J Am Board Fam Pract: first published as 10.3122/15572625-11-5-378 on 1 September 1998. Downloaded from MEDICAL PRACTICE Hormorial Therapy in the Management of Premenstrual Syndrome

Jeffrey D. Tiemstra, MD, and Krishna Patel, PharmD

Background: Premenstrual syndrome (PMS) is characterized by any of a number of physical and psychological symptoms consistently occurring in the late luteal phase. therapy is often recommended based on anecdotal evidence, although controlled studies have shown it to be ineffective. Oral contraceptives are more often used with mixed results. When hormonal therapy for PMS is indicated, the most appropriate choice remains a challenge. Methods: We describe a case report of successful hormonal therapy for PMS and a review of the literature on the effectiveness of hormonal therapies. Results and Conclusimls: is clearly effective in relieving 'symptoms of PMS, whereas progesterone is ineffective and might even worsen symptoms. Combination oral contraceptives are effective, undoubtedly as a result of the estrogen component. While little comparative data exist to guide choice of an oral contraceptive, maximizing the relative estrogenic potency of the oral contraceptive seems logical. Depressive symptoms might not respond to hormonal treatment, and new research suggests that selective serotonin reuptake inhibitors might be particularly effective. 0 Am Board Pam Pract 1998;11:378-81.)

Premenstrual syndrome (PMS), also referred to as elimination of caffeine, biofeedback, and relaxation late luteal phase dysphoric disorder, is a problem therapy. Because the beneficial effects of these copyright. commonly encountered in primary care, with up measures have not been proved, patients who seek to 10 percent of women of childbearing age experi­ medical attention for this syndrome will usually re­ encing symptoms severe enough to seek medical quire pharmacologic treatments. Patients with few treatment. I Despite decades of research, standard or closely related symptoms might benefit from criteria for the diagnosis of PMS are still debated. medication targeted at those symptoms. For in­ PMS can include any number of a wide variety of stance, headaches, cramping, and abdominal pain symptoms, such as depression, mood lability, can benefit from anti-inflammatory swelling, abdominal pain, breast tenderness, ; abdominal bloating and leg edema will re- ' headaches, or fatigue. Although the constellation spond to loop ; severe emotional symp­ http://www.jabfm.org/ of symptoms will vary from patient to patient, toms will respond to antidepressants or anxiolytics; characteristic of PMS is the timing of symptoms and migraine-type headaches can respond to mi­ around the menstrual cycle: symptoms predictably graine prophylactic or abortive therapy. Neverthe­ occur during the luteal phase, subside with menses, less, many patients either have too many symp­ and are absent during the follicular phase. toms to be targeted individually or have systemic

Management of this common syndrome re­ symptoms such as fatigue for which there is no on 25 September 2021 by guest. Protected mains a challenge because many popular treat­ specific therapy.2-4 ments are unproved or known to be ineffective. As a result of the temporal relation of PMS to NOflpharmacologic recommendations often in­ the menstrual cycle, much attention has been given clude high-carbohydrate diets, vitamin and min­ to estrogen and progesterone treatment, with the eral supplementation (particularly pyridoxine), expectation that the entire array ofPMS symptoms might respond. Results have been mixed, however, and misconceptions persist among both patients Submitted, revised, 24 Sept 1997. From the Family Practice Residency, Saint Joseph Medical and providers regarding the relative efficacies of Center (JDl),Joliet, and the Departtnent of Phamlacy Prac­ estrogen and progesterone therapy. Specifically, tice, University of Illinois at Chicago (KP), Ill. Address reprint requests to Jeffrey D. liemstra, MD, Family Practice progesterone continues to be widely used in treat­ Residency, 24024 \Y. Brancaster Dr, Naperville, IL 60564. ment despite convincing evidence that it is not

378 ]ABFP Sept.-Oct.1998 Vol. 11 No.5 J Am Board Fam Pract: first published as 10.3122/15572625-11-5-378 on 1 September 1998. Downloaded from beneficial, whereas estrogen remains underrecog­ ual patients neither the timing nor the severity of nized as effective therapy. The following case illus­ symptoms correlated with fluctuations in hormone trates some of these pitfalls. levels.5,6 A 1991 study confirmed that variations in hormone levels did not correlate with symptoms, Illustrative Case and additionally found that pharmacologically ter­ A 38-year-old woman came to the office for man­ minating the luteal phase did not prevent the nor­ agement of PMS symptoms. The patient de­ mal cyclic appearance ofPMS symptoms.7 Thus scribed fatigue, sore throat, cold sores, myalgias, although PMS symptoms correlate with the late and ,chills starting in the week before her men­ luteal phase of the menstrual cycle, luteal phase strual period and peaking in the first 2 days of her levels of and progesterone do not appear period. By days 3 and 4 her symptoms completely to be direct causative factors. resolved, and she would feel fine for the next 2 to More recent hypotheses on the cause of PMS 3 weeks. She specifically denied headaches, bloat­ suggest a role for fluctuating levels of neurotrans­ ing, swelling, or weight gain. Her menstrual flow mitters in the production of symptoms.8,9 Specifi­ was heavy (eight or more heavy pads) the first 3 cally, estrogen has been found to stimulate certain days, tapering within 3 more days. These symp­ populations of dopamine and serotonin receptors, toms had developed within the last 5 years; before the same receptors that are stimulated by selective that she had more severe cramping with her peri­ serotonin reuptake inhibitor (SSRI) antidepres­ ods but less severe systemic symptoms. During the sants. 1O In addition, drugs, such as ta­ past 3 years she had sought medical care repeatedly moxifen and progesterone, have been found to for this condition. Dietary modifications included stimulate PMS symptoms in susceptible patients. I I low-fat, high-complex-carbohydrate diet, elimina­ Thus, PMS symptoms appear to be mediated in tion of caffeine, and supplementation with a B­ the central nervous system, and declining levels of complex vitamin. Pharmacologic treatments had estrogen in the late luteal phase might indirectly included ibuprofen, naproxen, cyclic progesterone, contribute to the timing of their appearance. and most recently an oral contraceptive (Lo/Ovral) Because of the decline of both estrogen and copyright. for 3 months. None of these treatments had pro­ progesterone levels in the late luteal phase coinci­ duced any relief. dent with the timing of PMS symptoms, both have Recent evaluations had included physical exam­ been considered and studied as potential treat­ ination, pelvic sonogram, complete blood count, ments. New understanding of the role of neuro­ iron studies, monospot test, thyroid function tests, transmitters in PMS has helped explain what have and glucose tolerance tests, the results of which been in the past confusing and conflicting data. were normal. The patient's prescription for oral contracep­ Treatment http://www.jabfm.org/ tives was changed from a OJ-mg /0.03- Progesterone estradiol combination (Lo/Ovral) to a 1.0-mg Anecdotal case reports suggested that progesterone norethindrone/0.05-mg combination might be of benefit in the treatment ofPMS. Con­ (Ortho-Novum 1150) to achieve greater estrogenic trolled studies, however, have consistently shown activity relative to progestational activity. WIth this no benefit from progesterone; in fact, several stud­

change the patient had complete resolution of her ies have found a worsening of symptoms with prog­ on 25 September 2021 by guest. Protected symptoms within 2 months and was very pleased esterone. 12-14 Unfortunately, it continues to be rec­ with the res'ult. ommended as an option worth trying. Case reports of responses to progesterone treatment are most Causes of PMS likely explained by the fairly high placebo response Because PMS symptoms are by definition tempo­ seen in many PMS studies (up to 50 percent),9 al­ rally related to the menstrual cycle, it seems logical though weak estrogenic effects of therapeutic doses that cyclic fluctuations in estrogen or progesterone of progesterone through peripheral conversion to would be causally related. Studies in the 1980s, estrogen are theoretically possible. however, found that estrogen and progesterone \Vith the increasing use of long-acting forms of levels of patients with PMS did not differ from progesterone for contraception (Depo-Provera, asymptomatic control patients, and that in individ- Norplant), physicians should be aware of the po-

Hormonal Therapy for Premenstrual Syndrome 379 J Am Board Fam Pract: first published as 10.3122/15572625-11-5-378 on 1 September 1998. Downloaded from tential for these treatments to aggravate preexist­ effect will be more beneficial for PMS patients, ing PMS. Women with PMS symptoms of mood there are few data to guide the physician's choice of changes, food cravings, sleep disturbances, and a particular oral contraceptive. Studies looking at dysmenorrhea might experience worsening of monophasic and triphasic formulations individu­ symptoms with these progesterone-only forms of ally have found positive effects for both. One study contraception. that compared monophasic with triphasic formu­ lations included eight different pill formulations Estrogen in their two groups; no significant differences Given the new information on the effect of estro­ were found in symptoms, with the exception of gen on neurotransmitters, it is not surprising that breast tenderness (improved more in monophasic estrogen is effective in the treatment ofPMS. Con­ group).24 To date no controlled studies have been tinuous estrogen therapy has consistently proved done comparing a specific combination in differ­ effective in relieving PMS symptoms, whether ent ratios to assess the effect of varying the estro­ orally posthysterectomy, J5 transdermally,16 or gen-progesterone ratio. through implants.17 Unfortunately, the addition of Given the current data on for cyclic progesterone to these regimens results in a PMS, four conclusions can reasonably be drawn reduction of the beneficial effect. Because the car­ regarding choosing an oral contraceptive: (1) the cinogenic potential of unopposed estrogen pre­ initial choice should be based on suitability to the cludes its use for most women, the usefulness of es­ individual patient rather than that PMS is being trogen as a treatment of PMS is compromised by treated; (2) given the variability in response rates, if the need to add cyclic progesterone. the initial choice is ineffective, it is reasonable to try different combinations before abandoning hor­ Oral Contraceptives monal therapy; (3) when there is no clear-cut first Combination oral contraceptives have both estro­ choice, or when a change in oral contraceptive genic and progestational effects. Whereas the rela­ choice is indicated, a combination with a relatively copyright. tive strengths of these effects can be approximated higher estrogen activity and minimal progesterone for individual pill combinations by the incidence of activity might be more effective in treating PMS various side effects, the relative effects of any given symptoms, with the possible exception of breast combination on an individual patient are less pre­ tenderness; and (4) PMS-related depressive symp­ dictable. Relative absorption of the estrogen and toms might not respond well to hormonal therapy. progesterone, peripheral conversion of both hor­ If depressive symptoms are the only symptoms un­ mones, and individual susceptibility to specific side responsive to an oral contraceptive, adding an effects can all determine the patient's response. Be­ SSRI-type antidepressant might be preferable to cause the beneficial effect of oral contraceptives on switching the oral contraceptive. http://www.jabfm.org/ PMS is most likely due to the estrogen effect, it is not surprising that combination oral contracep­ Conclusion tives have produced less dramatic results in the Premenstrual syndrome is a common patient treatment of PMS than estrogen alone. problem in primary care. Many patients will come Both observationaP8-21 and controlled prospec­ to their family physician for treatment, often when tive22-24 studies have generally found that oral con­ nonpharmacologic treatment has failed. Symp­ on 25 September 2021 by guest. Protected traceptive users have fewer PMS symptoms than tom-specific therapy can provide relief for some nonusers. Symptoms likely to respond include patients, but hormone therapy should be consid­ more specific physical symptoms such as bloating, ered when a wide variety of symptoms are de­ headaches, abdominal pain, and breast tenderness. scribed, when symptom-specific therapy is inade­ Less clear has been the effect of oral contraceptives quate, or when contraception is desired. on such emotional symptoms as irritability and de­ Progesterone therapy should be avoided, be­ pression; some studies have suggested that PMS­ cause progesterone is no better than placebo and related depression might not respond as well as can actually worsen the condition. Unopposed es­ physical symptoms to oral contraceptives.20,22 trogen therapy can be considered in patients who Although it is reasonable to conclude that oral are not at risk for endometrial cancer. For most contraceptives with relatively greater estrogenic women, however, combination oral contraceptives

380 JABFP Sept.-Oct.1998 Vol. 11 No.5 J Am Board Fam Pract: first published as 10.3122/15572625-11-5-378 on 1 September 1998. Downloaded from will be the hormonal therapy of choice. As initial premenstrual syndrome. Clin Obstet GynecoI1992; therapy any monophasic or triphasic low-dose oral 35:587-98. contraceptive is appropriate. If a response is not 10. Fink G, Sumner BE, Rosie R, Grace 0, Quinn]P. Es­ trogen control of central neurotransmission: effect on seen initially, different combinations with increas­ mood, mental state, and memol)'. Cell Mol Neuro­ ing relative estrogenic activity can be tried before bioi 1996;16:325-44. abandoning oral contraceptive therapy altogether. 11. Arpels] C. The female brain hypoestrogenic contin­ When prescribing oral contraceptives, physicians uum from the premenstrual syndrome to menopause. should keep in mind the common risks and com- • A hypothesis and review of supporting data.] Reprod Med 1996;41:633-9. plications, particularly the increased risk of cardio­ 12. Moline ML. Pharmacologic strategies for managing vascular complications for women 30 to 50 years of premenstrual syndrome. Clin Pharm 1993;12:181-96. age. 13. Freeman E, Rickels K, Sondheimer S], Polansky M. If the response to oral contraceptives is inade­ Ineffectiveness of progesterone suppository treatment quate, other hormonally directed therapies such as for premenstrual syndrome.]AMA 1990;264:349-53. 14. Sampson GA. Premenstrual syndrome: a double­ , -releasing hormone ana­ blind controlled trial of progesterone and placebo. Br logues, and oophorectomy have clearly proved ef­ ] Psychiatry 1979;135:209-15. fectiveness, but their side effects and long-term 15. Casper RF, Hearn MT. The effect of hysterectomy consequences are greater, and they should gener­ and bilateral oophorectomy in women with severe ally be reserved for only the most extreme cases. 1 premenstrual syndrome. Am] Obstet Gynecol 1990; 162:105-9. One recent study found that fluoxetine was ef­ 16. Watson NR, StuddJw. Savvas M, Garnett T, Baber fective in the treatment of premenstrual dyspho­ R]. Treatment of severe premenstrual syndrome with ria. 25 This finding, combined with a better under­ oestradiol patches and cyclical oral . standing of the cause of PMS, suggests that for Lancet 1989;2:730-2. patients with predominantly emotional symptoms, 17. MagosAL, BrincatM, Studd]W. Treatment of the premenstrual syndrome by subcutaneous estradiol an SSRI antidepressant might be a better initial implants and cyclical oral norethisterone: placebo­ treatment choice than hormonal therapy. controlled study. Br Med] Clin Res Ed 1986;292: copyright. 1629-33. References 18. Bancroft], Rennie D. The impact of oral contracep­ 1. Hsia LS, Long MH. Premenstrual syndrome. Cur­ tives on the experience of perimenstrual mood, clum­ rent concepts in diagnosis and management.] Nurse siness, food craving and other symptoms.] Psycho­ MidwifeI)' 1990;35:351-7. som Res 1993;37:195-202. 2. Severino SK, Moline ML. Premenstrual syndrome. 19. Brown S, Vessey M, Stratton 1. The influence of Identification and management. Drugs 1995;49:71- method of contraception and cigarette smoking on 82. menstrual patterns. Br] Obstet Gynaecol 1988;95: 3. DeMonico SO, Brown CS, Ling nv, Premenstrual 905-10.

syndrome. Curr Opin Obstet Gynecol 1994;6:499- 20. Kutner S], Brown '\VL. Types of oral contraceptives, http://www.jabfm.org/ 502. depression, and premenstrual symptoms. J Nerv 4. Parker PD. Premenstrual syndrome. Am Fam Physi­ MentDis 1972;155:153-62. cian 1994;50:1309-17. 21. Herzberg B, Coppen A. Changes in psychological 5. Rubinow DR, Hoban MC, Grover GN, Galloway symptoms in women taking oral contraceptives. Br J DS, Roy-Bryne P, Andersen R, et al. Changes in Psychiatry 1970;116:161-3. plasma hormones across the menstrual cycle in pa­ 22. Graham CA, Sherwin BB. A prospective treatment tients with menstrually related mood disorders and in study of premenstrual symptoms using a triphasic oral control subjects. Am] Obstet GynecoI1988;158:5-11. contraceptive.) Psychosom Res 1992;36:257-66. on 25 September 2021 by guest. Protected 6. Backstrom T, Sanders D, Leask R, Davidson D, 23. Backstrom T, Hansson-Malmstrom Y, Lindhe BA, Warner P, Bancroft]. Mood, sexuality, hormones, and Cavalli-Bjorkman B, Nordenstrom S. Oral contra­ the menstrual cycle. II. Hormone levels and their re­ ceptives in premenstrual syndrome: a randomized lationship to the premenstrual syndrome. Psychosom comparison of triphasic and monophasic prepara­ Med 1983;45:503-7. tions. Contraception 1992;46:253-68. 7 Schmidt P], Neiman LK, Grover GN, Muller KL, 24. Walker A, Bancroft}. Relationship between premen­ Merriam GR, Rubinow DR. Lack of effect of induced strual symptoms and oral contraceptive use: a con­ menses on symptoms in women with premenstrual trolled study. Psychosom Med 1990;52:86-96. syndrome. N Engl] Med 1991;324: 1174-9. 25. Steiner M, Steinberg S, Stewart D, Carter D, Berger 8. Smith S, Schiff I, editors. Modem management of C, Reid R, et al. Fluoxetine in the treatment of pre­ premenstrual syndrome. New York: \V\V Norton & menstrual dysphoria. Canadian FluoxetinelPremen­ Company, 1993. strual Dysphoria Collaborative Study Group. N Engl 9. Mortola]E Issues in the diagnosis and research of J Med 1995;332:1529-34.

Hormonal Therapy for Premenstrual Syndrome 381

~------~------