FERTILITY AND STERILITY Copyright 1995 American Society for Reproductive Medicine Printed on acid-free paper in U. S. A. Medroxyprogesterone increases basal temperature: a placebo-controlled crossover trial in postmenopausal women* Jerilynn C. Prior, M.D.tt Donald W. McKay, Ph.D.tll Yvette M. Vigna, B.A.t Susan I. Barr, Ph.D. Division of Endocrinology, Department of Medicine, University of British Columbia and Vancouver Hospital and Health Sciences Centre, Vancouver, British Columbia, Canada Objective: To assess whether temperature is increased by medroxyprogesterone (MPA) and thus whether basal temperature records could be used to determine ovulation during cyclic MPA therapy. Design: A 2-month double-blind placebo-controlled crossover trial in which oral basal temperature was measured daily. Setting: Normal human volunteers in an academic medical environment. Subjects: Eleven postmenopausal women not taking gonadal hormones. Intervention: Medroxyprogesterone acetate (10 mg/d) or placebo, calendar days 16 to 25, with crossover. Main Outcome Measures: Comparison of mean temperature days 17 to 26 during MPA versus placebo; comparison of differences between temperatures days 7 to 16 and 17 to 26 in MP A versus placebo months; and analysis for a significant monthly thermal shift. Results: The mean temperatures during MPA treatment averaged 0.27 C higher than during the placebo phase and showed a significant change from pretreatment to "treatment" phases during MPA but not during placebo cycles. Eight of the MPA and one of the placebo cycles showed a shift from lower to higher temperatures days 16 to 25. Conclusion: Medroxyprogesterone acetate has a physiological progesterone-like thermal effect. Therefore basal temperature data cannot reliably indicate ovulation during cyclic MPA administration. Fertil Steril 1995;63:1222-6 Key Words: Medroxyprogesterone acetate, progestin, basal temperature, progesterone For at least 100 years, the menstrual cycle has been known to affect core body temperature (1). Received June 17, 1994; revised and accepted December 20, 1994. * Supported by University of British Columbia, Vancouver, British Columbia, Canada, Medical Research Account 5-54590, and Upjohn Company of Canada, Toronto, Ontario, Canada. t Correspondence: Jerilynn C. Prior, M.D., Department ofmedicine, University of British Columbia, Suite 3318, 910 West 10th Avenue, Vancouver, British Columbia, Canada, V5Z 1M9 (FAX: 604-875-5925). :j: Division of Endocrinology, Department of Medicine, University of British Columbia and Vancouver Hospital and Health Sciences Centre. Family and Nutritional Sciences, University of British Columbia. II Present address: Faculty of Medicine, Memorial University of Newfoundland, St. John's, Newfoundland, Canada. 1222 Prior et a!. Thermogenic effect of MPA Menstrual cycle records of BBT measurements have been used for scientific studies for 30 years (2-4) and have aided fertility planning for at least three generations. Progesterone's effect in raising core temperature is a reproducible phenomenon, an action not shared by other classes of steroids. Injected progesterone has been shown to raise temperature in humans (5, 6). Medroxyprogesterone acetate (MPA; Provera R; Upjohn Company, Kalamazoo, MI and Toronto, Ontario, Canada), a C-21-derived P derivative, was formulated for use in the diagnostic assessment of amenorrhea (7), in the assessment (8) and therapy of dysfunctional uterine bleeding (9), and in menopausal combined hormone treatment (10). More recent evidence indicates that cyclic MPA administration increases spinal bone density in women with abnormal menstrual cycles or subclini-
cal ovulatory disturbances (11, 12), thus increasing the potential indications for cyclic MPA therapy. We have used prospectively a quantitative analysis of list-recorded morning temperatures to assess ovulatory function in women in several studies (12-15). The least squares method of basal temperature analysis has been validated through a study in which the computer-generated day of significant temperature shift from a lower follicular to a higher luteal phase temperature was shown to highly correlate with the day on which the serum LH peak occurred (16). We needed to ascertain with more certainty whether MPA could increase the basal temperature to determine whether we would be justified in using quantitative temperature methods to assess cycle types during a randomized blinded study of cyclic MPA versus placebo in women with menstrual disturbances (12). In addition, should MPA be shown to not be thermogenic, BBT data could be used in clinical practice to assess endogenous hormonal changes in MPA-treated women. For practical as well as scientific and theoretical reasons, therefore, it has become useful to know whether the synthetic progestin MPA alters temperature regulation and thus would interfere with the simultaneous use of basal temperature records. Many resources, including clinical papers (17) and pharmacology texts (18), do not mention whether progestins of different chemical formulas, especially those derived from the nonthermogenic T molecule, have temperature effects. Finally, we have been unable to locate a source that has assessed systematically whether MPA has a thermogenic effect. This study was designed to determine whether MPA causes a significant thermogenic effect in the dose of 10 mg/d that is equivalent in its endometrial effect to the mean luteal phase P levels. The temperature investigation was part of a broader controlled MPA treatment study in which the enrolled postmenopausal women prospectively recorded feelings and experiences on a Daily Menopause Diary (19). Design MATERIALS AND METHODS This is a 2-month double-blind, random-ordered, placebo-controlled crossover trial in which women took 10 mg/d MPA or placebo during days 16 to 25 of the month with crossover during the next month. The order oftherapy was determined by the research pharmacist from a random numbers table. The active and placebo medications were identical capsules filled with white powder. Oral basal temperature was measured daily for 2 months. All women were given a digital Celsius thermometer (no. 4009; Bec- ton Dickinson Canada Inc., Mississauga, Ontario, Canada) and were instructed to take their morning oral temperatures at the usual time of rising. Temperatures were recorded in a list at the bottom of the Daily Menopause Diary form. Illness, disturbed sleep, or changed time of temperature recording also were noted and temperatures were omitted from analysis if a temperature exceeded the preceding one by >0.3 C and coincided with such a notation. Monthly temperature data (that were truncated at day 26 for consistency of analysis) were analyzed using quantitative least squares statistics to determine the occurrence of a significant temperature shift (16). Women were examined, had height and weight measurements taken that were used to derive the body mass index (BMI; weight in kg divided by height in meters squared), and had a blood sample for E2 (and FSH in one woman who had undergone hysterectomy) as reported previously (19). Subjects Fourteen menopausal women who were 2': 1 year beyond their last menstrual flow and who were not taking gonadal hormones were recruited by word of mouth or by advertisements posted in community centers, on hospital bulletin boards, and in clinical centers. Potential subjects were interviewed for hormone use, general health, possible allergy to progestin, and last menstrual period. Each volunteer signed informed consent to a protocol approved by the University of British Columbia Committee on Research Involving Human Subjects. Statistical Analysis Women began taking MPA or placebo on calendar day 16 and continued the intervention through day 25. However, because the BBT was taken in the morning, the day 16 temperature was recorded before starting the intervention. Thus, for the purposes of analysis, temperatures recorded on days 17 to 26 of the month were used as the time on MPA or placebo. Data were analyzed in three different ways: [1] the mean temperature during the 10-day block of time on MPA was compared with the temperature during the time on placebo; [2] the difference between mean temperature during the 10 days preceding MPA and placebo (days 7 to 16) and during taking of tablets (days 17 to 26) was obtained and these differences were compared within women; and [3] each month's temperature record was analyzed using the previously published least squares method (16). An intervention was considered to cause a significant temperature shift if the paired t-test of dif- Prior et al. Thermogenic effect of MPA 1223
Table 1 Mean Oral Morning Basal Temperatures During 10 Days Before Treatment and the 10 Days of Treatment With MPA or Placebo* Subject Before MPA MPA Difference Before placebo Placebo Difference 1 36.30 36.53 0.23 36.43 36.33-0.10 2 36.33 36.57 0.24 36.56 36.36-0.20 3 36.62 36.59-0.03 36.48 36.50 0.02 4 36.38 36.83 0.45 36.47 36.46-0.01 5 36.35 36.58 0.23 36.34 36.40 0.06 6 36.36 36.72 0.36 36.01 36.06 0.05 7 36.22 36.63 0.41 36.24 36.26 0.02 8 36.46 36.86 0.40 36.54 36.60 0.06 9 36.49 36.63 0.14 36.41 36.41 0.00 10 36.21 36.31 0.10 36.25 36.25 0.00 11 36.33 36.77 0.44 36.37 36.43 0.06 Mean:!: SD 36.37:!: 0.12 36.64 :!: 0.15 0.27 :!: 0.16 36.37 :!: 0.16 36.37:!: 0.14-0.00 :!: 0.08 * Values are in degrees Celsius. ference between temperatures before and after shift resulted in a t value 2: 2.0, if the t value was a positive number (thus indicating a shift from lower to higher temperature), and ifthe temperature shift occurred during days 17 to 26 of the month. Nonparametric two-sided tests were used for all analyses. The significance level was set at 0.05. RESULTS Of 14 women enrolled, 11 completed the 2-month study. The three women for whom data were incomplete (related to home and family responsibilities and travel) did not differ in demographic characteristics from those completing the study. None dropped out because of adverse symptoms. One of the 11 women experienced withdrawal flow-she was the youngest participant (43 years old) and was just 1 year beyond her last menstrual period (19). The 11 participants had a mean age of 53.5 years (range 43 to 63 years) and had been postmenopausal for a mean of 6.5 years (range 1 to 14 years). Their mean height was 162.6 cm (range 153 to 174.5 cm) and mean weight was 67.7 kg (range 47.4 to 116.0 kg) with a BMI of 25.4 kg/m2 (range 19.2 to 37.5 kg/ m2). The mean serum E2 level was 27.8 pg/ml with a range of 2.7 to 131.3 pg/ml (101.9 pmolll with a range of 10 to 482 pmolll). The highest E2 level (and the only one that was in the normal premenopausal range) was measured in a woman who was overweight (BMI 37.5 kg/m2), had undergone hysterectomy, had vasomotor symptoms, and also had a repeatedly elevated FSH level (19). The mean of daily temperatures for the 10 days before treatment and treatment phases in MPA and placebo treatment months are shown in Table 1. The mean of daily temperatures during the 10 days on active MPA was 36.64 :±: 0.15 e (mean:±: SD) and during the 10 days on placebo was 36.37 :±: 0.14 e (P = 0.001). 1224 Prior et al. Thermogenic effect of MPA The temperature change from the pre intervention to intervention phases is also shown in Table 1. The mean change during the month in which MPA was administered was +0.27 :±: 0.16 versus 0.00 :±: 0.08 e during the month in which placebo was given (P = 0.002). Finally, we used a least squares analysis (Maximina computer program [16]) for assessing a temperature shift within an individual's treatment month. As shown in Table 2, a significant temperature increase occurred during days 17 to 26 of drug administration in 7 of 11 MPA treatment months and during 1 of 11 placebo months. The difference in the occurrence of increased temperature shifts is significant by X 2 (P = 0.008). There is no relationship between the random order of interventions (7 of 11 had the MPA in the 1st month) and whether a significant temperature increase occurred (X2 = 0.898). DISCUSSION During the normal ovulatory menstrual cycle, a 0.318 :±: 0.09 e increase (95% confidence interval Table 2 Data for Subjects Showing a Significant Increase in Temperature by Maximina Program During MPA or Placebo Interventions Day of temperature Subject increase t value* Intervention 1 19 2.92 MPA 4 17 11.29 MPA 5 20 6.23 MPA 6 17 4.21 MPA 7 19 4.20 MPA 8 17 2.03 Placebo 9 23 3.53 MPA 11 17 5.36 MPA * Statistic derived from a paired t-test comparison of the mean temperature before with the mean temperature after the temperature shift. A t value"", +2.0 is significant at the P < 0.05 level.
0.301 to 0.334) in temperature occurs in the luteal phase under the influence of endogenous P (Prior JC, unpublished observations). This study indicates that the synthetic C-21 P derivative MPA has a significant and apparently physiological effect on the oral morning basal temperature. To our knowledge, this is the first clear documentation of the thermogenic effect ofmpa. It remains uncertain whether P exerts its temperature effects through mechanisms that are predominantly central (20) or peripheral (21). Recognition that MPA influences basal temperature becomes useful because women who are prescribed progestins for menometrorrhagia or oligomenorrhea also may want to use symptothermal methods to control fertility (22, 23). It also is useful for a woman who has been treated for some duration with cyclic MPA for ovulatory disturbances with or without osteopenia (12) to know whether she has become normally ovulatory. If MPA were to have no thermogenic effect on core temperature, diagnosis of the return of normally ovulatory cycles could be obtained by basal temperature records. For these reasons, it becomes important to know whether this commonly used exogenous progestin potentially can increase the basal temperature and thus obscure the thermogenic effect of endogenous P. The thermal effect of MPA appears to be neither totally consistent nor universally documented. We had shown previously that a woman with ovulation disturbances who took cyclic MPA on days 16 to 25 of her own cycle (doses and duration of treatment equal to those used in this study) had basal temperature changes that were sometimes indicative of ovulation and sometimes not (14). Similar variability is shown in this study. A few women had no temperature shift on MPA. Limits of the sample size preclude us from determining whether other factors (such as age or serum estrogen level) might influence the thermal response to MP A. We have applied the computerized least squares analysis to these data. We currently use this analysis in menstrual cycle temperature data sets as a way of documenting ovulation based on the thermogenic action ofp (16). By these criteria, 7 of 11 cycles had a significant temperature increase within the days of MPA administration. The one significant temperature increase during a placebo "cycle" would not be unexpected given that the acceptance level of a type 1 error was set at a = 0.05. Vasomotor symptoms are known to be associated with a transient small decrease in core temperature (24). Medroxyprogesterone acetate treatment also is known to be effective for vasomotor symptom control (25). For both reasons, women volunteers for this study were selected to have few vasomotor symptoms (19). Although it is possible that some of the variability in temperature change during MPA administration related to the mild vasomotor symptoms that were recorded (19), it is unlikely that they would have altered the results of this study. Although progestin treatment commonly is given either cyclically or continuously as part of combined menopausal hormone treatment, cyclic progestins also are given to younger women with menstrual disturbances (12). This makes it important for all progestins being introduced to the market to be evaluated for their thermal effects. Further, studies investigating the presence or absence of effects on basal temperature ofthe currently available progestins, such as norethindrone and norgestrol, are needed. Acknowledgments. We thank the women volunteers, Nenita Alojado, R.N., for study coordination and data collection, and Luciana Frighetto, B.Sc. Pharm., and the pharmacy staff at Vancouver Hospital and Health Sciences Centre for randomizing the order of administration and holding the code and the medications. Medroxyprogesterone acetate and placebo were supplied by Upjohn Company of Canada (Toronto, Ontario, Canada). REFERENCES 1. Squire W. Puerperal temperatures. Lancet 1867;2:293-5. 2. Marshall JR. Thermal changes in the normal menstrual cycle. Br Med J 1963;1:102-4. 3. Vollman RF. The menstrual cycle. In: Friedman EA, editor. Major problems in obstetrics and gynecology. Vol. 7. Toronto: W.B. Saunders Company, 1977:11-193. 4. Doring GK. The incidence of anovular cycles in women. J Reprod Fertil 1969;(Suppl 6):77-81. 5. Rothchild I, Barnes AC. The effects of dosage and of estrogen, androgen or salicylate administration on the degree of body temperature elevation induced by progesterone. Endocrinology 1952;50:485-96. 6. Israel SL, Schneller O. The thermogenic property of progesterone. Fertil Steril 1950;1:53-65. 7. Ranick LD, Shangold MM, Ahmed SW. Cervical mucus and serum estradiol as predictors of response to progestin challenge. Fertil Steril 1990;54:352-5. 8. Gambrell RD Jr, Massey FM, Castaneda TA, Ugenas AJ, Ricci CA, Wright JM. Use of the progestogen challenge test to reduce the risk of endometrial cancer. Obstet Gynecol 1980;55:732-8. 9. Cowan BD, Morrison JC. Management of abnormal genital bleeding in girls and women. N Engl J Med 1991;324: 1710-5. 10. Kirkham C, Hahn PM, Van Vugt DA, Carmichael JA, Reid RL. A randomized, double-blind, placebo-controlled, crossover trial to assess the side effects of medroxyprogesterone acetate in hormone replacement therapy. Obstet Gynecol 1991;78:93-7. 11. Prior JC. Progesterone and its role in bone remodelling. In: Ziegler R, Pfeilschifter J, Brautigam M, editors. Sex steroids and bone. Berlin: Springer-Verlag, 1993:29-56. 12. Prior JC, Vigna Y, Barr SI, Rexworthy C, Lentle BC. Cyclic medroxyprogesterone treatment increases bone density: a controlled trial in athletic women with menstrual cycle disturbances. Am J Med 1994;96:521-30. 13. Prior JC, Vigna YM, Schechter MT, Burgess AE. Spinal bone Prior et al. Thermogenic effect of MPA 1225
loss and ovulatory disturbances. N Engl J Med 1990; 323:1221-7. 14. Prior JC, Vigna YM. Ovulation disturbances and exercise training. Clin Obstet Gynecol1991;34:180-90. 15. Barr SI, Prior JC, Vigna Y. Restrained eating and ovulatory disturbances: possible implications for bone health. Am J Clin Nutr 1993;59:92-7. 16. Prior JC, Vigna YM, Schulzer M, Hall JE, Bonen A. Determination of luteal phase length by quantitative basal temperature methods: validation against the midcycle LH peak. Clin Invest Med 1990;13:123-31. 17. Cumming DD. Questions and answers: effects of exogenous steroids on basal body temperature. JAmMed Assoc 1990;264:2448-50. 18. Goodman AS, Gilman AG, Gilman A. The pharmacological basis of therapeutics. 6th ed. New York: Macmillan Publishing Company Inc, 1980. 19. Prior JC, Alojado N, McKay DW, Vigna YM. No adverse ef- fects of medroxyprogesterone treatment without estrogen in postmenopausal women: double-blind, placebo-controlled, cross-over trial. Obstet Gynecol 1994;83:24-8. 20. Nakayama T. Thermosensitive neurons in the brain. Jpn J Physiol 1985;35:375-89. 21. Frascarolo P, Schutz Y, Jequier E. Decreased thermal conductance during the luteal phase of the menstrual cycle in women. J Appl Physiol 1990;69:2029-33. 22. Billings EL, Brown JB, Billings JJ, Burger HG. Symptoms and hormonal changes accompanying ovulation. Lancet 1972;1:282-4. 23. Marshall JR. A field trial of the basal body temperature method of regulating births. Lancet 1968;2:8-10. 24. Kronenberg F, Downey JA. Thermoregulatory physiology of menopause hot flashes: a review. Can J Physiol Pharmacol 1987;65:1312-24. 25. Schiff I, Tulchinsky D, Cramer D, Ryan KJ. Oral medroxyprogesterone in the treatment of postmenopausal symptoms. JAmMed Assoc 1980;244:1443-5. 1226 Prior et al. Thermogenic effect of MPA