Menstrual cycle, tolerance and blood alcohol level discrimination ability

0306-4603/84$3.00 + .OO

AddictiveBehaviors, Vol. 9, pp. 67-77, 1984

Printed

Copyright

in the USA. All rights reserved.

0 1984 Pergamon

Press Ltd

MENSTRUAL CYCLE, TOLERANCE AND BLOOD ALCOHOL LEVEL DISCRIMINATION ABILITY WILLIAM

M. HAY, PETER E. NATHAN, HARRY and WILLIAM FRANKENSTEIN

Alcohol

Behavior

Research

Laboratory,

Rutgers,

W. HEERMANS,

The State University

Abstract-Relationships between variations in blood alcohol level (BAL) and blood alcohol level discrimination accuracy were investigated as a function of menstrual cycle, hormonal variations, and behavioral tolerance to alcohol in 20 female social drinkers. All subjects consumed a moderate dose of alcohol on three occasions during one complete menstrual cycle. Subjects estimated their level of intoxication eight times during each drinking session. Each subject’s behavioral tolerance to alcohol each session was assessed by a body sway procedure. Subjects also completed detailed questionnaires on symptoms related to menstrual period and familial history of alcohol use. Contrary to earlier reports, (1) no difference in total time of intoxication or in other indices of ethanol metabolism was found between women (N = 9) who were not taking birth control pills and those (N = 11) who were and (2) no difference was found in peak BAL as a function of menstrual cycle phase. Stage of menstrual cycle did not affect the accuracy of BAL estimation. A trend approaching significance suggested that, with increasing BAL, women in the oral contraceptive group were more accurate in their estimates of intoxication. A significant interaction between menstrual cycle phase and tolerance levels was found; high tolerant women were significantly less accurate than low tolerant women in estimating BAL during the midcycle phase of the menstrual cycle.

Research has suggested a relationship between variations in women’s responsiveness to alcohol as a function of hormonal level and stage in menstrual cycle. It has been suggested that women reach higher blood alcohol levels (BALs), absorb alcohol faster, and reach peak blood alcohol levels sooner than men, following equivalent dosages of alcohol (Dubowski, 1976; Jones & Jones, 1976a). Women tested during the premenstrual phase of the menstrual cycle have been reported to attain significantly higher peak blood alcohol levels, and to demonstrate faster absorption rates, than women tested during either the menstrual or intermenstrual phases (Jones & Jones, 1976a). Tests of three women on each day through a complete menstrual cycle indicated great variability in BAL following a constant alcohol dose; peak BALs occurred during premenstruum and about the time of ovulation (Jones & Jones, 1976b). Jones and Jones (1976a) also investigated possible relationships between levels of the sex steroids estrogen and progesterone, and metabolic variability in women. Ethanol metabolism in 11 women taking oral contraceptives and 11 controls was contrasted. The two groups of women achieved almost identical peak blood alcohol levels after a standard dose of alcohol (0.66 ml/kg). Women in the oral contraceptive group, however, were reported to have metabolized ethanol significantly slower and, as a consequence, to have remained intoxicated longer than the controls. No significant differences between the two groups in rate of ethanol metabolism as a function of menstrual cycle phase were found. Zeiner and Kegg (1980) also compared alcohol metabolism rates in women taking Preparation of this paper was made possible by Grant AA00259-13 from the National Institute on Alcohol and Alcohol Abuse and Grant 28-9178 from the Charles and Johanna Busch Foundation, Rutgers University. Requests for reprints should be sent to Peter E. Nathan, Ph.D., Alcohol Behavior Research Laboratory, Building 3530 Busch Campus, Rutgers, The State University, New Brunswick, NJ 08903. 67

68

WILLIAM

M. HAY

et al.

oral contraceptives (N = 10) and those (N = 10) not taking them. Both groups were tested the first day of menstrual flow (when concentrations of estrogen and progesterone are low), and on day 24 of the cycle, prior to the onset of the next menstrual period (when concentrations of estrogen and progesterone are high). The oral contraceptives 10 of the subjects were taking artificially elevated estrogen and progesterone concentrations on Day 24 with respect both to levels on day 1 and to the levels of subjects not taking the contraceptives. Peak blood alcohol levels were higher and alcohol clearance rates faster on day 1 (when hormone levels were low) than on day 24 (when they were high). In addition, subjects not taking oral contraceptives reached higher blood alcohol levels than subjects taking oral contraceptives on day 24 but not on day 1. These results suggested that high levels of female sex steroids lowered peak blood alcohol levels and slowed down ethanol metabolism. Estrogen and progesterone levels are both low throughout menstrual flow (Days l-4). Estrogen peaks at the time of ovulation (Day 14), declines moderately and rises to a sustained level on about Days 22-26. Progesterone levels also peak during Days 22-26, with both estrogen and progesterone decreasing rapidly at the onset of the premenstrual phase. As Zeiner and Kegg (1980) have noted, this pattern of hormone variation does not support Jones and Jones (1976b), who reported that women tested during the premenstrual phase achieved higher peak BALs with more rapid absorption than during menstrual flow. Since levels of both estrogen and progesterone would be declining or at sustained low levels during each of these menstrual phases (Days l-5 and Days 26-28), hormonal level variations could not have had a major impact on alcohol metabolism or peak BAL achieved. These conflicting results make it impossible to draw firm conclusions on the relationship between the kinetics of alcohol metabolism and menstrual cycle-related fluctuations in hormonal levels. The fact that another recent study (Sutker, Allain, Brantley & Randall, 1981) failed to confirm another of Jones and Jones’ findings (1976a, 1976b) to the effect that women reach higher peak BALs and demonstrate faster alcohol absorption rates than men after equivalent doses of alcohol simply adds to this research area’s methodologic and conceptual disarray. In addition, no research to our knowledge has examined relationships among variations in hormone levels as a function of menstrual cycle phase, BAL discrimination ability, and tolerance in women. The present study, therefore, had three main objectives. The first, a replication and extension of prior work by others, was to compare women using oral birth control methods with women experiencing regular menstrual cycles in terms of peak BAL, total intoxication time, and latency to peak BAL. The second objective was to determine whether stage of menstrual cycle influenced ability to make accurate BAL estimates and whether oral contraceptive use influenced accuracy of BAL estimation. The final objective was to assess the possible interrelationship of ethanol tolerance (a documented determinant of BAL estimation accuracy), stage of menstrual cycle, and BAL estimation ability. METHOD

Subjects Subjects were 20 female social drinkers, 18-30 years of age (M = 20.2 years), recruited from advertisements placed in the Rutgers University student newspaper. Selection criteria included a history of moderate social drinking (i.e., a score between 16 and 23 (M = 18.95) on the Oates-McCoy scale (Oates & McCoy, 1973); the absence of medical or psychological contraindications to alcohol administration (Lawson,

Menstrual

cycle and blood

alcohol

69

level

Nathan & Lipscomb, 1980); and a history of regular, normal menstrual cycles. Menstrual history was assessed via a detailed menstrual history-birth control questionnaire adapted from Moos (1968), followed by discussion with a staff nurse with extensive gynecological experience. Eleven women using birth control pills and nine women who reported regular menstrual periods and were using an alternative birth control method met these criteria. Brand name and number of months the contraceptive was taken by the 11 subjects on oral contraceptives are listed in Table 1. Average length of the menstrual cycle for these subjects was 28 days; mean duration of menstrual flow was 4.73 days. Average length of the menstrual cycle for nine subjects not taking oral birth control medication was 28.9 days; mean duration of menstrual flow was 5.44 days. The two groups did not differ in education, age, drinking history, height, or weight. Subjects were instructed to report any deviation from projected menstrual cycle, delay in menses, or suspicion of pregnancy immediately. Procedure

All subjects came to the Alcohol Behavior Research Laboratory (Nathan, Goldman, Lisman 8~ Taylor, 1972) once during each of three phases of a full menstrual cycle: premenstrual phase (days 26-28); menstrual flow (days l-5); and mid-cycle (days 12-15). Appointments for each session were scheduled after each subject’s menstrual calendar was analyzed to determine the optimum timing of each session. Between sessions, subjects were asked to maintain usual drinking patterns. Subjects followed the same experimental schedule during each experimental session. They arrived at 8 a.m., after an S-hour food and liquid fast and after refraining from taking any drugs for 12 hours. Vital signs and an initial blood alcohol level reading (zero, in all cases) were recorded. BAL was monitored throughout the study via breath samples analyzed by an Intoximeter Mark IV gas chromatograph. A diary, in which subjects recorded the amount and type of alcohol consumed since the previous session, was also collected. Subjects also completed a menstrual symptom questionnaire adapted from Moos (1968), which detailed symptoms experienced during the current cycle phase. Each of the three experimental sessions was divided into ten 40-minute experimental periods. The experimental schedule is summarized in Table 2. Periods 1 and 2 were given over to instructions and two baseline tolerance assessments. Assessment of behavioral tolerance to alcohol was by a body sway measure previously described by Lipscomb and Nathan (1980). It is an extremely sensitive indicator of intoxication level Table 1. Oral contraceptives: brand name, number of months taken. Subject 1 2 3 4 5 6 I 8 9 10 11

Oral

Contraceptive

Lo Ovral OvCON - 50 Loestrin 1.5/30 Ovral 50 Lo Ovral Lo Ovral Loestrin Medicon Lo Ovral Brevicon Ortho-novum l/50

Months 10

17 24 16 5 30 1 30 24

70

WILLIAM

Table 2.

M. HAY et al.

Experimental

schedule. Period

1 Alcohol Administration and Dose Estimates Blood Alcohol Level Measurements and Estimates Body Sway Tolerance Measures

x

(40 minutes

each)

2

3

4

5

6

7

8

9

10

x

x

xxxxxx x x X

x x

x x

x x

x x

x x

and appears, as well, to be a valuable measure of behavioral tolerance to alcohol (Lipscomb & Nathan, 1980; Moskowitz, Daily & Henderson, 1974). The apparatus reflects anterior/posterior and lateral body sway of 0.64 cm or greater separately while subjects stand as motionless as possible for a l-minute period with eyes open. Alcohol was administered six times each session, at the beginning of Periods 3 to 8. Before each drink, subjects gargled with an anesthetic mouthwash to reduce taste cues. Each drink contained a mixture of 80-proof Vodka and tomato juice to total 6 oz (180 ml) of liquid. Three of the drinks contained 1.5 oz (45 ml) vodka, two contained 1.O oz (30 ml) vodka, and one contained .5 oz (15 ml) vodka. While the number of high, medium and low doses of alcohol remained constant across the three sessions, the order of their administration was randomized in sessions 2 and 3. Subjects were required to consume each drink within 6 minutes. Mean peak blood alcohol level was 103 mg/%, ranging between 77 and 130 mg%. Eight BAL measurements were taken during Periods 3-10: 25 minutes after each of the six drinks, and 65 and 105 minutes after the final drink. Immediately before each BAL measurement, subjects estimated level of intoxication. Subjects were instructed to estimate level of intoxication on a scale ranging from 0 (“cold sober”) to 50 (“very high - as high as you have ever been”). These estimates were transformed to a 50-point BAL scale, where 0 = 0 mg% and 50 = 150 mg%. On the completion of the session’s last task, subjects were served a meal. They remained in the laboratory until their BAL reached zero mg%. RESULTS

Analyses of variance revealed no differences within any of the three sessions between women taking birth control pills and women not taking birth control pills in latency from onset of drinking to peak BAL, peak BAL to zero BAL, total time intoxicated, and peak BAL. Additional analyses of variance indicated no differences between or within the two groups on these indices as a function of phase in menstrual cycle. Analyses of peak BAL and latency to peak BAL across both groups did not indicate a significant difference between premenstrual and other menstrual cycle phases.

BAL discrimination Analysis of variance of estimation accuracy scores-contraceptive use (2) x sessions (3) x periods (8)-indicated significant main effects for sessions, F(2,36) = 6.13, p < .Ol, and for periods, F(7,126) = 50.33, p < .OOl. The main effect for periods, consistent with prior observation of social drinkers (Huber, Karlin & Nathan, 1976), indicates that BAL discrimination accuracy deteriorates as absolute BAL increases. Newman-Keuls post hoc analyses, comparing the means of sessions 1, 2, and 3, were not significant. Mean error scores were 0.44, 0.58 and 0.40 respectively. Significant interaction effects were found between contraceptive use and periods, F(7,126) = 2.10, p < .05 and between sessions and periods, F(14,252) = 6.23,

Menstrual cycle and blood alcohol level

71

p < .OOl. Newman-Keuls post hoc analyses did not yield significant differences between the means of the contraceptive use (“Pill”-“No Pill”) x periods interaction. Figure 1 illustrated this nonsignificant trend: Women who did not use oral contraceptive methods made greater absolute errors in BAL estimation than pill users as BAL increased (BAL peaked at Period 6). BAL discrimination accuracy was also assessed each session. Analyses of variance showed significant main effects for periods all three sessions; this anticipated finding affirmed prior findings to the effect that blood alcohol level discrimination accuracy varies inversely with blood alcohol level. A significant interaction effect - contraceptive use x period (07,126) = 2.08, p < .05)-was found only in Session 1. NewmanKeuls analyses of this interaction were not significant. To determine whether menstrual cycle phase exerted a significant effect on the ability to render accurate BAL estimations, a contraceptive use (2) x cycle phase (3) x periods (8) analysis of variance was performed (using estimation error scores). No significant main effect was found for phase in cycle. Results showed a main effect for periods, F(7,126) = 50.33, p < .OOOl,consistent with earlier analyses, as well as a periods by cycle phase interaction, F(7,126) = 2.10, p < .05. Post hoc analyses indicated a consistent (but nonsignificant) trend for the group of women who did not use

.05

,+---e w

No Pill Pill

I/

&

/

g .04

\

\

\

\

I= 2 I= z go3 E E w w I= .02 0 z Q

.o I

.oo

L

I

2

3

4 Periods

5

6

7

8

Fig. 1. Absolute errors in BAL estimation for women taking birth control pills versus women not taking birth control pills across experimental periods.

WILLIAM M. HAY et al.

72

oral contraceptives to make greater errors in BAL estimation than the oral contraceptive group across periods, with more errors occurring as BAL increased. Separate analyses to assess BAL accuracy within each cycle phase were also done. Each analysis revealed a significant main effect for periods. A significant interaction effect -contraceptive use x periods, F(7,126) = 2.24, p .05-was noted only in the mid-cycle phase analysis. Newman-Keuls post hoc analyses did not reveal significant differences between the means of the contraceptive use x periods interaction. Figure 2 illustrates this nonsignificant trend: this anticipated finding affirmed prior findings to the effect that blood alcohol level discrimination accuracy varies inversely with blood alcohol level. Women not using oral contraceptives made greater absolute errors in BAL estimation than pill users during the mid-cycle phase.

Tolerance analyses Two groups of subjects were formed on the basis of a median split on subjects’ mean sway change scores with eyes open during periods 4, 6 and 7 of session 1 (Lipscomb & Nathan, 1980). Prior analyses had shown these three scores to be maximally sensitive to BAL. The 10 subjects who showed the greatest mean change in body sway from

Cf-----?

;;pill

4 t

I

I

2

I

3

4

5

6

7

8

Periods Fig. 2. Absolute errors in BAL estimation for women taking birth control pills versus women not taking birth control pills across experimental periods at the mid-cycle phase.

Menstrual

cycle and blood

alcohol

73

level

baseline (nondrinking) levels were considered low tolerant while those 10 subjects who showed the least change were defined for this purpose as high tolerant. The high tolerant group contained six oral contraceptive users and four others while the low tolerant group contained five oral contraceptives users and five others. Mean baseline sway scores were 21.7 for the high tolerant group, and 28.2 for the low tolerant group. High tolerant subjects showed a minimal change in mean sway during rising BALs (M sway change = 5.71; SD = 6.1) while low tolerant subjects swayed substantially more in this condition (M sway change = 27.7; SD = 11.6). Additional analyses of variance showed that high and low tolerance groups did not differ in actual BALs achieved in any session. An initial overall analysis of BAL discrimination accuracy using ratio error scores - tolerance level (2) x sessions (3) x periods (8) - revealed significant main effects for sessions, F(2,36) = 6.15, p < .005, and for periods, F (7,126) = 45.40, p < .005, and a periods x sessions interaction, F (14,252) = 6.60, p < .OOl. No main effect for initial tolerance level across sessions and periods was found. NewmanKeuls post hoc analyses were non-significant. BAL discrimination accuracy was also assessed for each session. These analyses indicated a significant main effect for periods

.037c ,036 ,035 -

@--------0

High Tolrrancr Low Tolrranco

,034 .033,032 gE .031z

.030-

2 .029m E .029 ‘=: .027f

.026-

9

.025.024.023.022f

.oo ’

1

2

I PREMENSTRUAL

Fig. 3. function

Absolute error in BAL estimation of time in cycle.

3

MENSTRUAL FLOW

for high tolerant

MIDCYCLE

versus low tolerant

women

as a

WILLIAM M. HAY et al.

74

(p < .OOl) in all three sessions. A significant interaction effect-periods x tolerance level-was also found in Session 3, F(7,126) = 2.59, p < .Ol. Newman-Kerns comparisons did not reveal differences between the means of this interaction. When phase in menstrual cycle was considered-cycle phase (3) x tolerance level (2) x periods (8)-a main effect for periods, F(7,126) = 45.40, p < .OOl, and a significant interaction effect for cycle phase x tolerance F(2,36) = 3.78, p < .Ol, were found. Figure 3 illustrates this significant interaction. Figure 3 suggests that high tolerant women demonstrated much more variability in BAL estimation ability than low tolerant subjects. Newman-Keuls post hoc analyses indicated that low tolerant subjects were significantly more accurate in estimating BAL than high tolerant subjects during the midcycle phase of the menstrual cycle (p < .05). The effects of initial tolerance level on BAL discrimination accuracy were also assessed for each phase in the menstrual cycle. Analyses of variance showed, as expected, a significant main effect for periods (p < .Ol) during all menstrual phases. A significant interaction effect-tolerance level x period-was found only at the midcycle phase, F(7,126) = 3.74, p < .OOl. This interaction is shown in Figure 4. Newman-Keuls post hoc analyses indicated significant differences (p < .05) between the mean BAL

.05 l -----4

High Tolrranl Law Tolerant

.04

.-6 ‘0 .-E f w _I .03 S z k W

0) 3 3 $

.02

.Ol

.oo Periods Fig. 4. Absolute error in BAL estimation for high tolerant versus low tolerant women during the mid-cycle phase.

Menstrual

cycle and blood alcohol

level

15

discrimination accuracy of the low and high tolerant groups during periods 4, 5, 6 and 7 (when blood alcohol levels were highest), with the low tolerant group consistently more accurate during these times. Self-reported symptoms

Each subject completed a detailed menstrual symptom questionnaire (Moos, 1968) during each phase of the menstrual cycle. The questionnaire listed 47 frequentlyreported menstrual symptoms. Subjects indicated their experience with a symptom during each menstrual phase on a scale ranging from (l), no experience with symptom, to (6), symptom acute or partially disabling. Analysis of variance-contraceptive use (2) x cycle phase (3)-indicated a main effect for cycle phase, F (2,36) = 7.50, p < .Ol. Subjects in both contraceptive use groups reported a higher frequency of menstrual distress at midcycle phase (M midcycle = 15.50) than at premenstrual (M premenstrual = 12.10) or menstrual flow phases (M menstrual flow = 8.80). DISCUSSION

Contrary to earlier reports (Jones & Jones, 1976a), no difference was found in the total time of intoxication between women taking oral contraceptives and women not taking oral contraceptives. In addition, no differences were found between these two groups of women on the latency from onset of drinking to peak BAL or elapsed time from peak BAL to zero BAL. The study also failed to confirm additional findings by Jones and Jones (1976a): when women were tested during the pre-menstrual phase of the menstrual cycle, they did not achieve higher peak BALs or absorb alcohol more rapidly than when tested during midcycle or menstrual flow phases. Consistent with Jones and Jones (1976b) but contrary to Zeiner and Kegg (1980), there were no differences in peak BAL or in rate of alcohol metabolism as a function of phase in menstrual cycle between women using oral contraceptives and those not using them. One logical explanation for these diverse findings is the diversity in alcohol administration procedures utilized in these studies. Although subjects in the Jones and Jones (1976b), Zeiner and Kegg (1981) and Sutker et al. (1982) studies all consumed the same dose of ethanol (.66 ml/kg/beverage alcohol), mixes used and alcohol-mixer ratio chosen varied across the studies. As well, consumption times were 5 minutes in the first two studies and 10 minutes in the third. In the present study, alcohol administration was more gradual, with subjects consuming six drinks containing 210 mg of beverage alcohol at 45 minute intervals across the experimental day and achieving higher means BALs than subjects in the other studies. There is some evidence to support this explanation. Jones and Vega (1973) have already documented the differential impact of drinking time on cognitive performance and elimination rate. A second explanation for these diverse findings is that the actual day(s) of the menstrual cycle on which subjects were tested varied across studies, as did the means by which investigators attempted to ensure the accuracy of subjects’ reports of menstrual cycle pattern and variability. In the present study, even though cycle test days appeared to be comparable to the days sampled by Jones and Jones (1976b), we failed to replicate their finding of differences in alcohol metabolism between the premenstrual and other phases of the menstrual cycle. Zeiner and Kegg (1981), however, tested pill and no pill subjects on Day 24 (when estrogen and progesterone levels were high) and Day 1 (when concentrations of both hormones were low). This testing procedure maximized the probability of significantly different hormone levels and resulted in findings that support the hormonal mediation hypothesis: high levels of female sex hormones are associated with lower peak blood alcohol levels and slower alcohol metabolism.

76

WILLIAM

M. HAY et al.

Since actual hormonal levels were not measured be used when interpreting the data.

in the studies,

however,

caution

must

We also found that menstrual cycle phase did not exert a significant effect on the ability of our subjects to make accurate blood alcohol level estimations. In addition, blood alcohol level (BAL) discrimination ability in women using oral contraceptives and women who were not using them did not differ (although, with increasing BAL, women who were not on oral birth control pills tended to larger errors in their estimation accuracy). Unlike drinkers studied earlier (Lipscomb & Nathan, 1980), initial tolerance levels alone did not impact significantly on BAL estimation accuracy in this sample of female drinkers. When menstrual phase was analyzed in this context, however, initial tolerance levels were found to exert significant effects on BAL discrimination ability: High tolerant female social drinkers exhibited more variability in BAL estimation ability than low tolerant females, with the effect that high tolerant women were significantly less accurate than low tolerant women in estimating BAL during the midcycle phase of the menstrual cycle. To summarize: Research to date suggests that hormonal fluctuations during the menstrual cycle can influence a woman’s metabolic responsiveness to alcohol. The precise mechanism for this effect remains unclear, in large part because methodologic differences among relevant studies obscure the effect. Besides the impact of differences in alcohol administration and test day selection and validation, the results of the present study suggest that differences in behavioral tolerance to ethanol mediate hormonal effects and affect BAL discrimination ability in women. Clarification of the interaction of behavioral tolerance and hormone levels awaits additional research with both female social drinkers and alcoholics, designed to measure through one or more menstrual cycles, ethanol metabolism, tolerance changes, and to assess actual hormonal levels to validate self-reported menstrual cycle data. This research has important implications for both the etiology and treatment of women alcoholics. Differential sensitivity to blood alcohol level across the menstrual cycle may well play a role in what some (e.g., Mendelson & Mello, 1979) have speculated may be differences in the likelihood of extended drinking episodes among women problem drinkers. To this end, one survey of a sample of alcoholic women indicated that 67% of menstruating women and 46% of nonmenstruating women related their drinking to menstrual cycle, with drinking typically beginning or increasing during the premenstruum (Belfer & Shader, 1976). REFERENCES Belfer, M.L., & Shader, R.K. Premenstrual factors as determinants of alcoholism in women. In M. Greenblatt & M.A. Schuckit (Eds.), A/coholismproblems in women and children. New York: Grune & Stratton, 1976. Dubowski, K.M. Human pharmacokinetics of ethanol, 1. Peak blood concentrations and elimination in male and female subject. Alcohol Technical Reports, 1976, 5, 55-72. Huber. H.. Karlin, R.. & Nathan. P.E. Blood alcohol level discrimination by nonalcoholics: The role of internal and external cues. Journal of Studies on Alcohol, 1976, 37, 27-39. Jones, B.M., & Jones, M.K. Alcohol effects in women during the menstrual cycle. Anna/s of New York York Academy of Science, 1976, 213, 567-587.(a) Jones, B.M., & Jones, M.K. Women and alcohol: Intoxication, metabolism and the menstrual cycle. In M. Greenblatt & M.A. Schuckit (Eds.), Alcoholism problems in women and children. New York: Grune & Stratton, 1976.(b) Jones, B.M., & Vega, A. Fast and slow drinkers, blood alcohol variables and cognitive performance. Quarterly Journal of Studies on Alcohol, 1973, 34, 779-806. Lawson, D.M., Nathan, P.E., & Lipscomb, T.R. Guidelines for the administration of alcohol to human subjects in behavioral research. Journal of Studieson Alcohol, 1980, 41, 871-888.

Menstrual

cycle and blood

alcohol

level

77

Lipscomb, T.R., & Nathan, P.E. Effect of family history of alcoholism, drinking pattern, and tolerance of blood ethanol level discrimination. Archives of General Psychiatry, 1980, 37, 576-596. Mendelson, J.H., & Mello, N.K. The diagnosis and treatment of alcoholism. New York: McGraw-Hill, 1979. Moos, R.H. The development of menstrual distress questionnaire. Psychosomatic Medicine, 1968, 30, 853-867. Moskowitz, H., Daily, J., & Henderson, R. Acute tolerance to behavioral impairment by alcohol in moderate and heavv drinkers. Reoort to the Highway Research Institute, National Highway Traffic Safety Administration. Washington, DC: Department of Transportation, 1974. Nathan, P.E., Goldman, M.S., Lisman, S.A., &Taylor, G. Alcohol and alcoholics: A behavioral approach. Transactions of New York Academy of Sciences, 1972, 34, 602-627. Oates, J.F., & McCoy, R.T. Laboratory evaluation of alcohol safety interlock systems: Instrument performance at high BALs. Report to the Highway Research Institute, National Highway Traffic Safety Administration. Washington, DC: Department of Transportation, 1973. Sutker, P.B., Allain, A.N., Brantley, P.J., & Randall, C.L. Acute alcohol intoxication, negative affect and autonomic arousal in women and men. Addictive Behaviors, 1982, 7, 17-25. Zeiner, A.R., & Kegg, P.S. Menstrual cycle and oral contraceptive effects on alcohol pharmacokinetics in Caucasian females. Alcoholism: Clinical and Experimental Research, 1980, 4, 233-237,