference preparation. The assay sensitivity was 0.12 ng/ml. Intra- and interassay CVs were 6.6% and 7.5%, respectively.

Data Analysis

The cocaine self-administration monkeys (Group 1) and the control monkeys (Group 2) were compared with respect to menstrual cycle duration. In the cocaine self-administration monkeys, we examined the effect of dose and duration of cocaine exposure on menstrual cycle duration and the frequency of ovulatory and anovulatory cycles.

Menstrual cycle duration. The duration of the menstrual cycles during intervals of cocaine self-administration and cocaine withdrawal was compared with the average duration of precocaine base-line menstrual cycles for individual monkeys. Menstrual cycles were classified as short, long or of normal length relative to the average length of each monkey's base-line menstrual cycles. Menstrual cycles that were at least 3 days shorter than the base-line menstrual cycle average were defined as short cycles. Menstrual cycles that were at least 5 days longer than the base-line menstrual cycle average were defined as long cycles. Menstrual cycles that fell within this range relative to the base-line cycles were classified as of normal length. These classification criteria were more conservative than using one standard deviation from the base-line menstrual cycle average because standard deviations from base-line cycles ranged from 0.58 to 2.65 days. Moreover, there is a limit to how short a menstrual cycle can be, whereas amenorrheic menstrual cycles can last for months. Amenorrheic cycles were defined as 60 days or more without menstruation. The frequency of occurrence of normal, amenorrheic, long and short menstrual cycles was tabulated for each monkey and compared to precocaine base-line cycles. Menstrual cycle lengths (days) during cocaine self-administration and withdrawal (Group 1) were compared with menstrual cycle durations observed in the control monkeys (Group 2) with a Contingency Table analysis (Winer, 1971).

Endocrine characteristics of the menstrual cycle. The endocrine characteristics of the menstrual cycle were evaluated by measuring changes in progesterone through time. The following criteria were used to classify each menstrual cycle as ovulatory, anovulatory and/or with luteal phase dysfunction. Ovulation can occur in menstrual cycles of any duration, including near the end of an amenorrheic cycle.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Eight rhesus females self-administered cocaine for an average of 27 ± 3.6 menstrual cycles, and the experimental history of each monkey is summarized in table 1. The first exposure to cocaine lasted for 256 to 776 days, and monkeys self-administered an average of 3.51 to 6.89 mg/kg/day of cocaine. The average dose of cocaine self-administered by these rhesus monkeys was equivalent to or greater than cocaine use reported by human cocaine abusers. In clinical studies, cocaine abusers reported using about 2 g/week of i.v. cocaine, which is equal to 4.08 mg/kg/day in a 70-kg man or 5.71 mg/kg/day in a 50-kg woman (Mendelson et al., 1988, 1989). The second period of cocaine exposure lasted for 103 to 623 days, and monkeys self-administered significantly higher average doses of cocaine (P < .02) than during the first exposure (table 1). Body weight during base-line menstrual cycles did not differ significantly from body weight during the first or second exposure to cocaine (P = .1146). Although the average number of banana pellets self-administered varied between subjects, all monkeys consistently ate the supplemental fruit, vegetables and chow provided each day.

Menstrual cycle duration during the precocaine base-line. All females had two or more normal ovulatory menstrual cycles during the drug-free base-line conditions before chronic cocaine exposure (Group 1) or initiation of acute endocrine studies (Group 2). Base-line menstrual cycle duration did not differ significantly between the two groups. In the cocaine self-administration group, base-line menstrual cycles averaged 27.4 ± 0.80 days (range 23.5 ± 0.5 to 35.6 ± 1.2 days). In the control group, base-line menstrual cycle duration averaged 26.7 ± 0.57 (range 23.3 ± 0.32 to 30.3 ± 0.8 days).

Menstrual cycle duration during chronic cocaine self-administration. Figure 1 compares the distribution of menstrual cycles of different duration between the control group and the cocaine self-administration group during chronic cocaine exposure and cocaine withdrawal. The control group had significantly more menstrual cycles of normal duration than the cocaine self-administration group (P < .001). In the control group, 94% of the 155 cycles were of normal duration, and the number of cycles that were one standard deviation longer or shorter than the base-line cycles accounted for 5% and 1%, respectively. Moreover, the control group did not have any amenorrheic cycles (defined as 60 days or more without menses). In contrast, during cocaine self-administration, menstrual cycle duration was quite variable, and 48% of the 217 cycles were abnormally short, abnormally long or amenorrheic (fig. 1).

View larger version (41K): [in this window] [in a new window]   Fig. 1.   Distribution of menstrual cycle durations under control, cocaine self-administration and withdrawal conditions. Each panel shows the percent of the total number of menstrual cycles studied in the control group (N = 6) and in the cocaine group (N = 8) that were classified as 1) of normal duration compared with the length of the average base-line cycle (striped bar); 2) short, defined as at least 3 days less than each monkey's base-line cycle average (light gray bar); 3) long, defined as at least 5 days longer than each monkey's base-line cycle average (dark gray bar) or 4) amenorrheic, defined as more than 60 days without menstruation (black bar). The number of monkeys in each group and the number of menstrual cycles on which these calculations were based are shown on the abscissa.

Menstrual cycle duration after abrupt withdrawal from cocaine. Cocaine self-administration continued until the i.v. catheter became occluded, requiring implantation of a new catheter. The distribution of abnormal menstrual cycle durations during cocaine withdrawal was similar to that during cocaine self-administration (fig. 1) and also differed significantly from the distribution of menstrual cycle durations in normal controls (P < .001). During cocaine withdrawal, 44% of the 82 cycles were of abnormal duration, compared with the precocaine base-line. Amenorrheic cycles ranging from 61 to 190 days in duration accounted for 6% of the menstrual cycles during cocaine self-administration, whereas during cocaine withdrawal, 7% of the cycles were amenorrheic. The transition cycles, when catheter occlusion occurred part way through a menstrual cycle, often were associated with amenorrhea.

Menstrual cycle duration in individual monkeys. The effects of chronic cocaine self-administration and cocaine withdrawal on the duration of consecutive menstrual cycles in six individual monkeys are summarized in figures 2, 3, 4. The pattern of menstrual cycle disruptions varied within individuals across time, and there was no consistent tendency for cycle length to normalize during an episode of chronic cocaine exposure. Figure 2 shows changes in menstrual cycle duration and average cocaine self-administration (mg/kg/day) during consecutive menstrual cycles for two monkeys that had several abnormally short, as well as long, menstrual cycles during cocaine exposure. The top panel shows menstrual cycle duration and cocaine dose over 17 menstrual cycles (431 days) for monkey CH553. During the first menstrual cycle when cocaine was available, this monkey self-administered an average of 3.81 (± 0.09) mg/kg/day of cocaine, and her menstrual cycle was only 17 days long. Cocaine self-administration increased to 4.29 (± 0.43) and 6.14 (± 0.24) mg/kg/day during the second and third menstrual cycles and then decreased to 3.3 (± 0.52) mg/kg/day during menstrual cycle 4, which lasted only 19 days. Her seventh and eighth menstrual cycles were 16 and 18 days long, and she self-administered an average of 7.45 (± 0.23) and 5.93 (± 0.45) mg/kg/day of cocaine, respectively. In the ensuing months, her menstrual cycles varied from 21 to 39 days in length, and cocaine self-administration averaged between 4.66 (± 0.38) and 7.20 (± 0.23) mg/kg/day.

View larger version (41K): [in this window] [in a new window]   Fig. 2.   Menstrual cycle duration during cocaine self-administration and withdrawal in two individual monkeys (CH553 and CH548). Center panel: Consecutive menstrual cycles are shown on the left ordinate. The average duration of the base-line menstrual cycle length (days) ( ± S.E.) is shown at the top of each panel and indicated graphically as a single vertical line above B on the abscissa. Because the average duration of the base-line menstrual cycles differed for individual monkeys, changes in cycle lengths from a normalized base-line are displayed to facilitate comparisons between monkeys. The dotted vertical lines on each side of the base-line indicate one standard deviation from each monkey's average base-line menstrual cycle duration. The number of days that each successive menstrual cycle was shorter (left center panel) or longer (right center panel) than the average precocaine base-line cycle duration is shown on the abscissa. Menstrual cycles during cocaine self-administration are shown as open circles on a solid black area, and menstrual cycles during cocaine withdrawal are shown as open circles on a solid gray area. Right panel: The average amount of cocaine self-administered during each menstrual cycle is shown as a horizontal black bar ( ± S.E.), and the cocaine dose (mg/kg/day) is shown on the right abscissa.

View larger version (42K): [in this window] [in a new window]   Fig. 3.   Menstrual cycle duration during cocaine self-administration and withdrawal in two individual monkeys (12740 and CH712). Details of figure 3 are the same as described in the legend for figure 2.

View larger version (46K): [in this window] [in a new window]   Fig. 4.   Menstrual cycle duration during cocaine self-administration and withdrawal in two individual monkeys (13312 and CH577). Details of figure 4 are the same as described in the legend for figure 2.

Endocrine Characteristics of the Menstrual Cycle

Ovulation and luteal phase adequacy during cocaine self-administration and withdrawal. Ovulation was inferred from a mid-luteal phase increase in progesterone to 8.5 ng/ml or above (Filicori et al., 1984). After ovulation, a corpus luteum forms at the site of the ruptured oocyte and secretes increasing levels of progesterone during the early luteal phase. The mid-luteal phase elevation of progesterone levels usually lasts for more than 6 days (Knobil, 1980; Knobil and Hotchkiss, 1988), and this is a reliable indicator of the occurrence of ovulation. The periovulatory LH surge usually lasts for only 24 hr, and in the present study, collection of blood samples on alternate days did not permit consistent detection of peak LH levels at mid-cycle. In the control group, peak progesterone levels averaged 15.36 ± 0.95 ng/ml during menstrual cycles classified as ovulatory.

View larger version (38K): [in this window] [in a new window]   Fig. 5.   Frequency of ovulation in menstrual cycles of different length during cocaine self-administration and withdrawal. The left ordinate shows the percent of the total number of normal, short or long menstrual cycles that could be accurately classified as ovulatory or anovulatory on the basis of progesterone levels during the mid-luteal phase of the menstrual cycle. The abscissa of each panel shows the actual number of menstrual cycles in each category. The top panel shows the distribution of normal, short and long menstrual cycles classified as ovulatory or anovulatory during cocaine self-administration. The bottom panel presents the same information for menstrual cycles during cocaine withdrawal.

Cycle length and ovulation as a function of cocaine dose. The average dose of cocaine self-administered during the menstrual cycle did not reliably predict menstrual cycle length or the presence or absence of ovulation (fig. 6, top panel). Monkeys self-administered an average of 6.36 (± 0.19) mg/kg/day of cocaine during ovulatory cycles and an average of 6.35 (± 0.25) mg/kg of cocaine during anovulatory cycles (fig. 6, top, column 1). The amount of cocaine self-administered during the menstrual cycle that immediately preceded an anovulatory cycle also did not predict ovulation or anovulation (fig. 6, lower panel). Monkeys self-administered an average of 5.75 (± 0.27) mg/kg/day of cocaine during the menstrual cycles that preceded ovulatory cycles and 5.78 (± 0.42) mg/kg/day during the menstrual cycles that preceded anovulatory cycles (fig. 6, lower panel, column 1). Cocaine dose during the same cycle and during the immediately preceding cycle was also analyzed as a function of menstrual cycle duration (fig. 6, columns 2, 3 and 4). There were no statistically significant differences between the average cocaine dose self-administered during ovulatory and anovulatory menstrual cycles regardless of cycle duration (fig. 6).

View larger version (46K): [in this window] [in a new window]   Fig. 6.   The effect of cocaine dose on ovulation. The average dose of cocaine self-administered during the menstrual cycles classified as ovulatory (open bars) and anovulatory (gray bars) is shown in the top panel. The average dose of cocaine self-administered during the immediately preceding menstrual cycle is shown in the lower panel. The average cocaine dose self-administered during menstrual cycles of all lengths is shown at the left of each panel. The average cocaine dose (mg/kg/day) is shown on the left ordinate. Menstrual cycle length (short, long and normal) is shown on the abscissa.

Amenorrhea during cocaine self-administration and withdrawal. Amenorrheic cycles occurred with approximately equal frequency during cocaine self-administration and withdrawal (see fig. 1). Sixteen of the 19 amenorrheic cycles appeared to be anovulatory, and peak progesterone values averaged 2.31 (± 0.37) ng/ml (range 0.69-4.24 ng/ml). Only three amenorrheic cycles appeared to be ovulatory, and each of the ovulatory cycles occurred during cocaine self-administration. Ovulation was inferred from peak progesterone levels of 14, 14.7 and 15.8 ng/ml, which were measured 5 to 8 days before the end of the amenorrheic cycle. Figure 7 shows elevations in progesterone during the three ovulatory amenorrheic cycles that lasted for 63 to 117 days. The antecedent LH surge was not detected, perhaps because of the relative infrequency of blood sample collection. Monkey CH712 self-administered 6.7 (± 0.18) and 7.9 (± 0.19) mg/kg/day of cocaine during these amenorrheic cycles, and monkey 996B self-administered 5.58 (± 0.19) mg/kg/day of cocaine. At the beginning of these amenorrheic cycles, monkey CH712 had self-administered cocaine for a total of 138 and 830 days, respectively, and monkey 996B had self-administered cocaine for a total of 291 days.

View larger version (24K): [in this window] [in a new window]   Fig. 7.   Patterns of LH and progesterone in amenorrheic cycles during cocaine self-administration. LH levels (ng/ml) are shown on the left ordinate, and progesterone levels (ng/ml) are shown on the right ordinate. The final days of each amenorrheic cycle are shown on the abscissa of each panel. The total duration (days) of the amenorrheic cycle and the monkey number are shown at the upper left of each panel. The sustained elevations in progesterone are interpreted as evidence of prior ovulation, even though sampling frequency did not permit detection of an LH surge.

View larger version (22K): [in this window] [in a new window]   Fig. 8.   Month of onset and duration of amenorrheic cycles during cocaine self-administration and withdrawal. The month in which each amenorrheic menstrual cycle began is shown on the abscissa. The left ordinate shows the duration of each amenorrheic cycle in days. Amenorrheic cycles during cocaine self-administration are shown as closed circles, and amenorrheic cycles during cocaine withdrawal are shown as open circles. The open squares superimposed on the circles indicate elevated progesterone levels consistent with ovulation as shown in figure 8.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Chronic cocaine self-administration and cocaine withdrawal were both associated with severe disruptions of menstrual cycle regularity compared with control conditions (fig. 1), and there was considerable variability in the pattern of menstrual cycle abnormalities observed within and between individuals (figs. 2, 3, 4). Ovulatory menstrual cycles of normal length were interspersed with anovulatory menstrual cycles and cycles of abnormal duration (figs. 2 and 3). Abnormally short cycles with low progesterone levels suggestive of luteal phase dysfunction were frequently observed (table 2; figs. 5 and 6). However, 2 of 8 monkeys were relatively resistant to the disruptive effects of cocaine (fig. 4). Chronic cocaine exposure, as well as the abrupt discontinuation of cocaine availability, was sometimes associated with amenorrhea 61 to 190 days in duration (figs. 7 and 8). The persistent menstrual cycle abnormalities observed during cocaine withdrawal suggest that cocaine's disruptive effects on the menstrual cycle do not reflect acute intoxication alone. Rather, daily cocaine exposure may induce long-lasting disruptions of the neuroendocrine regulation of the menstrual cycle. Persistent effects of cocaine after chronic exposure have been observed in several models (Mello and Mendelson, 1997). In human cocaine abusers, hyperprolactinemia may occur during cocaine abstinence as well as during occasional cocaine use (Cocores et al., 1986; Mendelson et al., 1988). In rats, estrous cycles did not return to normal after high doses of cocaine over 5 to 6 weeks of observation (King et al., 1993).

Daily cocaine exposure had variable effects on the menstrual cycle in these rhesus monkeys, and similar variations are reported in humans. Not all cocaine-dependent women report irregular menstrual cycles, and many cocaine abusers become pregnant, as evidenced by recent concerns over the teratogenic effects of cocaine abuse (Hutchings, 1989; Mayes et al., 1992). However, because many human cocaine abusers are also polydrug abusers, it is impossible to attribute any menstrual cycle disorders or impairment of fertility to cocaine alone (Mello and Mendelson, 1997; Smith and Smith, 1990; Teoh et al., 1994). Cocaine-induced menstrual cycle disruptions in rhesus monkeys are consistent with reports that chronic exposure to high doses of cocaine (10-20 mg/kg/day) for up to 6 weeks disrupted estrous cyclicity in rats and decreased rates of ovulation (King et al., 1990, 1993).

Factors Contributing to Menstrual Cycle Abnormalities

The effect of cocaine dose and duration of exposure on the menstrual cycle. The cocaine dose level was not well correlated with the extent or type of menstrual cycle disruptions observed in individual monkeys. Some monkeys self-administered high daily doses of cocaine with minimal toxic effects (fig. 4), whereas in other monkeys, menstrual cycles were disrupted by both low and high levels of cocaine self-administration (figs. 2 and 3). These findings suggest that there may be a threshold for dose and duration of cocaine exposure that is sufficient to disrupt the menstrual cycle but that this threshold varies across individual monkeys. Analysis of group data also indicated that the average dose of cocaine self-administered did not reliably predict either menstrual cycle duration or anovulation (fig. 6). Because events during one menstrual cycle may influence the subsequent menstrual cycle, we also examined the average dose of cocaine self-administered during the cycles immediately before ovulatory and anovulatory cycles of different lengths. However, the dose of cocaine self-administered did not differ significantly between menstrual cycles that preceded ovulatory and anovulatory menstrual cycles (fig. 6).

Effects of other factors on menstrual cycle abnormalities. Because a number of factors other than chronic drug exposure can influence menstrual cycle duration and ovulation, it is important to consider the possibility that the abnormalities observed in the present study might not be attributable to cocaine alone. For example, malnutrition, concurrent medical conditions, excessive exercise, and stress may result in amenorrhea in women (Bullen et al., 1985; Frisch, 1982; McArthur et al., 1980; Sherman, 1984; Warren, 1992). In rhesus monkeys, seasonal factors may also contribute to anovulatory menstrual cycles under some conditions (Hartman, 1932; Walker et al., 1983). However, on the basis of the following considerations, we conclude that chronic cocaine self-administration, and not other, uncontrolled factors, was primarily responsible for the menstrual cycle disruptions observed.

Possible mechanisms underlying cocaine's effects on the menstrual cycle. The ways in which cocaine disrupts neuroendocrine regulation of the menstrual cycle are poorly understood (Mello, in press; Mello and Mendelson, 1997; Teoh et al., 1994). Moreover, analysis of these disorders is complicated by the fact that each clinically defined syndrome may result from hormonal disruptions that occurred earlier in the same menstrual cycle or in the previous menstrual cycle. Acute administration of cocaine changes basal levels of anterior pituitary hormones that are important for menstrual cycle normalcy. For example, acute administration of cocaine stimulates the release of LH, FSH and ACTH and suppresses prolactin in rhesus monkeys (Mello et al., 1990a and b; Sarnyai et al., 1996) as well as humans (Heesch et al., 1996; Mendelson et al., 1989; Teoh et al., 1994). Frequent repetition of these acute hormonal effects of cocaine might contribute to the menstrual cycle abnormalities observed during chronic cocaine self-administration. Data relating elevations in gonadotropins and ACTH and changes in prolactin levels to regulation of the menstrual cycle are summarized below.

Gonadotropin and Ovarian Steroid Hormones. Abnormally high levels of LH and/or estradiol during the follicular phase may suppress FSH and delay follicle maturation and subsequent ovulation (Dierschke et al., 1985, 1987; Zeleznik, 1981). Acute cocaine administration increases LH (Mello et al., 1990a and b, 1993a), so it is possible that recurrent LH stimulation during chronic cocaine exposure disrupts follicle development. Although FSH is only one determinant of normal folliculogenesis, adequate FSH levels are necessary for follicle development and maturation of the preovulatory follicle (Goodman and Hodgen, 1983). Suppression of FSH during the follicular phase also may result in luteal phase dysfunction after timely ovulation. High estrogen levels during the early luteal phase may lead to premature regression of the corpus luteum, resulting in a short menstrual cycle (Hutchison et al., 1987). Cocaine's interactions with progesterone remain to be determined, but suppression of progesterone could contribute to the luteal phase defects observed. In clinical endocrinology, criteria for the differential diagnosis and pathogenesis of luteal phase dysfunction remain controversial (McNeely and Soules, 1988; Stouffer, 1990).

ACTH. It is also possible that the acute stimulatory effects of cocaine on ACTH and, by inference, on CRH contribute to the menstrual cycle disorders observed during chronic cocaine self-administration. It is well established that CRF administration to rhesus monkeys suppresses the release of LH and FSH, which are essential for normal ovulation (Olster and Ferin, 1987; Xiao and Ferin, 1988). Thus repeated stimulation of ACTH during chronic cocaine self-administration may contribute to anovulation as well as amenorrhea. The effects of CRF on the hypothalamic-pituitary-adrenal axis in rodent models have recently been reviewed (Rivest and Rivier, 1995).

Prolactin. Abnormally low or high prolactin levels may also be associated with luteal phase dysfunction (McNeely and Soules, 1988). Hyperprolactinemia is sometimes associated with chronic cocaine abuse as well as with cocaine abstinence (Cocores et al., 1986; Dackis and Gold, 1985; Mello et al., 1994; Mendelson et al., 1989). Hyperprolactinemia also may be associated with amenorrhea, but both conditions can occur independently (Buchanan and Tredway, 1979; Mello et al., 1988; Tolis, 1980). Cocaine-related hyperprolactinemia would seem to be inconsistent with cocaine's acute effects on prolactin. Single doses of cocaine decrease prolactin levels in rhesus males and females (Mello et al., 1990a; 1993a) and cocaine-naive men (Heesch et al., 1996), presumably as a result of increasing dopamine levels. Prolactin is under inhibitory dopaminergic control (Ben-Jonathan, 1985; Neill et al., 1981; Yen 1979, 1991), and cocaine acts as an indirect dopamine agonist by binding to the dopamine transporter and blocking the reuptake of dopamine (Kuhar et al., 1988; Ritz et al., 1987). It has been suggested that chronic cocaine exposure may lead to a "down-regulation" of dopamine receptors that impairs the sensitive regulatory feedback relationship between hypothalamic dopamine and prolactin to result in hyperprolactinemia (Dackis and Gold, 1985; Wyatt et al., 1988).