Clinical Review

Endometrial polyps are a relatively common pathology, occurring in 24% to 41% of women who have abnormal bleeding, and in about 10% of asymptomatic women.^1,2 Endometrial polyps may be associated with leiomyomas in women who have abnormal bleeding.^1-3

Polyps originate as focal hyperplasia of basal endometrium and contain variable amounts of glands, stroma, and blood vessels. Glandular epithelium has higher estrogen- and progesterone-receptor expression than surrounding endometrium, whereas the stromal component of a polyp has hormone receptors similar to endometrium. This suggests that a polyp represents focal hyperplasia that is more glandular than stromal.⁴

In this Update, I outline the basics of diagnosis and treatment and report on several recent investigations:

• a retrospective analysis from Italy that found that endometrial polyps are associated with advancing age and that any apparent association between polyps and diabetes, hypertension, or obesity is likely age-related
• a cross-sectional study from Norway that found that some asymptomatic polyps regress spontaneously, usually when their length is 10.7 mm or less
• three studies that explore the variables associated with premalignant and malignant polyps
• an investigation of the relationship between endometrial polyps and the background endometrium that found atypical hyperplasia in endometrium remote from the polyp in a significant percentage of women.

Age is the most important variable when assessing a patient for endometrial pathology

Nappi L, Indraccolo U, Di Spiezio Sardo A, et al. Are diabetes, hypertension, and obesity independent risk factors for endometrial polyps? J Minim Invasive Gynecol. 2009;16(2):157–162.

In this retrospective analysis of 353 women who underwent office hysteroscopy, Nappi and co-workers set out to ascertain whether endometrial polyps are associated with diabetes, hypertension, or obesity, independent of age and menopausal status. They did find an association between age, menopause, hypertension, obesity, and the presence of endometrial polyps. However, after multivariable logistic regression, all variables except age lost statistical significance. The median age at which polyps were present was 53 years (range: 29–86 years).

Details of the trial

A total of 394 consecutive Caucasian women underwent hysteroscopy to assess abnormal uterine bleeding, infertility, cervical polyps, or abnormal sonographic patterns (e.g., postmenopausal endometrial thickness >5 mm, endometrial hyperechogenic spots). Of these women, 353 were included in the study, and demographic characteristics and data on diabetes, hypertension, and menopausal status were collected. Anthropometric parameters were also analyzed. When a polyp was detected, it was removed via office hysteroscopy, and histologic analysis was performed.

Small, asymptomatic uterine polyps may regress without treatment

Lieng M, Istre O, Sandvik L, Qvigstad E. Prevalence, 1-year regression rate, and clinical significance of asymptomatic endometrial polyps: cross-sectional study. J Minim Invasive Gynecol. 2009;16(4):465–471.

The treatment of asymptomatic polyps is controversial because their clinical consequences, malignant potential, and spontaneous regression rate are unknown. In this study, Lieng and colleagues prospectively estimated the prevalence and 1-year regression rate of incidentally diagnosed endometrial polyps in women 40 to 45 years old, as well as bleeding patterns and intensity.

They found polyps in 31 (12.1%) of 257 randomly selected women. At 1 year, the regression rate was 27%.

Details of the trial

At study inception, a standard 10-point visual analog scale was used to quantify each participant’s periodic bleeding, and a physical examination was performed. Transvaginal ultrasonography (US) and saline infusion sonography (SIS) were also performed. When a polyp was detected, researchers measured its length and used Doppler US to visualize the vessel feeding the polyp. An endometrial biopsy was also obtained.

The mean length of polyps was 14 mm (standard deviation [SD], 5.2 mm; 95% confidence interval [CI], 12.1–15.9; median, 13.4 mm; range, 6.7–28.7 mm), and the feeding vessel was identified for 22 of 31 polyps (71%). (For comparison, consider the findings of Clevenger and associates, who reported mean polyp diameters of 13.9 mm and 8.5 mm (P = .064), respectively, among women who had abnormal bleeding and women who did not.¹)

When researchers compared women who had polyps with those who had none, they found no significant differences in age, body mass index, blood pressure, gynecologic symptoms, menopausal status, use of hormone therapy, or use of the levonorgestrel-releasing intrauterine device (Mirena). Women who had endometrial polyps scored significantly higher, however, than women who did not on the visual analog scale for periodic bleeding and on the Pictorial Blood Assessment Chart—even when women who had myomas were excluded from the analysis. Although mean hemoglobin levels were similar between groups, women who had polyps had a significantly lower mean ferritin level (25 μg/L vs 41 μg/L; P = .05).

Polyps that persisted were larger from the start

Polyps regressed spontaneously in eight women, six of whom had the feeding vessel visualized at the initial consultation. Polyps that persisted after 12 months were significantly larger (mean polyp length, 15.1 mm; SD, 5.3 mm; 95% CI, 12.7–17.5) at study inception than were those that regressed (mean polyp length, 10.7 mm; SD, 3.9 mm; 95% CI, 7.5–14.0). Polyps that persisted beyond 1 year became significantly longer during follow-up, increasing from a mean length of 15.1 mm to 18.1 mm (SD, 7.9 mm; 95% CI, 0.7–5.3; P = .01).

Twenty of the 22 women who had persistent polyps underwent transcervical resection, one underwent laparoscopic supracervical hysterectomy, and one refused treatment. There were no complications.

Histology revealed that the polyps were benign in 16 women (80%), polypoid in two women (10%), and myomas in two women (10%). No atypical or malignant changes were observed in the polypectomy patients or among all participants.

A small, separate series (three patients) found all polyps to be 5 mm to 8 mm in length at detection, with a regression rate of 100% over several months.⁵

What variables signal a greater risk of malignancy?

Baiocchi G, Manci N, Pazzaglia M, et al. Malignancy in endometrial polyps: a 12-year experience. Am J Obstet Gynecol. 2009;201(5):462.e1–e4.

Gregoriou O, Konidaris S, Vrachnis N, et al. Clinical parameters linked with malignancy in endometrial polyps. Climacteric. 2009;12(5):454–458.

Wang JH, Zhao J, Lin J. Opportunities and risk factors for premalignant and malignant transformation of endometrial polyps: management strategies. J Minim Invasive Gynecol. 2010;17(1):53–58.

These three studies explore various aspects of a fundamental challenge: how to discriminate between polyps likely to undergo malignant transformation and those that will not.

The answer: Look for menopausal status, abnormal uterine bleeding, diabetes, obesity, and hypertension. Polyps larger than 1 cm also appear more likely to become malignant.

Details of the trials

In a retrospective study involving 1,242 women who had endometrial polyps, Baiocchi and colleagues identified 95.2% of the polyps as benign, 1.3% as premalignant, and 3.5% as malignant. Four clinical variables were significantly associated with premalignant and malignant features:

• age
• menopausal status
• abnormal uterine bleeding
• hypertension.

In their series of 516 cases, Gregoriou and associates found 96.9% of polyps to be benign, 1.2% to be premalignant, and 1.9% to be malignant. Four variables were associated with premalignant and malignant features:

• age above 60 years
• menopausal status
• obesity
• diabetes.

And in a study involving 766 patients, Wang and colleagues found 96.2% of polyps to be benign, 3.26% to involve hyperplasia with atypia, and 0.52% to be malignant. Among the variables associated with premalignant and malignant polyps were:

• polyp diameter larger than 1 cm
• menopausal status
• abnormal uterine bleeding.

Diagnosis may be trickier than you think, but treatment is straightforward

When a patient complains of abnormal uterine bleeding, evaluation often begins with transvaginal ultrasonography (US). Among the challenges of assessing the endometrium using US is the unreliability of endometrial thickness as a predictor of pathology. For example, Breitkopf and colleagues found that transvaginal US missed intracavitary lesions in one of six premenopausal women who had abnormal bleeding and an endometrial stripe thinner than 5 mm, for a sensitivity of 74%.⁶

In a separate study, Marello and colleagues used the combination of hysteroscopy and directed biopsy—the gold standard of diagnosis—to evaluate 212 postmenopausal women who had an endometrial thickness of 4 mm or less.⁷ (This parameter has been suggested as a cutoff for symptomatic postmenopausal women.⁸) Of these 212 women, 10% were found to have histologically confirmed intracavitary pathology (16 polyps and 4 submucous myomas).⁷ Among 13 symptomatic women in this study, three (23%) were found to have an endometrial polyp.⁷

These studies suggest that endometrial thickness alone should not be used to exclude benign endometrial pathology in symptomatic women, be they premenopausal or postmenopausal. No data back routine US to measure endometrial thickness in asymptomatic postmenopausal women.

Hysteroscopy and SIS are preferred

Both hysteroscopy and saline infusion sonography (SIS) have significantly better sensitivity and specificity in the diagnosis of intracavitary pathology than transvaginal US alone in women who have abnormal bleeding (FIGURE 1; VIDEOS 1, 2, AND 3 related to this article in the OBG Management Video Library at obgmanagement.com).^1,9 Hysteroscopy and SIS detect polyps with equal accuracy.¹⁰ However, hysteroscopy allows for removal of endometrial polyps and directed biopsy at the time of diagnosis.⁹

FIGURE 1 Imaging of polyps: go beyond transvaginal ultrasonography for optimal visualization

A. Hysteroscopic view of an endometrial polyp. B. The view with saline-infusion sonography.

In symptomatic women, resect the polyp

Polypectomy improves abnormal bleeding, according to a systematic review by Nathani and associates.¹¹ All studies included in the review, which involved follow-up intervals between 2 and 52 months, reported such an improvement.¹¹

When it is performed in the office, polypectomy offers several advantages over its inpatient counterpart:

• higher cost-effectiveness
• greater convenience
• avoidance of general anesthesia.

In both settings, it can be performed using mechanical or bipolar electrosurgical instrumentation (VIDEO 4).¹²

Segmental resection of the polyp while it is partially attached to the uterine wall is the optimal removal technique for large polyps (FIGURE 2). A grasping forceps can then be used to remove the polyp completely (VIDEO 5). Instruments such as a basket and snare are helpful in removing the polyp effectively.¹³

FIGURE 2 segmental resection of a polyp

During hysteroscopic polypectomy, the polyp is resected in segments while it is still partially attached to the endometrial wall.

Is histologic analysis of a polyp sufficient risk assessment?

Rahimi S, Marani C, Renzi C, Natale ME, Giovannini P, Zeloni R. Endometrial polyps and the risk of atypical hyperplasia on biopsies of unremarkable endometrium: a study on 694 patients with benign endometrial polyps. Int J Gynecol Pathol. 2009;28(6):522–528.

This study involved 694 consecutive patients who had benign endometrial polyps. Investigators sought to clarify the relationship between polyps and the underlying endometrium—specifically, to determine whether a polyp is a “circumscribed pathology of the endometrium or a polypoid expression of endometrial hyperplasia.” In describing the rationale for the study, the authors observe that the association between polyps and premalignant and malignant changes remains unclear.

Participants underwent hysteroscopy for removal of the polyps, at which time two biopsies of “unremarkable” endometrium, far from the base of the polyp, were also obtained.

Overall, endometrial hyperplasia without atypia was identified on hysteroscopically unremarkable endometrium in 18% of women, and atypia was identified in 7.3%. Among postmenopausal women, hyperplasia without atypia was identified in 21.6% of cases, atypia in 12%, and adenocarcinoma in 1.2%.

Multivariable analysis revealed that postmenopausal women who had polyps heavier than 1 g were 3.6 times more likely to have atypia (95% CI, 1.3–10.3). Among premenopausal women, the likelihood of atypia increased when the polyp weighed more than 0.4 g (odds ratio [OR], 3.5; 95% CI, 1.1–10.9) or the patient was older than 40 years (OR, 3.82; 95% CI, 1.1–13.2).

Don’t miss the videos that accompany this article!
They’re accessible in the OBG Management Video Library at obgmanagement.com

VIDEO 1: Saline-infusion sonographic imaging of a polyp

VIDEO 2: Hysteroscopic imaging of a polyp in a menopausal patient. A flexible 3-mm scope is used to assess an asymptomatic menopausal woman in whom an enlarged endometrial stripe was identified during earlier imaging.

VIDEO 3: Hysteroscopic imaging of a polyp and associated hyperplasia. A flexible 3-mm scope is used to evaluate a menopausal women who has uterine bleeding, revealing a 2-cm polyp.

VIDEO 4: Office polypectomy. An 8-mm, apparently benign polyp is removed from a premenopausal woman using continuous-flow operative hysteroscopy in an office setting.

VIDEO 5: Removal of a large polyp. A large (2 cm x 2.5 cm) polyp is removed in pieces, with the polyp partially attached to the endometrial wall, from a premenopausal woman who has abnormal bleeding.

More: Watch Dr. Brent Seibel perform OR-based hysteroscopic polypectomy

Endometrial polyps are a relatively common pathology, occurring in 24% to 41% of women who have abnormal bleeding, and in about 10% of asymptomatic women.^1,2 Endometrial polyps may be associated with leiomyomas in women who have abnormal bleeding.^1-3

Polyps originate as focal hyperplasia of basal endometrium and contain variable amounts of glands, stroma, and blood vessels. Glandular epithelium has higher estrogen- and progesterone-receptor expression than surrounding endometrium, whereas the stromal component of a polyp has hormone receptors similar to endometrium. This suggests that a polyp represents focal hyperplasia that is more glandular than stromal.⁴

In this Update, I outline the basics of diagnosis and treatment and report on several recent investigations:

• a retrospective analysis from Italy that found that endometrial polyps are associated with advancing age and that any apparent association between polyps and diabetes, hypertension, or obesity is likely age-related
• a cross-sectional study from Norway that found that some asymptomatic polyps regress spontaneously, usually when their length is 10.7 mm or less
• three studies that explore the variables associated with premalignant and malignant polyps
• an investigation of the relationship between endometrial polyps and the background endometrium that found atypical hyperplasia in endometrium remote from the polyp in a significant percentage of women.

Age is the most important variable when assessing a patient for endometrial pathology

Nappi L, Indraccolo U, Di Spiezio Sardo A, et al. Are diabetes, hypertension, and obesity independent risk factors for endometrial polyps? J Minim Invasive Gynecol. 2009;16(2):157–162.

In this retrospective analysis of 353 women who underwent office hysteroscopy, Nappi and co-workers set out to ascertain whether endometrial polyps are associated with diabetes, hypertension, or obesity, independent of age and menopausal status. They did find an association between age, menopause, hypertension, obesity, and the presence of endometrial polyps. However, after multivariable logistic regression, all variables except age lost statistical significance. The median age at which polyps were present was 53 years (range: 29–86 years).

Details of the trial

A total of 394 consecutive Caucasian women underwent hysteroscopy to assess abnormal uterine bleeding, infertility, cervical polyps, or abnormal sonographic patterns (e.g., postmenopausal endometrial thickness >5 mm, endometrial hyperechogenic spots). Of these women, 353 were included in the study, and demographic characteristics and data on diabetes, hypertension, and menopausal status were collected. Anthropometric parameters were also analyzed. When a polyp was detected, it was removed via office hysteroscopy, and histologic analysis was performed.

Small, asymptomatic uterine polyps may regress without treatment

Lieng M, Istre O, Sandvik L, Qvigstad E. Prevalence, 1-year regression rate, and clinical significance of asymptomatic endometrial polyps: cross-sectional study. J Minim Invasive Gynecol. 2009;16(4):465–471.

The treatment of asymptomatic polyps is controversial because their clinical consequences, malignant potential, and spontaneous regression rate are unknown. In this study, Lieng and colleagues prospectively estimated the prevalence and 1-year regression rate of incidentally diagnosed endometrial polyps in women 40 to 45 years old, as well as bleeding patterns and intensity.

They found polyps in 31 (12.1%) of 257 randomly selected women. At 1 year, the regression rate was 27%.

Details of the trial

At study inception, a standard 10-point visual analog scale was used to quantify each participant’s periodic bleeding, and a physical examination was performed. Transvaginal ultrasonography (US) and saline infusion sonography (SIS) were also performed. When a polyp was detected, researchers measured its length and used Doppler US to visualize the vessel feeding the polyp. An endometrial biopsy was also obtained.

The mean length of polyps was 14 mm (standard deviation [SD], 5.2 mm; 95% confidence interval [CI], 12.1–15.9; median, 13.4 mm; range, 6.7–28.7 mm), and the feeding vessel was identified for 22 of 31 polyps (71%). (For comparison, consider the findings of Clevenger and associates, who reported mean polyp diameters of 13.9 mm and 8.5 mm (P = .064), respectively, among women who had abnormal bleeding and women who did not.¹)

When researchers compared women who had polyps with those who had none, they found no significant differences in age, body mass index, blood pressure, gynecologic symptoms, menopausal status, use of hormone therapy, or use of the levonorgestrel-releasing intrauterine device (Mirena). Women who had endometrial polyps scored significantly higher, however, than women who did not on the visual analog scale for periodic bleeding and on the Pictorial Blood Assessment Chart—even when women who had myomas were excluded from the analysis. Although mean hemoglobin levels were similar between groups, women who had polyps had a significantly lower mean ferritin level (25 μg/L vs 41 μg/L; P = .05).

Polyps that persisted were larger from the start

Polyps regressed spontaneously in eight women, six of whom had the feeding vessel visualized at the initial consultation. Polyps that persisted after 12 months were significantly larger (mean polyp length, 15.1 mm; SD, 5.3 mm; 95% CI, 12.7–17.5) at study inception than were those that regressed (mean polyp length, 10.7 mm; SD, 3.9 mm; 95% CI, 7.5–14.0). Polyps that persisted beyond 1 year became significantly longer during follow-up, increasing from a mean length of 15.1 mm to 18.1 mm (SD, 7.9 mm; 95% CI, 0.7–5.3; P = .01).

Twenty of the 22 women who had persistent polyps underwent transcervical resection, one underwent laparoscopic supracervical hysterectomy, and one refused treatment. There were no complications.

Histology revealed that the polyps were benign in 16 women (80%), polypoid in two women (10%), and myomas in two women (10%). No atypical or malignant changes were observed in the polypectomy patients or among all participants.

A small, separate series (three patients) found all polyps to be 5 mm to 8 mm in length at detection, with a regression rate of 100% over several months.⁵

What variables signal a greater risk of malignancy?

Baiocchi G, Manci N, Pazzaglia M, et al. Malignancy in endometrial polyps: a 12-year experience. Am J Obstet Gynecol. 2009;201(5):462.e1–e4.

Gregoriou O, Konidaris S, Vrachnis N, et al. Clinical parameters linked with malignancy in endometrial polyps. Climacteric. 2009;12(5):454–458.

Wang JH, Zhao J, Lin J. Opportunities and risk factors for premalignant and malignant transformation of endometrial polyps: management strategies. J Minim Invasive Gynecol. 2010;17(1):53–58.

These three studies explore various aspects of a fundamental challenge: how to discriminate between polyps likely to undergo malignant transformation and those that will not.

The answer: Look for menopausal status, abnormal uterine bleeding, diabetes, obesity, and hypertension. Polyps larger than 1 cm also appear more likely to become malignant.

Details of the trials

In a retrospective study involving 1,242 women who had endometrial polyps, Baiocchi and colleagues identified 95.2% of the polyps as benign, 1.3% as premalignant, and 3.5% as malignant. Four clinical variables were significantly associated with premalignant and malignant features:

• age
• menopausal status
• abnormal uterine bleeding
• hypertension.

In their series of 516 cases, Gregoriou and associates found 96.9% of polyps to be benign, 1.2% to be premalignant, and 1.9% to be malignant. Four variables were associated with premalignant and malignant features:

• age above 60 years
• menopausal status
• obesity
• diabetes.

And in a study involving 766 patients, Wang and colleagues found 96.2% of polyps to be benign, 3.26% to involve hyperplasia with atypia, and 0.52% to be malignant. Among the variables associated with premalignant and malignant polyps were:

• polyp diameter larger than 1 cm
• menopausal status
• abnormal uterine bleeding.

Diagnosis may be trickier than you think, but treatment is straightforward

When a patient complains of abnormal uterine bleeding, evaluation often begins with transvaginal ultrasonography (US). Among the challenges of assessing the endometrium using US is the unreliability of endometrial thickness as a predictor of pathology. For example, Breitkopf and colleagues found that transvaginal US missed intracavitary lesions in one of six premenopausal women who had abnormal bleeding and an endometrial stripe thinner than 5 mm, for a sensitivity of 74%.⁶

In a separate study, Marello and colleagues used the combination of hysteroscopy and directed biopsy—the gold standard of diagnosis—to evaluate 212 postmenopausal women who had an endometrial thickness of 4 mm or less.⁷ (This parameter has been suggested as a cutoff for symptomatic postmenopausal women.⁸) Of these 212 women, 10% were found to have histologically confirmed intracavitary pathology (16 polyps and 4 submucous myomas).⁷ Among 13 symptomatic women in this study, three (23%) were found to have an endometrial polyp.⁷

These studies suggest that endometrial thickness alone should not be used to exclude benign endometrial pathology in symptomatic women, be they premenopausal or postmenopausal. No data back routine US to measure endometrial thickness in asymptomatic postmenopausal women.

Hysteroscopy and SIS are preferred

Both hysteroscopy and saline infusion sonography (SIS) have significantly better sensitivity and specificity in the diagnosis of intracavitary pathology than transvaginal US alone in women who have abnormal bleeding (FIGURE 1; VIDEOS 1, 2, AND 3 related to this article in the OBG Management Video Library at obgmanagement.com).^1,9 Hysteroscopy and SIS detect polyps with equal accuracy.¹⁰ However, hysteroscopy allows for removal of endometrial polyps and directed biopsy at the time of diagnosis.⁹

FIGURE 1 Imaging of polyps: go beyond transvaginal ultrasonography for optimal visualization

A. Hysteroscopic view of an endometrial polyp. B. The view with saline-infusion sonography.

In symptomatic women, resect the polyp

Polypectomy improves abnormal bleeding, according to a systematic review by Nathani and associates.¹¹ All studies included in the review, which involved follow-up intervals between 2 and 52 months, reported such an improvement.¹¹

When it is performed in the office, polypectomy offers several advantages over its inpatient counterpart:

• higher cost-effectiveness
• greater convenience
• avoidance of general anesthesia.

In both settings, it can be performed using mechanical or bipolar electrosurgical instrumentation (VIDEO 4).¹²

Segmental resection of the polyp while it is partially attached to the uterine wall is the optimal removal technique for large polyps (FIGURE 2). A grasping forceps can then be used to remove the polyp completely (VIDEO 5). Instruments such as a basket and snare are helpful in removing the polyp effectively.¹³

FIGURE 2 segmental resection of a polyp

During hysteroscopic polypectomy, the polyp is resected in segments while it is still partially attached to the endometrial wall.

Is histologic analysis of a polyp sufficient risk assessment?

Rahimi S, Marani C, Renzi C, Natale ME, Giovannini P, Zeloni R. Endometrial polyps and the risk of atypical hyperplasia on biopsies of unremarkable endometrium: a study on 694 patients with benign endometrial polyps. Int J Gynecol Pathol. 2009;28(6):522–528.

This study involved 694 consecutive patients who had benign endometrial polyps. Investigators sought to clarify the relationship between polyps and the underlying endometrium—specifically, to determine whether a polyp is a “circumscribed pathology of the endometrium or a polypoid expression of endometrial hyperplasia.” In describing the rationale for the study, the authors observe that the association between polyps and premalignant and malignant changes remains unclear.

Participants underwent hysteroscopy for removal of the polyps, at which time two biopsies of “unremarkable” endometrium, far from the base of the polyp, were also obtained.

Overall, endometrial hyperplasia without atypia was identified on hysteroscopically unremarkable endometrium in 18% of women, and atypia was identified in 7.3%. Among postmenopausal women, hyperplasia without atypia was identified in 21.6% of cases, atypia in 12%, and adenocarcinoma in 1.2%.

Multivariable analysis revealed that postmenopausal women who had polyps heavier than 1 g were 3.6 times more likely to have atypia (95% CI, 1.3–10.3). Among premenopausal women, the likelihood of atypia increased when the polyp weighed more than 0.4 g (odds ratio [OR], 3.5; 95% CI, 1.1–10.9) or the patient was older than 40 years (OR, 3.82; 95% CI, 1.1–13.2).

Don’t miss the videos that accompany this article!
They’re accessible in the OBG Management Video Library at obgmanagement.com

VIDEO 1: Saline-infusion sonographic imaging of a polyp

VIDEO 2: Hysteroscopic imaging of a polyp in a menopausal patient. A flexible 3-mm scope is used to assess an asymptomatic menopausal woman in whom an enlarged endometrial stripe was identified during earlier imaging.

VIDEO 3: Hysteroscopic imaging of a polyp and associated hyperplasia. A flexible 3-mm scope is used to evaluate a menopausal women who has uterine bleeding, revealing a 2-cm polyp.

VIDEO 4: Office polypectomy. An 8-mm, apparently benign polyp is removed from a premenopausal woman using continuous-flow operative hysteroscopy in an office setting.

VIDEO 5: Removal of a large polyp. A large (2 cm x 2.5 cm) polyp is removed in pieces, with the polyp partially attached to the endometrial wall, from a premenopausal woman who has abnormal bleeding.

More: Watch Dr. Brent Seibel perform OR-based hysteroscopic polypectomy

Endometrial polyps are a relatively common pathology, occurring in 24% to 41% of women who have abnormal bleeding, and in about 10% of asymptomatic women.^1,2 Endometrial polyps may be associated with leiomyomas in women who have abnormal bleeding.^1-3

Polyps originate as focal hyperplasia of basal endometrium and contain variable amounts of glands, stroma, and blood vessels. Glandular epithelium has higher estrogen- and progesterone-receptor expression than surrounding endometrium, whereas the stromal component of a polyp has hormone receptors similar to endometrium. This suggests that a polyp represents focal hyperplasia that is more glandular than stromal.⁴

In this Update, I outline the basics of diagnosis and treatment and report on several recent investigations:

• a retrospective analysis from Italy that found that endometrial polyps are associated with advancing age and that any apparent association between polyps and diabetes, hypertension, or obesity is likely age-related
• a cross-sectional study from Norway that found that some asymptomatic polyps regress spontaneously, usually when their length is 10.7 mm or less
• three studies that explore the variables associated with premalignant and malignant polyps
• an investigation of the relationship between endometrial polyps and the background endometrium that found atypical hyperplasia in endometrium remote from the polyp in a significant percentage of women.

Age is the most important variable when assessing a patient for endometrial pathology

Nappi L, Indraccolo U, Di Spiezio Sardo A, et al. Are diabetes, hypertension, and obesity independent risk factors for endometrial polyps? J Minim Invasive Gynecol. 2009;16(2):157–162.

In this retrospective analysis of 353 women who underwent office hysteroscopy, Nappi and co-workers set out to ascertain whether endometrial polyps are associated with diabetes, hypertension, or obesity, independent of age and menopausal status. They did find an association between age, menopause, hypertension, obesity, and the presence of endometrial polyps. However, after multivariable logistic regression, all variables except age lost statistical significance. The median age at which polyps were present was 53 years (range: 29–86 years).

Details of the trial

A total of 394 consecutive Caucasian women underwent hysteroscopy to assess abnormal uterine bleeding, infertility, cervical polyps, or abnormal sonographic patterns (e.g., postmenopausal endometrial thickness >5 mm, endometrial hyperechogenic spots). Of these women, 353 were included in the study, and demographic characteristics and data on diabetes, hypertension, and menopausal status were collected. Anthropometric parameters were also analyzed. When a polyp was detected, it was removed via office hysteroscopy, and histologic analysis was performed.

Small, asymptomatic uterine polyps may regress without treatment

Lieng M, Istre O, Sandvik L, Qvigstad E. Prevalence, 1-year regression rate, and clinical significance of asymptomatic endometrial polyps: cross-sectional study. J Minim Invasive Gynecol. 2009;16(4):465–471.

The treatment of asymptomatic polyps is controversial because their clinical consequences, malignant potential, and spontaneous regression rate are unknown. In this study, Lieng and colleagues prospectively estimated the prevalence and 1-year regression rate of incidentally diagnosed endometrial polyps in women 40 to 45 years old, as well as bleeding patterns and intensity.

They found polyps in 31 (12.1%) of 257 randomly selected women. At 1 year, the regression rate was 27%.

Details of the trial

At study inception, a standard 10-point visual analog scale was used to quantify each participant’s periodic bleeding, and a physical examination was performed. Transvaginal ultrasonography (US) and saline infusion sonography (SIS) were also performed. When a polyp was detected, researchers measured its length and used Doppler US to visualize the vessel feeding the polyp. An endometrial biopsy was also obtained.

The mean length of polyps was 14 mm (standard deviation [SD], 5.2 mm; 95% confidence interval [CI], 12.1–15.9; median, 13.4 mm; range, 6.7–28.7 mm), and the feeding vessel was identified for 22 of 31 polyps (71%). (For comparison, consider the findings of Clevenger and associates, who reported mean polyp diameters of 13.9 mm and 8.5 mm (P = .064), respectively, among women who had abnormal bleeding and women who did not.¹)

When researchers compared women who had polyps with those who had none, they found no significant differences in age, body mass index, blood pressure, gynecologic symptoms, menopausal status, use of hormone therapy, or use of the levonorgestrel-releasing intrauterine device (Mirena). Women who had endometrial polyps scored significantly higher, however, than women who did not on the visual analog scale for periodic bleeding and on the Pictorial Blood Assessment Chart—even when women who had myomas were excluded from the analysis. Although mean hemoglobin levels were similar between groups, women who had polyps had a significantly lower mean ferritin level (25 μg/L vs 41 μg/L; P = .05).

Polyps that persisted were larger from the start

Polyps regressed spontaneously in eight women, six of whom had the feeding vessel visualized at the initial consultation. Polyps that persisted after 12 months were significantly larger (mean polyp length, 15.1 mm; SD, 5.3 mm; 95% CI, 12.7–17.5) at study inception than were those that regressed (mean polyp length, 10.7 mm; SD, 3.9 mm; 95% CI, 7.5–14.0). Polyps that persisted beyond 1 year became significantly longer during follow-up, increasing from a mean length of 15.1 mm to 18.1 mm (SD, 7.9 mm; 95% CI, 0.7–5.3; P = .01).

Twenty of the 22 women who had persistent polyps underwent transcervical resection, one underwent laparoscopic supracervical hysterectomy, and one refused treatment. There were no complications.

Histology revealed that the polyps were benign in 16 women (80%), polypoid in two women (10%), and myomas in two women (10%). No atypical or malignant changes were observed in the polypectomy patients or among all participants.

A small, separate series (three patients) found all polyps to be 5 mm to 8 mm in length at detection, with a regression rate of 100% over several months.⁵

What variables signal a greater risk of malignancy?

Baiocchi G, Manci N, Pazzaglia M, et al. Malignancy in endometrial polyps: a 12-year experience. Am J Obstet Gynecol. 2009;201(5):462.e1–e4.

Gregoriou O, Konidaris S, Vrachnis N, et al. Clinical parameters linked with malignancy in endometrial polyps. Climacteric. 2009;12(5):454–458.

Wang JH, Zhao J, Lin J. Opportunities and risk factors for premalignant and malignant transformation of endometrial polyps: management strategies. J Minim Invasive Gynecol. 2010;17(1):53–58.

These three studies explore various aspects of a fundamental challenge: how to discriminate between polyps likely to undergo malignant transformation and those that will not.

The answer: Look for menopausal status, abnormal uterine bleeding, diabetes, obesity, and hypertension. Polyps larger than 1 cm also appear more likely to become malignant.

Details of the trials

In a retrospective study involving 1,242 women who had endometrial polyps, Baiocchi and colleagues identified 95.2% of the polyps as benign, 1.3% as premalignant, and 3.5% as malignant. Four clinical variables were significantly associated with premalignant and malignant features:

• age
• menopausal status
• abnormal uterine bleeding
• hypertension.

In their series of 516 cases, Gregoriou and associates found 96.9% of polyps to be benign, 1.2% to be premalignant, and 1.9% to be malignant. Four variables were associated with premalignant and malignant features:

• age above 60 years
• menopausal status
• obesity
• diabetes.

And in a study involving 766 patients, Wang and colleagues found 96.2% of polyps to be benign, 3.26% to involve hyperplasia with atypia, and 0.52% to be malignant. Among the variables associated with premalignant and malignant polyps were:

• polyp diameter larger than 1 cm
• menopausal status
• abnormal uterine bleeding.

Diagnosis may be trickier than you think, but treatment is straightforward

When a patient complains of abnormal uterine bleeding, evaluation often begins with transvaginal ultrasonography (US). Among the challenges of assessing the endometrium using US is the unreliability of endometrial thickness as a predictor of pathology. For example, Breitkopf and colleagues found that transvaginal US missed intracavitary lesions in one of six premenopausal women who had abnormal bleeding and an endometrial stripe thinner than 5 mm, for a sensitivity of 74%.⁶

In a separate study, Marello and colleagues used the combination of hysteroscopy and directed biopsy—the gold standard of diagnosis—to evaluate 212 postmenopausal women who had an endometrial thickness of 4 mm or less.⁷ (This parameter has been suggested as a cutoff for symptomatic postmenopausal women.⁸) Of these 212 women, 10% were found to have histologically confirmed intracavitary pathology (16 polyps and 4 submucous myomas).⁷ Among 13 symptomatic women in this study, three (23%) were found to have an endometrial polyp.⁷

These studies suggest that endometrial thickness alone should not be used to exclude benign endometrial pathology in symptomatic women, be they premenopausal or postmenopausal. No data back routine US to measure endometrial thickness in asymptomatic postmenopausal women.

Hysteroscopy and SIS are preferred

Both hysteroscopy and saline infusion sonography (SIS) have significantly better sensitivity and specificity in the diagnosis of intracavitary pathology than transvaginal US alone in women who have abnormal bleeding (FIGURE 1; VIDEOS 1, 2, AND 3 related to this article in the OBG Management Video Library at obgmanagement.com).^1,9 Hysteroscopy and SIS detect polyps with equal accuracy.¹⁰ However, hysteroscopy allows for removal of endometrial polyps and directed biopsy at the time of diagnosis.⁹

FIGURE 1 Imaging of polyps: go beyond transvaginal ultrasonography for optimal visualization

A. Hysteroscopic view of an endometrial polyp. B. The view with saline-infusion sonography.

In symptomatic women, resect the polyp

Polypectomy improves abnormal bleeding, according to a systematic review by Nathani and associates.¹¹ All studies included in the review, which involved follow-up intervals between 2 and 52 months, reported such an improvement.¹¹

When it is performed in the office, polypectomy offers several advantages over its inpatient counterpart:

• higher cost-effectiveness
• greater convenience
• avoidance of general anesthesia.

In both settings, it can be performed using mechanical or bipolar electrosurgical instrumentation (VIDEO 4).¹²

Segmental resection of the polyp while it is partially attached to the uterine wall is the optimal removal technique for large polyps (FIGURE 2). A grasping forceps can then be used to remove the polyp completely (VIDEO 5). Instruments such as a basket and snare are helpful in removing the polyp effectively.¹³

FIGURE 2 segmental resection of a polyp

During hysteroscopic polypectomy, the polyp is resected in segments while it is still partially attached to the endometrial wall.

Is histologic analysis of a polyp sufficient risk assessment?

Rahimi S, Marani C, Renzi C, Natale ME, Giovannini P, Zeloni R. Endometrial polyps and the risk of atypical hyperplasia on biopsies of unremarkable endometrium: a study on 694 patients with benign endometrial polyps. Int J Gynecol Pathol. 2009;28(6):522–528.

This study involved 694 consecutive patients who had benign endometrial polyps. Investigators sought to clarify the relationship between polyps and the underlying endometrium—specifically, to determine whether a polyp is a “circumscribed pathology of the endometrium or a polypoid expression of endometrial hyperplasia.” In describing the rationale for the study, the authors observe that the association between polyps and premalignant and malignant changes remains unclear.

Participants underwent hysteroscopy for removal of the polyps, at which time two biopsies of “unremarkable” endometrium, far from the base of the polyp, were also obtained.

Overall, endometrial hyperplasia without atypia was identified on hysteroscopically unremarkable endometrium in 18% of women, and atypia was identified in 7.3%. Among postmenopausal women, hyperplasia without atypia was identified in 21.6% of cases, atypia in 12%, and adenocarcinoma in 1.2%.

Multivariable analysis revealed that postmenopausal women who had polyps heavier than 1 g were 3.6 times more likely to have atypia (95% CI, 1.3–10.3). Among premenopausal women, the likelihood of atypia increased when the polyp weighed more than 0.4 g (odds ratio [OR], 3.5; 95% CI, 1.1–10.9) or the patient was older than 40 years (OR, 3.82; 95% CI, 1.1–13.2).

Don’t miss the videos that accompany this article!
They’re accessible in the OBG Management Video Library at obgmanagement.com

VIDEO 1: Saline-infusion sonographic imaging of a polyp

VIDEO 2: Hysteroscopic imaging of a polyp in a menopausal patient. A flexible 3-mm scope is used to assess an asymptomatic menopausal woman in whom an enlarged endometrial stripe was identified during earlier imaging.

VIDEO 3: Hysteroscopic imaging of a polyp and associated hyperplasia. A flexible 3-mm scope is used to evaluate a menopausal women who has uterine bleeding, revealing a 2-cm polyp.

VIDEO 4: Office polypectomy. An 8-mm, apparently benign polyp is removed from a premenopausal woman using continuous-flow operative hysteroscopy in an office setting.

VIDEO 5: Removal of a large polyp. A large (2 cm x 2.5 cm) polyp is removed in pieces, with the polyp partially attached to the endometrial wall, from a premenopausal woman who has abnormal bleeding.

More: Watch Dr. Brent Seibel perform OR-based hysteroscopic polypectomy

A busy gynecology practice sees more migraineurs in a day than a neurology practice sees in a month. But most of the migraineurs who visit gynecology offices have a chief complaint other than headache, and most leave without having mentioned the migraine—returning home to continue treating themselves with the same remedies their mothers used. Many of these women fail to seek a specific diagnosis, or treatment, until they develop chronic daily headaches or become unable to work.

Current theories of migraine implicate spreading cortical depression; trigeminal nerve activation with attendant vasodilation; and neurogenic inflammation.

Menstrual migraine (MM) is, arguably, the most common disabling condition encountered in women’s health. These migraines are more severe, last longer, and are more resistant to treatment than those that occur at other times in the cycle.^1,2 And, although headache specialists are adept at diagnosing migraine and prescribing a host of antiepileptic drugs and other migraine preventives, many of these specialists are uncomfortable manipulating the hormonal underpinnings of these “super migraines.”

This article outlines diagnostic criteria, treatment options, and preventive strategies, including hormonal therapy. In the process, it demonstrates how, in many cases, a gynecologist’s expertise in managing hormonal triggers may be the determinant of successful treatment.

Migraine prefers women

Migraine is the most common of all disabling headaches and afflicts 13% of the US population, but with a markedly skewed distribution: It preferentially attacks women in a 3:1 preponderance. The lifetime prevalence of migraine in women is 33%. Migraine strikes women primarily during reproductive years, when many women seek routine health care from a gynecologist rather than an internist or general practitioner.³

The migraine trigger appears to be estrogen withdrawal in susceptible persons—either during the natural menstrual cycle or as a result of cycling onto inert pills in an oral contraceptive (OC) regimen. A population study found that 39% of menstruating women experience headaches with menses, and almost one third of these headaches meet established criteria for MM or menstrual-related migraine.⁴

The distinction between these two entities, MM and menstrual-related migraine, is largely semantic. With MM, the menstrual attack is the patient’s only migraine. The broader term implies that she may have other attacks in addition to the menstrual ones. In this article, I’ve gathered both categories under the term MM, which also includes headaches that arise at the time of estrogen withdrawal associated with OC use.

Regardless of what we call them, these headaches have proved to be particularly vexing to headache medicine specialists who hesitate to address hormonal factors.

Diagnostic criteria are clear

Formal diagnosis of migraine requires that at least two of four signature characteristics plus at least one of two associated symptoms be present.⁵ The four characteristics are:

• moderate or severe pain
• throbbing
• unilateral location
• intensification of headache upon activity.

Any combination of two suffices for diagnosis—much to the astonishment of many patients who mistakenly believe that migraine must be severe or one-sided.

Associated symptoms include either nausea or both photophobia and phonophobia, the latter often signified by the simple preference to be in a dark, quiet room during an attack.

Untreated, migraine usually lasts between 4 and 72 hours.

A practical, clinical approach to diagnosis is to look for the episodic disabling headache. By disabling, I mean the presence of associated nausea or the need to stop one’s activities and lie down. A stable history of attacks with predictable menstrual association offers further confirmation.

Neck tension may be present

Symptoms of neck tension do not rule out migraine. In fact, neck pain is far more common at the time of treatment than is nausea, and migraineurs frequently describe neck pain that radiates forward during an attack.⁶

Treatment options begin with NSAIDs

Initial treatment of MM is no different than that of other migraines. Mild and moderate attacks can often be managed with nonsteroidal anti-inflammatory drugs (NSAIDs), which concomitantly treat comorbid dysmenorrhea.

A small randomized trial found that mefenamic acid (500 mg every 8 hours, as needed) was superior to placebo for acute MM, and a small crossover trial compared two formulations of diclofenac—50-mg sachets and tablets—with placebo in the treatment of acute migraine (TABLE). Almost one quarter of the patients taking the sachet form were pain-free at 2 hours, compared with 18.5% of patients taking the tablets and 11.7% of those who received placebo.^7,8

Injectable ketorolac is more potent and is proving to be as effective as or more effective than triptans or opioids for severe, persistent migraine.

When migraine is more severe, or when NSAIDs no longer suffice, treatment advances to migraine-specific medications—specifically, ergotamines and triptans.

Brand names of drugs mentioned in this article

Ergots and ergotamines have varied safety profiles in pregnancy

Dihydroergotamine belongs to the oldest family of migraine-specific drugs, although it is not widely used today. One reason may be that most migraineurs are women of reproductive age, and ergots are oxytocic and potential teratogens. Furthermore, they are not recommended in lactation. Ergotamines, however, are quite diverse, and some have more acceptable FDA use-in-pregnancy ratings. Injectable dihydroergotamine appears to be as effective as or less effective than triptans for migraine pain, but more effective than other drugs, such as NSAIDs or analgesics, for acute attacks.⁹ The nasal spray is more effective than placebo, but less effective than triptans.⁹

Triptans are relatively safe for women of reproductive age

Most information on use of triptans in pregnancy concerns sumatriptan. Like all triptans, it falls into Category C; however, its published profile is reassuring.^10-12 Data from Sweden on 2,027 first-trimester exposures show a 3.6% risk of birth defects (major and minor), compared with an identical 3.6% risk in the general Swedish population.¹² Sumatriptan is considerably less lipophilic than most of the drugs in its class and has been rated as compatible with breastfeeding.¹³

All triptans except eletriptan have been studied in the treatment of acute MM and found to be superior to placebo.¹⁴ These serotonin 1_B/1_D agonists are tailored therapy for acute migraine because, unlike analgesics, they address the underlying pathology of the attack, inhibiting the release of vasoactive peptides, promoting selective meningeal vasoconstriction, and blocking pain pathways in the brainstem. They also resolve the associated symptoms of nausea and photophobia. Triptans should not be used in the presence of untreated hypertension or vascular disease (cardiovascular, cerebrovascular, or peripheral vascular disease).

Watch for “rebound” headache. Like all acute treatment—be it a simple analgesic, NSAID, caffeine-containing compound, butalbital, ergotamine, or opioid—too-frequent use of triptans can produce medication-overuse headache, also referred to as rebound headache. In general, try to limit the use of agents for acute migraine, whether prescription or over-the-counter, to no more than 2 days a week to avoid this consequence. Also, be aware that some agents—notably, butalbital-containing compounds—may cause rebound headache, even when given as infrequently as 5 days of the month.¹⁵

Some preventive strategies are “nonspecific”

When treatment of acute MM is inadequate, management shifts toward prevention. Options include nonspecific (those that do not address the hormonal trigger) and specific (hormonal) strategies. Nonspecific strategies rely on the predictable onset of MM for proper timing of the therapeutic intervention; therefore, women who have irregular cycles are not good candidates for this approach. The presence of comorbid conditions such as dysmenorrhea, menometrorrhagia, and endometriosis may argue for early adoption of specific strategies that improve both conditions at once.

Nonhormonal (nonspecific) therapy NSAIDs may provide some relief. In one small study, naproxen sodium (550 mg twice daily) was administered for 2 weeks, beginning 1 week before the anticipated onset of menses. It modestly reduced the overall duration and severity of menstrual migraine¹⁶—a benefit that must be weighed against the risk of adverse events and the more recent FDA black box warning about its potentially serious heart and gastrointestinal risks.

Triptan regimens may prevent MM. Several triptans have been investigated for prevention of MM. A small, open-label study evaluated the use of oral sumatriptan (25 mg three times daily), beginning 2 to 3 days before the anticipated onset of MM and continuing for 5 days.¹⁷ Menstrual attacks were prevented in just over 50% of cases.

Naratriptan (1 mg twice daily), beginning 2 days before anticipated menses and continuing for 5 days, reduced the number of MM attacks by 50%, with side effects comparable to placebo.¹⁸ A higher dosage (2.5 mg twice daily) did not prove to be superior to placebo, for unexplained reasons.

In a similar trial, frovatriptan (2.5 mg daily or 2.5 mg twice daily), beginning 2 days before the anticipated onset of MM and continuing for 6 days, prevented at least 50% of MM.¹⁹ Twice-daily dosing was superior to daily administration.

Magnesium may shorten MM. In one small study, oral magnesium (360 mg daily), beginning on the 15^th day of the cycle and continuing through the menses, shortened the duration of MM and improved menstrual complaints better than placebo.²⁰

Hormonal therapy

The goal of hormonal therapy is to eliminate or sufficiently minimize the premenstrual decline in estrogen that is believed to precipitate MM.²¹ An observational study of 229 women found that hormonal strategies prevented MM in 73% of cases (81% when taken as directed).²²

It is fortunate that MM, by definition, is migraine without aura, because the use of combined OCs has been controversial in the setting of migraine with aura. International studies have reported a small but increased risk of stroke associated with their use, although a subset of women with migraine had no increased risk.^23-25 In contrast, no U.S. study since 1975 has found an increased risk of stroke associated with the use of OCs.²⁶ One large domestic study reviewed 3.6 million woman-years of use and found no increased risk of ischemic stroke with the low-dose OCs currently available, nor did a pooled analysis of U.S. studies.^27,28

The discrepancy is likely explained by the strong relative contraindication, in the United States, to the use of OCs in smokers older than 35 years (the smoking prevalence in most of the European case series was more than 50%), as well as the more prevalent use of high-dose OCs in the international studies. High-dose pills were implicated in the majority of stroke cases in the World Health Organization (WHO) study, but were used by only 0.7% of cases and controls in the pooled U.S. studies.²³ Nevertheless, both ACOG and WHO concluded that the risk of OC use usually outweighs the benefit in women older than 35 years whose migraines are complicated by focal neurologic deficits.

Extended-cycle OCs may offer a lengthy reprieve. Regimens that forego monthly withdrawal bleeds and provide extended administration of active pills can afford migraineurs a lengthy reprieve from MM.²⁹ Breakthrough bleeding is the most common side effect but tends to decrease over time. It is preferable for the patient to take the pill at bedtime to avoid a drug nadir during susceptible stages of sleep, which may be associated with migraine generation. It also is prudent to avoid concomitant administration of drugs that might increase the rate of hepatic metabolism of estrogen (e.g., a high dosage of topiramate) and lead to a more rapid decline in concentration.

When using an extended-cycle regimen that allows for periodic withdrawal bleeds, give the patient supplemental estrogen during the withdrawal week if the decline in ethinyl estradiol (EE) exceeds 10 μg, a threshold that appears to elicit MM in susceptible women.^21,22 For example, with an extended-cycle regimen that contains 30 μg of EE, it may be necessary to add 20 μg of estrogen during the placebo week to prevent MM, whereas an extended-cycle regimen that contains 20 μg EE, declining to 10 μg during the 13^th week, would be adequate as packaged.

21/7 hormonal OCs may require supplemental estrogen to prevent MM. The late luteal-phase decline in estradiol concentration during the natural menstrual cycle is equivalent to the decline experienced with the transition to inert pills in 20- to 25-μg EE formulations. Therefore, these regimens confer the same risk of MM as the woman’s natural cycle. Products that contain incrementally higher dosages of EE (30, 35, and 50 μg) confer an increasingly greater risk of MM. Supplementation of estrogen during the placebo week may prevent MM by limiting the decline in estrogen.

In a small, open-label study, women took an OC containing 20 μg EE at bedtime on days 1 to 21 of the cycle, followed by 0.9 mg of conjugated equine estrogens (CEE) on days 22 to 28. All patients reported a reduction in migraine frequency of at least 50%, with a mean reduction of 78%.³⁰

Parenteral options can be created utilizing a transdermal 20-μg EE/norelgestromin patch or a 15-μg EE/etonogestrel vaginal ring. With the first approach, I recommend adding a 0.1-mg EE patch during the withdrawal week to prevent MM. With the latter, a 0.075-mg EE patch may be used during the week following ring removal.³¹

Consider a menstrually targeted estrogen supplement. Some women who have contraindications to use of an OC may still be candidates for targeted strategies using lower dosages of supplemental estrogen rather than a combination OC. Among the options is perimenstrual administration of an estradiol patch or gel.³² One study found a 0.1-mg estrogen patch (worn 7 days and applied just before the expected onset of menses) to be effective, but lower dosages (0.025 to 0.05 mg EE) were not.³³ Daily application of 1.5 mg estradiol gel for 7 days, beginning before the onset of MM, was also effective in preventing MM.^34-36

With targeted strategies, timing is critical; if estrogen is begun too early, the incidence of migraine may rise after cessation.³⁶

GnRH agonists may benefit women who have isolated MM. Administration of gonadotropin-releasing hormone (GnRH) agonists has been shown to ease MM, but the baseline frequency of headaches appears to be an important variable in the success of these agents.^37-39 Leuprolide acetate markedly diminished MM in women who seldom experienced headaches other than their menstrual attacks.³⁷ It did not benefit women who experienced migraine in the setting of chronic headache (≥15 days/month), probably because chronic migraine is influenced by other variables, such as nonrestorative sleep and overuse of medication.^15,38,40

No evidence that progestin-only contraceptives are effective. Although these agents have been proposed for treatment of migraine, evidence of their benefit is lacking. It is my opinion that, more often than not, they exacerbate migraine.

A busy gynecology practice sees more migraineurs in a day than a neurology practice sees in a month. But most of the migraineurs who visit gynecology offices have a chief complaint other than headache, and most leave without having mentioned the migraine—returning home to continue treating themselves with the same remedies their mothers used. Many of these women fail to seek a specific diagnosis, or treatment, until they develop chronic daily headaches or become unable to work.

Current theories of migraine implicate spreading cortical depression; trigeminal nerve activation with attendant vasodilation; and neurogenic inflammation.

Menstrual migraine (MM) is, arguably, the most common disabling condition encountered in women’s health. These migraines are more severe, last longer, and are more resistant to treatment than those that occur at other times in the cycle.^1,2 And, although headache specialists are adept at diagnosing migraine and prescribing a host of antiepileptic drugs and other migraine preventives, many of these specialists are uncomfortable manipulating the hormonal underpinnings of these “super migraines.”

This article outlines diagnostic criteria, treatment options, and preventive strategies, including hormonal therapy. In the process, it demonstrates how, in many cases, a gynecologist’s expertise in managing hormonal triggers may be the determinant of successful treatment.

Migraine prefers women

Migraine is the most common of all disabling headaches and afflicts 13% of the US population, but with a markedly skewed distribution: It preferentially attacks women in a 3:1 preponderance. The lifetime prevalence of migraine in women is 33%. Migraine strikes women primarily during reproductive years, when many women seek routine health care from a gynecologist rather than an internist or general practitioner.³

The migraine trigger appears to be estrogen withdrawal in susceptible persons—either during the natural menstrual cycle or as a result of cycling onto inert pills in an oral contraceptive (OC) regimen. A population study found that 39% of menstruating women experience headaches with menses, and almost one third of these headaches meet established criteria for MM or menstrual-related migraine.⁴

The distinction between these two entities, MM and menstrual-related migraine, is largely semantic. With MM, the menstrual attack is the patient’s only migraine. The broader term implies that she may have other attacks in addition to the menstrual ones. In this article, I’ve gathered both categories under the term MM, which also includes headaches that arise at the time of estrogen withdrawal associated with OC use.

Regardless of what we call them, these headaches have proved to be particularly vexing to headache medicine specialists who hesitate to address hormonal factors.

Diagnostic criteria are clear

Formal diagnosis of migraine requires that at least two of four signature characteristics plus at least one of two associated symptoms be present.⁵ The four characteristics are:

• moderate or severe pain
• throbbing
• unilateral location
• intensification of headache upon activity.

Any combination of two suffices for diagnosis—much to the astonishment of many patients who mistakenly believe that migraine must be severe or one-sided.

Associated symptoms include either nausea or both photophobia and phonophobia, the latter often signified by the simple preference to be in a dark, quiet room during an attack.

Untreated, migraine usually lasts between 4 and 72 hours.

A practical, clinical approach to diagnosis is to look for the episodic disabling headache. By disabling, I mean the presence of associated nausea or the need to stop one’s activities and lie down. A stable history of attacks with predictable menstrual association offers further confirmation.

Neck tension may be present

Symptoms of neck tension do not rule out migraine. In fact, neck pain is far more common at the time of treatment than is nausea, and migraineurs frequently describe neck pain that radiates forward during an attack.⁶

Treatment options begin with NSAIDs

Initial treatment of MM is no different than that of other migraines. Mild and moderate attacks can often be managed with nonsteroidal anti-inflammatory drugs (NSAIDs), which concomitantly treat comorbid dysmenorrhea.

A small randomized trial found that mefenamic acid (500 mg every 8 hours, as needed) was superior to placebo for acute MM, and a small crossover trial compared two formulations of diclofenac—50-mg sachets and tablets—with placebo in the treatment of acute migraine (TABLE). Almost one quarter of the patients taking the sachet form were pain-free at 2 hours, compared with 18.5% of patients taking the tablets and 11.7% of those who received placebo.^7,8

Injectable ketorolac is more potent and is proving to be as effective as or more effective than triptans or opioids for severe, persistent migraine.

When migraine is more severe, or when NSAIDs no longer suffice, treatment advances to migraine-specific medications—specifically, ergotamines and triptans.

Brand names of drugs mentioned in this article

Ergots and ergotamines have varied safety profiles in pregnancy

Dihydroergotamine belongs to the oldest family of migraine-specific drugs, although it is not widely used today. One reason may be that most migraineurs are women of reproductive age, and ergots are oxytocic and potential teratogens. Furthermore, they are not recommended in lactation. Ergotamines, however, are quite diverse, and some have more acceptable FDA use-in-pregnancy ratings. Injectable dihydroergotamine appears to be as effective as or less effective than triptans for migraine pain, but more effective than other drugs, such as NSAIDs or analgesics, for acute attacks.⁹ The nasal spray is more effective than placebo, but less effective than triptans.⁹

Triptans are relatively safe for women of reproductive age

Most information on use of triptans in pregnancy concerns sumatriptan. Like all triptans, it falls into Category C; however, its published profile is reassuring.^10-12 Data from Sweden on 2,027 first-trimester exposures show a 3.6% risk of birth defects (major and minor), compared with an identical 3.6% risk in the general Swedish population.¹² Sumatriptan is considerably less lipophilic than most of the drugs in its class and has been rated as compatible with breastfeeding.¹³

All triptans except eletriptan have been studied in the treatment of acute MM and found to be superior to placebo.¹⁴ These serotonin 1_B/1_D agonists are tailored therapy for acute migraine because, unlike analgesics, they address the underlying pathology of the attack, inhibiting the release of vasoactive peptides, promoting selective meningeal vasoconstriction, and blocking pain pathways in the brainstem. They also resolve the associated symptoms of nausea and photophobia. Triptans should not be used in the presence of untreated hypertension or vascular disease (cardiovascular, cerebrovascular, or peripheral vascular disease).

Watch for “rebound” headache. Like all acute treatment—be it a simple analgesic, NSAID, caffeine-containing compound, butalbital, ergotamine, or opioid—too-frequent use of triptans can produce medication-overuse headache, also referred to as rebound headache. In general, try to limit the use of agents for acute migraine, whether prescription or over-the-counter, to no more than 2 days a week to avoid this consequence. Also, be aware that some agents—notably, butalbital-containing compounds—may cause rebound headache, even when given as infrequently as 5 days of the month.¹⁵

Some preventive strategies are “nonspecific”

When treatment of acute MM is inadequate, management shifts toward prevention. Options include nonspecific (those that do not address the hormonal trigger) and specific (hormonal) strategies. Nonspecific strategies rely on the predictable onset of MM for proper timing of the therapeutic intervention; therefore, women who have irregular cycles are not good candidates for this approach. The presence of comorbid conditions such as dysmenorrhea, menometrorrhagia, and endometriosis may argue for early adoption of specific strategies that improve both conditions at once.

Nonhormonal (nonspecific) therapy NSAIDs may provide some relief. In one small study, naproxen sodium (550 mg twice daily) was administered for 2 weeks, beginning 1 week before the anticipated onset of menses. It modestly reduced the overall duration and severity of menstrual migraine¹⁶—a benefit that must be weighed against the risk of adverse events and the more recent FDA black box warning about its potentially serious heart and gastrointestinal risks.

Triptan regimens may prevent MM. Several triptans have been investigated for prevention of MM. A small, open-label study evaluated the use of oral sumatriptan (25 mg three times daily), beginning 2 to 3 days before the anticipated onset of MM and continuing for 5 days.¹⁷ Menstrual attacks were prevented in just over 50% of cases.

Naratriptan (1 mg twice daily), beginning 2 days before anticipated menses and continuing for 5 days, reduced the number of MM attacks by 50%, with side effects comparable to placebo.¹⁸ A higher dosage (2.5 mg twice daily) did not prove to be superior to placebo, for unexplained reasons.

In a similar trial, frovatriptan (2.5 mg daily or 2.5 mg twice daily), beginning 2 days before the anticipated onset of MM and continuing for 6 days, prevented at least 50% of MM.¹⁹ Twice-daily dosing was superior to daily administration.

Magnesium may shorten MM. In one small study, oral magnesium (360 mg daily), beginning on the 15^th day of the cycle and continuing through the menses, shortened the duration of MM and improved menstrual complaints better than placebo.²⁰

Hormonal therapy

The goal of hormonal therapy is to eliminate or sufficiently minimize the premenstrual decline in estrogen that is believed to precipitate MM.²¹ An observational study of 229 women found that hormonal strategies prevented MM in 73% of cases (81% when taken as directed).²²

It is fortunate that MM, by definition, is migraine without aura, because the use of combined OCs has been controversial in the setting of migraine with aura. International studies have reported a small but increased risk of stroke associated with their use, although a subset of women with migraine had no increased risk.^23-25 In contrast, no U.S. study since 1975 has found an increased risk of stroke associated with the use of OCs.²⁶ One large domestic study reviewed 3.6 million woman-years of use and found no increased risk of ischemic stroke with the low-dose OCs currently available, nor did a pooled analysis of U.S. studies.^27,28

The discrepancy is likely explained by the strong relative contraindication, in the United States, to the use of OCs in smokers older than 35 years (the smoking prevalence in most of the European case series was more than 50%), as well as the more prevalent use of high-dose OCs in the international studies. High-dose pills were implicated in the majority of stroke cases in the World Health Organization (WHO) study, but were used by only 0.7% of cases and controls in the pooled U.S. studies.²³ Nevertheless, both ACOG and WHO concluded that the risk of OC use usually outweighs the benefit in women older than 35 years whose migraines are complicated by focal neurologic deficits.

Extended-cycle OCs may offer a lengthy reprieve. Regimens that forego monthly withdrawal bleeds and provide extended administration of active pills can afford migraineurs a lengthy reprieve from MM.²⁹ Breakthrough bleeding is the most common side effect but tends to decrease over time. It is preferable for the patient to take the pill at bedtime to avoid a drug nadir during susceptible stages of sleep, which may be associated with migraine generation. It also is prudent to avoid concomitant administration of drugs that might increase the rate of hepatic metabolism of estrogen (e.g., a high dosage of topiramate) and lead to a more rapid decline in concentration.

When using an extended-cycle regimen that allows for periodic withdrawal bleeds, give the patient supplemental estrogen during the withdrawal week if the decline in ethinyl estradiol (EE) exceeds 10 μg, a threshold that appears to elicit MM in susceptible women.^21,22 For example, with an extended-cycle regimen that contains 30 μg of EE, it may be necessary to add 20 μg of estrogen during the placebo week to prevent MM, whereas an extended-cycle regimen that contains 20 μg EE, declining to 10 μg during the 13^th week, would be adequate as packaged.

21/7 hormonal OCs may require supplemental estrogen to prevent MM. The late luteal-phase decline in estradiol concentration during the natural menstrual cycle is equivalent to the decline experienced with the transition to inert pills in 20- to 25-μg EE formulations. Therefore, these regimens confer the same risk of MM as the woman’s natural cycle. Products that contain incrementally higher dosages of EE (30, 35, and 50 μg) confer an increasingly greater risk of MM. Supplementation of estrogen during the placebo week may prevent MM by limiting the decline in estrogen.

In a small, open-label study, women took an OC containing 20 μg EE at bedtime on days 1 to 21 of the cycle, followed by 0.9 mg of conjugated equine estrogens (CEE) on days 22 to 28. All patients reported a reduction in migraine frequency of at least 50%, with a mean reduction of 78%.³⁰

Parenteral options can be created utilizing a transdermal 20-μg EE/norelgestromin patch or a 15-μg EE/etonogestrel vaginal ring. With the first approach, I recommend adding a 0.1-mg EE patch during the withdrawal week to prevent MM. With the latter, a 0.075-mg EE patch may be used during the week following ring removal.³¹

Consider a menstrually targeted estrogen supplement. Some women who have contraindications to use of an OC may still be candidates for targeted strategies using lower dosages of supplemental estrogen rather than a combination OC. Among the options is perimenstrual administration of an estradiol patch or gel.³² One study found a 0.1-mg estrogen patch (worn 7 days and applied just before the expected onset of menses) to be effective, but lower dosages (0.025 to 0.05 mg EE) were not.³³ Daily application of 1.5 mg estradiol gel for 7 days, beginning before the onset of MM, was also effective in preventing MM.^34-36

With targeted strategies, timing is critical; if estrogen is begun too early, the incidence of migraine may rise after cessation.³⁶

GnRH agonists may benefit women who have isolated MM. Administration of gonadotropin-releasing hormone (GnRH) agonists has been shown to ease MM, but the baseline frequency of headaches appears to be an important variable in the success of these agents.^37-39 Leuprolide acetate markedly diminished MM in women who seldom experienced headaches other than their menstrual attacks.³⁷ It did not benefit women who experienced migraine in the setting of chronic headache (≥15 days/month), probably because chronic migraine is influenced by other variables, such as nonrestorative sleep and overuse of medication.^15,38,40

No evidence that progestin-only contraceptives are effective. Although these agents have been proposed for treatment of migraine, evidence of their benefit is lacking. It is my opinion that, more often than not, they exacerbate migraine.

A busy gynecology practice sees more migraineurs in a day than a neurology practice sees in a month. But most of the migraineurs who visit gynecology offices have a chief complaint other than headache, and most leave without having mentioned the migraine—returning home to continue treating themselves with the same remedies their mothers used. Many of these women fail to seek a specific diagnosis, or treatment, until they develop chronic daily headaches or become unable to work.

Current theories of migraine implicate spreading cortical depression; trigeminal nerve activation with attendant vasodilation; and neurogenic inflammation.

Menstrual migraine (MM) is, arguably, the most common disabling condition encountered in women’s health. These migraines are more severe, last longer, and are more resistant to treatment than those that occur at other times in the cycle.^1,2 And, although headache specialists are adept at diagnosing migraine and prescribing a host of antiepileptic drugs and other migraine preventives, many of these specialists are uncomfortable manipulating the hormonal underpinnings of these “super migraines.”

This article outlines diagnostic criteria, treatment options, and preventive strategies, including hormonal therapy. In the process, it demonstrates how, in many cases, a gynecologist’s expertise in managing hormonal triggers may be the determinant of successful treatment.

Migraine prefers women

Migraine is the most common of all disabling headaches and afflicts 13% of the US population, but with a markedly skewed distribution: It preferentially attacks women in a 3:1 preponderance. The lifetime prevalence of migraine in women is 33%. Migraine strikes women primarily during reproductive years, when many women seek routine health care from a gynecologist rather than an internist or general practitioner.³

The migraine trigger appears to be estrogen withdrawal in susceptible persons—either during the natural menstrual cycle or as a result of cycling onto inert pills in an oral contraceptive (OC) regimen. A population study found that 39% of menstruating women experience headaches with menses, and almost one third of these headaches meet established criteria for MM or menstrual-related migraine.⁴

The distinction between these two entities, MM and menstrual-related migraine, is largely semantic. With MM, the menstrual attack is the patient’s only migraine. The broader term implies that she may have other attacks in addition to the menstrual ones. In this article, I’ve gathered both categories under the term MM, which also includes headaches that arise at the time of estrogen withdrawal associated with OC use.

Regardless of what we call them, these headaches have proved to be particularly vexing to headache medicine specialists who hesitate to address hormonal factors.

Diagnostic criteria are clear

Formal diagnosis of migraine requires that at least two of four signature characteristics plus at least one of two associated symptoms be present.⁵ The four characteristics are:

• moderate or severe pain
• throbbing
• unilateral location
• intensification of headache upon activity.

Any combination of two suffices for diagnosis—much to the astonishment of many patients who mistakenly believe that migraine must be severe or one-sided.

Associated symptoms include either nausea or both photophobia and phonophobia, the latter often signified by the simple preference to be in a dark, quiet room during an attack.

Untreated, migraine usually lasts between 4 and 72 hours.

A practical, clinical approach to diagnosis is to look for the episodic disabling headache. By disabling, I mean the presence of associated nausea or the need to stop one’s activities and lie down. A stable history of attacks with predictable menstrual association offers further confirmation.

Neck tension may be present

Symptoms of neck tension do not rule out migraine. In fact, neck pain is far more common at the time of treatment than is nausea, and migraineurs frequently describe neck pain that radiates forward during an attack.⁶

Treatment options begin with NSAIDs

Initial treatment of MM is no different than that of other migraines. Mild and moderate attacks can often be managed with nonsteroidal anti-inflammatory drugs (NSAIDs), which concomitantly treat comorbid dysmenorrhea.

A small randomized trial found that mefenamic acid (500 mg every 8 hours, as needed) was superior to placebo for acute MM, and a small crossover trial compared two formulations of diclofenac—50-mg sachets and tablets—with placebo in the treatment of acute migraine (TABLE). Almost one quarter of the patients taking the sachet form were pain-free at 2 hours, compared with 18.5% of patients taking the tablets and 11.7% of those who received placebo.^7,8

Injectable ketorolac is more potent and is proving to be as effective as or more effective than triptans or opioids for severe, persistent migraine.

When migraine is more severe, or when NSAIDs no longer suffice, treatment advances to migraine-specific medications—specifically, ergotamines and triptans.

Brand names of drugs mentioned in this article

Ergots and ergotamines have varied safety profiles in pregnancy

Dihydroergotamine belongs to the oldest family of migraine-specific drugs, although it is not widely used today. One reason may be that most migraineurs are women of reproductive age, and ergots are oxytocic and potential teratogens. Furthermore, they are not recommended in lactation. Ergotamines, however, are quite diverse, and some have more acceptable FDA use-in-pregnancy ratings. Injectable dihydroergotamine appears to be as effective as or less effective than triptans for migraine pain, but more effective than other drugs, such as NSAIDs or analgesics, for acute attacks.⁹ The nasal spray is more effective than placebo, but less effective than triptans.⁹

Triptans are relatively safe for women of reproductive age

Most information on use of triptans in pregnancy concerns sumatriptan. Like all triptans, it falls into Category C; however, its published profile is reassuring.^10-12 Data from Sweden on 2,027 first-trimester exposures show a 3.6% risk of birth defects (major and minor), compared with an identical 3.6% risk in the general Swedish population.¹² Sumatriptan is considerably less lipophilic than most of the drugs in its class and has been rated as compatible with breastfeeding.¹³

All triptans except eletriptan have been studied in the treatment of acute MM and found to be superior to placebo.¹⁴ These serotonin 1_B/1_D agonists are tailored therapy for acute migraine because, unlike analgesics, they address the underlying pathology of the attack, inhibiting the release of vasoactive peptides, promoting selective meningeal vasoconstriction, and blocking pain pathways in the brainstem. They also resolve the associated symptoms of nausea and photophobia. Triptans should not be used in the presence of untreated hypertension or vascular disease (cardiovascular, cerebrovascular, or peripheral vascular disease).

Watch for “rebound” headache. Like all acute treatment—be it a simple analgesic, NSAID, caffeine-containing compound, butalbital, ergotamine, or opioid—too-frequent use of triptans can produce medication-overuse headache, also referred to as rebound headache. In general, try to limit the use of agents for acute migraine, whether prescription or over-the-counter, to no more than 2 days a week to avoid this consequence. Also, be aware that some agents—notably, butalbital-containing compounds—may cause rebound headache, even when given as infrequently as 5 days of the month.¹⁵

Some preventive strategies are “nonspecific”

When treatment of acute MM is inadequate, management shifts toward prevention. Options include nonspecific (those that do not address the hormonal trigger) and specific (hormonal) strategies. Nonspecific strategies rely on the predictable onset of MM for proper timing of the therapeutic intervention; therefore, women who have irregular cycles are not good candidates for this approach. The presence of comorbid conditions such as dysmenorrhea, menometrorrhagia, and endometriosis may argue for early adoption of specific strategies that improve both conditions at once.

Nonhormonal (nonspecific) therapy NSAIDs may provide some relief. In one small study, naproxen sodium (550 mg twice daily) was administered for 2 weeks, beginning 1 week before the anticipated onset of menses. It modestly reduced the overall duration and severity of menstrual migraine¹⁶—a benefit that must be weighed against the risk of adverse events and the more recent FDA black box warning about its potentially serious heart and gastrointestinal risks.

Triptan regimens may prevent MM. Several triptans have been investigated for prevention of MM. A small, open-label study evaluated the use of oral sumatriptan (25 mg three times daily), beginning 2 to 3 days before the anticipated onset of MM and continuing for 5 days.¹⁷ Menstrual attacks were prevented in just over 50% of cases.

Naratriptan (1 mg twice daily), beginning 2 days before anticipated menses and continuing for 5 days, reduced the number of MM attacks by 50%, with side effects comparable to placebo.¹⁸ A higher dosage (2.5 mg twice daily) did not prove to be superior to placebo, for unexplained reasons.

In a similar trial, frovatriptan (2.5 mg daily or 2.5 mg twice daily), beginning 2 days before the anticipated onset of MM and continuing for 6 days, prevented at least 50% of MM.¹⁹ Twice-daily dosing was superior to daily administration.

Magnesium may shorten MM. In one small study, oral magnesium (360 mg daily), beginning on the 15^th day of the cycle and continuing through the menses, shortened the duration of MM and improved menstrual complaints better than placebo.²⁰

Hormonal therapy

The goal of hormonal therapy is to eliminate or sufficiently minimize the premenstrual decline in estrogen that is believed to precipitate MM.²¹ An observational study of 229 women found that hormonal strategies prevented MM in 73% of cases (81% when taken as directed).²²

It is fortunate that MM, by definition, is migraine without aura, because the use of combined OCs has been controversial in the setting of migraine with aura. International studies have reported a small but increased risk of stroke associated with their use, although a subset of women with migraine had no increased risk.^23-25 In contrast, no U.S. study since 1975 has found an increased risk of stroke associated with the use of OCs.²⁶ One large domestic study reviewed 3.6 million woman-years of use and found no increased risk of ischemic stroke with the low-dose OCs currently available, nor did a pooled analysis of U.S. studies.^27,28

The discrepancy is likely explained by the strong relative contraindication, in the United States, to the use of OCs in smokers older than 35 years (the smoking prevalence in most of the European case series was more than 50%), as well as the more prevalent use of high-dose OCs in the international studies. High-dose pills were implicated in the majority of stroke cases in the World Health Organization (WHO) study, but were used by only 0.7% of cases and controls in the pooled U.S. studies.²³ Nevertheless, both ACOG and WHO concluded that the risk of OC use usually outweighs the benefit in women older than 35 years whose migraines are complicated by focal neurologic deficits.

Extended-cycle OCs may offer a lengthy reprieve. Regimens that forego monthly withdrawal bleeds and provide extended administration of active pills can afford migraineurs a lengthy reprieve from MM.²⁹ Breakthrough bleeding is the most common side effect but tends to decrease over time. It is preferable for the patient to take the pill at bedtime to avoid a drug nadir during susceptible stages of sleep, which may be associated with migraine generation. It also is prudent to avoid concomitant administration of drugs that might increase the rate of hepatic metabolism of estrogen (e.g., a high dosage of topiramate) and lead to a more rapid decline in concentration.

When using an extended-cycle regimen that allows for periodic withdrawal bleeds, give the patient supplemental estrogen during the withdrawal week if the decline in ethinyl estradiol (EE) exceeds 10 μg, a threshold that appears to elicit MM in susceptible women.^21,22 For example, with an extended-cycle regimen that contains 30 μg of EE, it may be necessary to add 20 μg of estrogen during the placebo week to prevent MM, whereas an extended-cycle regimen that contains 20 μg EE, declining to 10 μg during the 13^th week, would be adequate as packaged.

21/7 hormonal OCs may require supplemental estrogen to prevent MM. The late luteal-phase decline in estradiol concentration during the natural menstrual cycle is equivalent to the decline experienced with the transition to inert pills in 20- to 25-μg EE formulations. Therefore, these regimens confer the same risk of MM as the woman’s natural cycle. Products that contain incrementally higher dosages of EE (30, 35, and 50 μg) confer an increasingly greater risk of MM. Supplementation of estrogen during the placebo week may prevent MM by limiting the decline in estrogen.

In a small, open-label study, women took an OC containing 20 μg EE at bedtime on days 1 to 21 of the cycle, followed by 0.9 mg of conjugated equine estrogens (CEE) on days 22 to 28. All patients reported a reduction in migraine frequency of at least 50%, with a mean reduction of 78%.³⁰

Parenteral options can be created utilizing a transdermal 20-μg EE/norelgestromin patch or a 15-μg EE/etonogestrel vaginal ring. With the first approach, I recommend adding a 0.1-mg EE patch during the withdrawal week to prevent MM. With the latter, a 0.075-mg EE patch may be used during the week following ring removal.³¹

Consider a menstrually targeted estrogen supplement. Some women who have contraindications to use of an OC may still be candidates for targeted strategies using lower dosages of supplemental estrogen rather than a combination OC. Among the options is perimenstrual administration of an estradiol patch or gel.³² One study found a 0.1-mg estrogen patch (worn 7 days and applied just before the expected onset of menses) to be effective, but lower dosages (0.025 to 0.05 mg EE) were not.³³ Daily application of 1.5 mg estradiol gel for 7 days, beginning before the onset of MM, was also effective in preventing MM.^34-36

With targeted strategies, timing is critical; if estrogen is begun too early, the incidence of migraine may rise after cessation.³⁶

GnRH agonists may benefit women who have isolated MM. Administration of gonadotropin-releasing hormone (GnRH) agonists has been shown to ease MM, but the baseline frequency of headaches appears to be an important variable in the success of these agents.^37-39 Leuprolide acetate markedly diminished MM in women who seldom experienced headaches other than their menstrual attacks.³⁷ It did not benefit women who experienced migraine in the setting of chronic headache (≥15 days/month), probably because chronic migraine is influenced by other variables, such as nonrestorative sleep and overuse of medication.^15,38,40

No evidence that progestin-only contraceptives are effective. Although these agents have been proposed for treatment of migraine, evidence of their benefit is lacking. It is my opinion that, more often than not, they exacerbate migraine.