Menopause and Hypothalamic-Pituitary Sensitivity to Estrogen

ORIGINAL CONTRIBUTION

Gerson Weiss, MD Context The onset of human menopause is thought to be caused solely by ovarian Joan H. Skurnick, PhD failure and oocyte depletion. However, clinical symptoms and certain recent data in Laura T. Goldsmith, PhD perimenopausal women suggest central nervous system involvement. Objective To determine if modifications of hypothalamic-pituitary response to es- Nanette F. Santoro, MD trogen feedback mechanisms occur in older reproductive-age women as a mecha- Susanna J. Park, MD nism of the onset of menopause. Design, Setting, and Participants The Study of Women’s Health Across the Na- Y THE YEAR 2030, MORE THAN 1.2 tion (SWAN) is a multiethnic observational cohort study of the menopausal transition billion women in the world will in 3302 women at 7 US sites. Of the subcohort of 840 women who participated in be at least 50 years old.1 This in- the Daily Hormone Study between 1997 and 1999, 680 women had evidence of lu- creasing proportion of the fe- teal activity. The remaining 160 women (19%) did not have luteal activity and are the maleB population will be experiencing the subject of this report. menopausal transition with its accom- Main Outcome Measures Daily urinary hormone levels of estrogen and proges- panying physiology and pathophysiol- terone metabolites, luteinizing hormone (LH), and follicle-stimulating hormone (FSH). ogy. Reproductive aging in women and Results Three groups of women were studied: those with estrogen increases with the hormonal changes that occur dur- an LH surge, those with estrogen increases without an LH surge, and those with nei- ing the onset of human menopause have ther. There were no differences in age or ethnicity among the 3 groups of women. been ascribed solely to ovarian failure Women in the third group (no increases) experienced more menopausal symptoms and oocyte depletion.2 However, in other (hot flashes) than did women in the other groups with estrogen increases. In older species, the central nervous system is the reproductive-age women, the frequent existence of anovulatory cycles with estrogen major regulator of age-related reproduc- peaks, equivalent to those that result in LH surges in younger women, yet in which no LH surges occur, indicates failure of estrogen-positive feedback on LH secretion. In tive dysfunction.3 other anovulatory cycles, follicular-phase estrogen levels did not lower LH secretion There are 4 events involving the hy- as occurs in cycles of younger women, indicating decreased estrogen-negative feed- pothalamic-pituitary-ovarian axis that back on LH secretion. control the human menstrual cycle: (1) Conclusion Our findings are compatible with hypothalamic-pituitary insensitivity to The secretion of follicle-stimulating estrogen in aging perimenopausal women. hormone (FSH), responsible for the de- JAMA. 2004;292:2991-2996 www.jama.com velopment of ovarian follicles and pro- duction of estradiol.4 Throughout the cycle, estrogen maintains low gonado- luteinization of the follicle, triggered by sites across the United States designed tropin levels via its negative feedback the LH surge, forming a corpus lu- to enhance understanding of the fac- effect on hypothalamic gonadotropin- teum. This is an ovarian response that tors that influence the health of women releasing hormone and consequently lu- results in progesterone secretion nec- of diverse race and ethnicity.10 The de- teinizing hormone (LH) and FSH se- essary for the establishment of a preg- tails of enrollment have been previ- cretion.5 (2) The FSH-induced increase nancy.9 ously reported.10 Race/ethnicity was in ovarian estrogen secretion to levels How these events may be altered dur- of sufficient strength and duration trig- ing menopausal transition has not been Author Affiliations: Department of Obstetrics, Gy- gering an LH surge (positive feed- well established. necology and Women’s Health, New Jersey Medical 6 School of UMDNJ, Newark, NJ (Drs Weiss, Skurnick, back). (3) The LH surge, a hypotha- Goldsmith, and Park) and Department of Obstetrics, lamic-pituitary response to the estrogen METHODS Gynecology and Women’s Health, Albert Einstein Col- lege of Medicine, New York, NY (Dr Santoro). stimulus. This positive feedback re- The Study of Women’s Health Across Corresponding Author: Laura T. Goldsmith, PhD, De- sponse of estrogen on LH secretion has the Nation (SWAN) is a multiethnic ob- partment of Obstetrics, Gynecology and Women’s Health, New Jersey Medical School of UMDNJ, 185 been used as a test of hypothalamic- servational cohort study of the meno- S Orange Ave, Newark, NJ 07103 (goldsmit@umdnj pituitary function.7,8 (4) Ovulation and pausal transition in 3302 women at 7 .edu).

Downloaded From: https://jamanetwork.com/ on 10/01/2021 MENOPAUSE AND HYPOTHALAMIC-PITUITARY SENSITIVITY TO ESTROGEN

self-determined by study participants dence of luteal activity based on a vali- times the mean baseline level (of 4 con- and was obtained by asking the follow- dated algorithm for pregnanediol gluc- secutive days starting 5 days earlier), ing open-ended question: “How would uronide.12 The algorithm locates the 5 (3) baseline mean plus 3 standard de- you describe your primary racial or eth- nadir days of pregnanediol glucuro- viations of the 4-day baseline levels, (4) nic group?” The responses then were nide in the follicular phase using mov- baseline mean plus 2 standard devia- categorized as Caucasian (white), Af- ing averages throughout the cycle. A tions of levels on days 2 through 6 african American, Chinese, Japanese, or 3-fold increase in pregnanediol gluc- ter the peak, and (5) 0.8 times the maxi- Hispanic. This study was approved by uronide concentrations above this na- mum level in that cycle. In addition, the all of the sites’ institutional review dir for at least 3 consecutive days was LH surge was required to culminate in board, and written informed consent considered evidence of luteal activity. a drop to no more than 1.5 times base- was obtained from each participant. The cycles of these women have been line within 6 days following the peak. A subcohort participated in the Daily reported previously.12 The remaining Cycles were classified by these algo- Hormone Study (DHS) from 1997 to 160 women (19% of the total 840 wom- rithms as falling into 1 of 3 distinct pat- 1999. The women in the DHS in- en) did not have luteal activity. One terns: (1) both estrogen increase and LH cluded 257 Caucasian (white) women, woman could not be subclassified due surge (coincident within 2 days), (2) 175 African American women, 152 Chi- to missing data points. The 159 remain- estrogen increase only, and (3) nei- nese women, 170 Japanese women, and ing women are the subjects of this re- ther. Visual inspection of the cycle data 86 women of Hispanic origin. The port. plots by 2 observers (G.W., J.H.S.) re- cohort has been described previ- All women in the SWAN DHS study vealed that 20 cycles had apparent es- ously.11,12 Inclusion criteria were age 42 had levels of FSH equal to or greater trogen increases that were missed by the to 52 years; an intact uterus and at least than those previously demonstrated in defining algorithm due to an increase one ovary; at least one menstrual pe- younger women throughout the en- too early in a short cycle to establish a riod in the prior 3 months; no use of tire cycle.12 This has been described pre- baseline or a slow decline. Four cycles sex-steroid hormones in the previous viously and is due to the decreased se- with an algorithmically determined es- 3 months; and not being pregnant. DHS cretion of ovarian inhibin in older trogen increase were reclassified to “nei- enrollees completed a daily collection women.14,15 Decreased gonadotropin se- ther.” Thus, of 159 cycles, 29 were clas- of morning voided urine for an entire cretion, as is found in some anovula- sified as “both,” 32 as “estrogen increase menstrual cycle ending in bleeding or tory cycles in premenopausal young only,” and 98 as “neither.” These cycle to 50 days, whichever came first. Dur- women in their teens to 30s, clearly did classifications were based solely on hor- ing the cycle that they collected daily not occur in these women. mone levels and not age or meno- urine specimens, DHS enrollees also Conceptually, an estrogen increase pausal symptoms. completed a daily diary, a question- is a high level compared with base- Analysis of variance was conducted naire in which they answered, once a line, in absolute terms and relative to to compare cycle classification groups day, whether they had experienced observed variability, followed by sub- on women’s ages and body mass in- within the preceding 24 hours any stantial decline. The specific criteria dex; ␹2 tests were conducted to com- trouble sleeping and any hot flashes or used here were adapted from previ- pare groups on ethnicity, reason for night sweats. ously established definitions for mid- ending collection, and experience of Urinary LH, FSH, the estradiol uri- reproductive age women.13 An estro- symptoms during the cycle. Rank- nary metabolites estrone conjugates gen increase was defined as an E1c level sum tests were conducted to compare (E1c), and the progesterone urinary me- of at least (1) 50 pg/mg creatinine, (2) E1c levels of groups 1 and 2 by cycle tabolite pregnanediol glucuronide were twice the baseline level (the mean of 5 day and to compare the 3 groups on measured using chemiluminescent as- consecutive days starting 9 days ear- women’s percentage of cycle days with says as described previously.11,12 Con- lier), and (3) 3 standard deviations of symptom occurrence. Reported P val- centrations were normalized for cre- the baseline levels above the baseline ues are 2-sided, without adjustment for atinine excretion. Previous studies have mean. In addition, the estrogen peak multiple comparisons. Statistical analy- demonstrated that urinary levels of was required to culminate in a drop to ses were performed with SAS version these hormones, collected and mea- no more than 1.5 times baseline at some 8.2 (SAS Institute, Cary, NC). Statisti- sured by the methods used herein, mir- time within the next 5 days. An LH cal significance was set at 2-sided ror serum hormone patterns during the surge is considered present when a high PϽ.05. menstrual cycle in eumenorrheic con- level is observed relative to baseline in trols so closely that patterns of serum absolute terms and in excess of day to RESULTS and urinary gonadotropins and sex ste- day variability both before and after the Women in the 3 classification catego- roids are superimposible.13 peak, established by a drop in levels. An ries presented here were compared by Of the 840 women who completed LH surge was defined as an LH level of ethnicity, age, and body mass index. the DHS study, 680 women had evi- at least (1) 6 mIU/mg creatinine, (2) 3 As shown in TABLE 1, there was no

Downloaded From: https://jamanetwork.com/ on 10/01/2021 MENOPAUSE AND HYPOTHALAMIC-PITUITARY SENSITIVITY TO ESTROGEN

significant correlation between cat- Table 1. Characteristics of Women by Cycle Category* egory of cycle and any of these charac- Cycle Category teristics. The results from the 29 individuals 1 (Estrogen Increase 2 (Estrogen 3 (Neither Estrogen P who had an estrogen increase fol- + LH Surge) Increase Only) Nor LH Increase) Value† lowed by an LH surge (group 1, “both”) No. of women 29 32 98 Women with menses by 21 (72) 24 (75) 41 (42) are shown in FIGURE 1, left panels, in cycle day 50, No. (%) which hormone levels are synchro- Age, mean (SD), y 48 (2.5) 49 (2.5) 49 (2.6) .10 nized to the LH peak. As seen in ovu- BMI, mean (SD) 27 (6.8) 30 (7.4) 29 (8.2) .59 latory cycles, these LH surges are ac- Abbreviations: BMI, body mass index, calculated as weight in kilograms divided by height in meters squared; LH, luteinizing hormone. companied by FSH surges. Hormone *Groups did not differ significantly by ethnicity (P=.22, ␹2, data not shown). Classification groups differ significantly in levels are similar to those of the previ- the reason for ending daily collection of samples: onset of menses or end of 50-day collection interval (PϽ.001, ␹2). ously reported women in DHS who had †Analysis of variance. luteal phases,12 indicating an adequate hypothalamic-pituitary response. In sented in FIGURE 2. For ease of illus- trogen nor LH increases (group 3) par- women in group 1, the follicle or fol- tration of this similarity (between E1c ticipated in the daily diary component licles that secreted sufficient estrogen levels in the women with gonadotro- of the study. For each symptom— to elicit an LH and FSH surge did not pin surges and those in women with- trouble sleeping or hot flashes or night luteinize as documented by lack of an out surges), the corresponding data for sweats—the percentage of days a woman increase in pregnanediol glucuronide group 2 from Figure 1 are also pro- reported that she experienced the symp- levels. This is a defect at the ovarian vided in Figure 2. These E1c patterns tom was computed as the total number level because hypothalamic-pituitary re- in women in groups 1 and 2 are also of days she reported presence of the sponses were similar to those of women similar to E1c secretion in the previ- symptom divided by the total number with apparently normal cycles. ously reported SWAN DHS study in of days she reported either yes or no for The results from the 32 women who women with luteal activity.12 Thus, in that symptom ϫ 100. These percent- had clear estrogen increases but no LH group 2 women there is adequate ovar- ages of days with positive reports for surges (group 2) are shown in Figure 1, ian response, but the LH surge, a hy- women in the 3 categories were com- middle panel. Hormone levels are syn- pothalamic-pituitary phenomenon, did pared by rank-sum tests. chronized to the E1c peak because no not occur in the presence of an estro- There were no differences in the per- LH surges were seen in these women. gen stimulus that is adequate to elicit centages of days with trouble sleeping The estrogen increases in these women an LH surge in ovulating women and among the 3 groups (TABLE 2). Com- were equivalent to those seen in ovu- in younger women. This clearly dem- parison of the percentages of days with latory women and to those in group 1. onstrates unresponsiveness of the hy- hot flashes or night sweats among the However, in contrast to those of group pothalamic-pituitary axis to an estro- 3 groups revealed significant group dif- 1 women, these estrogen increases did gen peak. Gonadotropin levels dropped ferences (Table 2). The percentages of not produce an LH surge. Compari- in the latter part of the cycles, likely due days with hot flashes or night sweats sons of the E1c levels at each cycle day to negative feedback from the estro- were significantly higher for group 3 in group 1 women with those in group gen increase. women than for either group 1 women 2 women were performed using rank- Hormone secretion in the 98 women (P=.01) or group 2 women (P=.02). sum tests. At all cycle days, E1c levels who had no estrogen peaks or LH The percentages of days with hot flashes in group 2 women who had no LH surges (group 3) are shown in Figure 1, or night sweats in group 1 women and surge were not lower than E1c levels right panels. Levels of LH are higher group 2 women did not differ (P=.73). in group 1 women who had LH surges. than those seen in either SWAN peri- In fact, no differences at any cycle day menopausal women with luteal phases COMMENT were observed, with the exception of or in the other 2 groups presented here. Hormone secretion patterns in older re- day –15, when E1c levels in group 2 Estrogen levels are comparable to those productive-age women demonstrate sig- women were higher (P=.04) than those in the early follicular phase of DHS lu- nificant alterations of hypothalamic- in women in group 1. teal women12 and group 1 or 2 women. pituitary feedback mechanisms in That the secretion pattern of E1c in Thus, group 3 women still have ovar- addition to decreased ovarian func- group 2 women was equivalent to that ian function but are unable to pro- tion. Cycles exist in which failure to of women in group 1 is not apparent duce an estrogen peak. mount an LH surge occurs in the face in Figure 1, in which hormone levels Twenty-eight of the 29 women with of adequate estrogen stimulation. These are synchronized to the LH surge as is LH surges (group 1), 31 of the 32 findings support the hypothesis that convention. Therefore, the E1c levels women without LH surges (group 2), there is a relative hypothalamic- synchronized to the E1c peak are pre- and 97 of the 98 women with neither es- pituitary insensitivity to estrogen in ag-

Downloaded From: https://jamanetwork.com/ on 10/01/2021 MENOPAUSE AND HYPOTHALAMIC-PITUITARY SENSITIVITY TO ESTROGEN

ing women that is manifested by both latory women and in group 1 women, LH, as is seen in postmenopausal positive and negative feedback mecha- fail to do so in group 3 women, who have women. Levels of estrogen capable of nisms. Estrogen levels and patterns that elevated LH in the presence of early fol- lowering LH in cycling women were not produce LH surges in younger women licular-phase levels of E1c. This situa- able to cause negative feedback of es- fail to do so in some older women. In tion may represent a later stage of the trogen on LH secretion. Because con- addition, levels of estrogen similar to menopausal transition because there is trol of FSH secretion is more complex those in younger women, which cause opening of the negative feedback loop than LH and includes major influences negative feedback of LH in normal ovu- between ovarian estrogen and pituitary by inhibins and activins, FSH is not a

Figure 1. Daily Urinary Hormone Levels in Anovulatory Older Reproductive-Age Women With and Without Estrogen Increases

Women With Both Estrogen Increase Women With Estrogen Increase Women With No Estrogen Increase and LH Surge (n = 29) but No LH Surge (n = 32) or LH Surge (n = 98) (Day of LH Peak = Day 0) (Day of Estrogen Peak = Day 0)

Estrone Conjugates 250

200

150

100

E1c, ng/mg Creatinine 50

Luteinizing Hormone

LH, mIU/mg Creatinine 5

Follicle-Stimulating Hormone

140

120

100

FSH, mIU/mg Creatinine 20

0 –21 –14 –7 0 7 14 21 –21 –14 –7 0 7 14 21 1 8 15 22 29 36 43 50 Day of Cycle Day of Cycle Day of Collection

First morning voided urine samples were collected by each woman daily for an entire menstrual cycle or to 50 days. Levels of estrone conjugates (E1c), luteinizing hormone (LH), follicle-stimulating hormone (FSH), and creatinine were measured in each sample. Hormone levels in each sample were normalized to creatinine values. Levels are presented as mean (SEM) plotted against cycle day, where day 0 is the day of the LH or estrogen peak for women with an estrogen increase. Mean (SEM) levels are plotted against day of collection for women with no estrogen increase.

Downloaded From: https://jamanetwork.com/ on 10/01/2021 MENOPAUSE AND HYPOTHALAMIC-PITUITARY SENSITIVITY TO ESTROGEN

good marker for estrogen-negative feed- these women12 do not get hot flashes, patients’ symptoms. An appreciation of back control of gonadotropin secre- the flashes may be due to opening of these changes may assist patients in un- tion. Decreased LH pulse frequency has the negative feedback loop of estrogen derstanding and dealing with their own been observed in the presence of nor- on gonadotropin secretion. menopause. Other symptoms, such as mal, midreproductive sex steroid lev- All women have hormonal changes mood and changes in affect, may also els in both the follicular16 and luteal17 during their menopausal transition, but have similar explanations. phases of the menstrual cycles of older not all women experience symptoms. Demonstration of hypothalamic- premenopausal women, supporting this It is likely that the gonadotropin con- pituitary insensitivity to estrogen in hypothesis as well. trol centers in the brain differ from the perimenopausal women shows that cer- A predominant hypothesis to ex- areas involved in the symptoms as- tain human menopausal responses are plain the onset of puberty is the occur- sessed in our study. These regions may similar to those of other species, in- rence of a gradual decrease in estrogen have different steroid sensitivities and cluding rats. Age-related decreased ex- sensitivity of the hypothalamic- control mechanisms, accounting for our pression of estrogen receptor ␤ in some pituitary axis, such that small levels of findings. However, certain changes in hypothalamic areas in the rat has been circulating estrogen, which suppress gonadotropin levels may be permis- described.21 These results suggest that gonadotropin secretion prepuberty, are sive to alterations in central nervous sys- the rat may be a useful model for the unable to do so in the pubertal transi- tem function, which result in symp- study of human central nervous sys- tion.18 Decreased sensitivity in later life toms. Physicians should be aware of the tem aging and that the mechanisms of may simply be a continuation of the central nervous system changes in- reproductive aging may be more simi- same pattern of progressive age- volved in the menopausal transition be- lar in diverse mammalian species than related estrogen insensitivity. Levels of cause these changes best explain their previously thought. LH are higher in perimenopausal women than in younger women, even Figure 2. Daily Urinary E1c Levels in Anovulatory Older Reproductive-Age Women With in the presence of estrogen concentra- Estrogen Increases tions that result in lower LH levels in 14 Women With Both Estrogen Increase Women With Estrogen Increase younger women. Symptoms such as and LH Surge (n = 29) but No LH Surge (n = 32) (Day of Estrogen Peak = Day 0) (Day of Estrogen Peak = Day 0) hot flashes and sleep disturbances 250 occur more commonly in perimenopausal women than in postmeno- 200 pausal women.19 Yet, the perimenopausal transition is a time when 150 circulating estrogen levels are equivalent or higher than levels observed in 100 younger women.14 Additionally, exog- enous estrogen is therapeutic in peri- E1c, ng/mg Creatinine 50 20 menopausal women. These observa- 0 tions are consistent with the hypothesis –21 –14 –7 0 7 14 21 –21 –14 –7 0 7 14 21 that a decrease in estrogen sensitivity Day of Cycle Day of Cycle occurs as women age through the meno- Comparison of E1c levels (estrone conjugates) in women who had an LH surge (group 1) (left panel) vs those pausal transition. who did not (group 2) (right panel). E1c levels (mean [SEM]) for women with both estrogen increases and LH We found no differences in re- surges are shown here, where day 0 is the day of maximum E1c. ported symptoms of sleep distur- Table 2. Daily Diary Symptom Reporting bances or vasomotor changes in women Cycle Category with different LH-positive feedback responses to equivalent circulating es- 1 (Estrogen 3 (Neither Increase 2 (Estrogen Estrogen Nor P trogen peaks (categories 1 and 2). Al- + LH Surge) Increase Only) LH Increase) Value though group 3 women did not differ No. of women 28 31 97 in the prevalence of sleep distur- Trouble sleeping, mean % 21 22 23 .93* bances, they had a significantly higher of days in cycle prevalence of hot flashes or night Hot flashes or night sweats sweats. While all women in this study Mean % of days in cycle 9 11 26 .009* Women not experiencing, 15 (54) 15 (48) 34 (35) .14† had similar baseline estrogen levels, the No. (%) group 3 women did not have mid- Abbreviation: LH, luteinizing hormone. cycle estrogen peaks. Since younger *Rank-sum test. †␹2 test. women with estrogen levels similar to

Downloaded From: https://jamanetwork.com/ on 10/01/2021 MENOPAUSE AND HYPOTHALAMIC-PITUITARY SENSITIVITY TO ESTROGEN

Author Contributions: Dr Weiss had full access to all Luke’s Medical Center, Chicago, Ill: Lynda Powell, PI Steering Committee: Chris Gallagher, Chair, 1995- of the data in the study and takes responsibility for (U01 AG012505); University of California, Davis/ 1997; Jenny Kelsey, Chair, 1997-2002; Susan Johnson, the integrity of the data and the accuracy of the data Kaiser: Ellen Gold, PI (U01 AG012554); University of Chair, 2002- present. analysis. California, Los Angeles: Gail Greendale, PI (U01 Funding/Support: The Study of Women’s Health Study concept and design: Weiss, Skurnick, Goldsmith, AG012539); University of Medicine and Dentistry/ Across the Nation (SWAN) was funded by the Na- Santoro, Park. New Jersey Medical School, Newark: Gerson Weiss, tional Institute on Aging. Acquisition of data: Weiss, Santoro. PI 1995-2004, Nanette Santoro, PI, 2004-present Role of the Sponsor: The National Institute on Aging Analysis and interpretation of data: Weiss, Skurnick, (U01 AG012535); and the University of Pittsburgh, funded the SWAN study. The current study used data Goldsmith, Santoro, Park. Pittsburgh, Pa: Karen Matthews, PI (U01 generated by the SWAN Daily Hormone Study (DHS). Drafting of the manuscript: Weiss, Skurnick, Goldsmith, AG012546). The DHS was designed and approved by the SWAN Park. NIH Program Office: National Institute on Aging, Steering Committee. The design and conduct of the cur- Critical revision of the manuscript for important in- Bethesda, Md: Sherry Sherman, 1994-present, Mar- rent study were a product of the authors and the samples tellectual content: Weiss, Skurnick, Goldsmith, Santoro, cia Ory, 1994-2001; National Institute of Nursing Re- were collected according to the SWAN DHS protocol Park. search, Bethesda, Md: Janice Phillips, 2002-present, and occurred at all 7 clinical sites. Data management Statistical analysis: Skurnick. Carole Hudgings, 1997-2002. was performed by the SWAN Coordinating Center. Obtained funding: Weiss, Skurnick, Santoro. Central Laboratory: University of Michigan, Ann Ar- Analysis and interpretation of the data were accom- Administrative, technical, or material support: Weiss, bor: Rees Midgley, PI 1995-2000, Daniel McCon- plished by the authors. Preparation of the manuscript Goldsmith, Santoro, Park. nell, 2000-present (U01 AG012495, Central Ligand was performed by the authors. The SWAN Presenta- Study supervision: Weiss, Santoro. Assay Satellite Services). tion and Publication Committee reviewed and approved SWAN Clinical Centers: University of Michigan, Ann Coordinating Centers: University of Pittsburgh, Pitts- the concept, study plan, and manuscript draft for con- Arbor: MaryFran Sowers, PI (U01 NR04061); Massa- burgh, Pa: Kim Sutton-Tyrrell, PI (U01 AG012546) sistency with the SWAN study and appropriateness of chusetts General Hospital, Boston: Robert Neer, PI 2001-present and New England Research Institutes, the current work in accordance with SWAN policy. 1995-1999, Joel Finkelstein, PI 1999- present (U01 Watertown, Mass: Sonja McKinlay, PI (U01 Acknowledgment: We thank the study staff at each AG012531); Rush University, Rush-Presbyterian-St AG012553), 1995-2001. site and all the women who participated in SWAN.

REFERENCES 1. World Health Organization. Research on the Meno- 9. Vande Wiele RL, Bogumil J, Dyrenfurth I, et al. 15. Burger HG, Dudley EC, Hopper JL, et al. The en- pause in the 1990s: Report of a WHO Scientific Group. Mechanisms regulating the menstrual cycle in women. docrinology of the menopausal transition: a cross- Geneva, Switzerland: WHO; 1996. WHO Technical Recent Prog Horm Res. 1970;26:63-103. sectional study of a population-based sample. J Clin Report Series 866. 10. Sowers MF, Crawford S, Sternfeld B, et al. De- Endocrinol Metab. 1995;80:3537-3545. 2. Berkow R. The Merck Manual of Diagnosis and sign, survey, sampling and recruitment methods of 16. Matt DW, Kauma SW, Pincus SM, et al. Charac- Therapy. Whitehouse Station, NJ: Merck & Co; 1987. SWAN: a multi-center, multi-ethnic, community- teristics of luteinizing hormone secretion in younger 3. Wise PM, Krajnak KM, Kashon ML. Menopause: based cohort study of women and the menopausal versus older premenopausal women. Am J Obstet the aging of multiple pacemakers. Science. 1996;273: transition. In: Lobo RA, Kelsey J, Marcus R, eds. Meno- Gynecol. 1998;178:504-510. 67-70. pause: Biology and Pathobiology. San Diego, Calif: 17. Reame NE, Kelch RP, Beitins IZ, et al. Age effects 4. Hillier SG, Reichert LE, Van Hall EV. Control of pre- Academic Press; 2000:175-188. on follicle-stimulating hormone and pulsatile lutein- ovulatory follicular estrogen biosynthesis in the hu- 11. Santoro N, Crawford SL, Allsworth JE, et al. As- izing hormone secretion across the menstrual cycle of man ovary. J Clin Endocrinol Metab. 1981;52:847-856. sessing menstrual cycles with urinary hormone assay. premenopausal women. J Clin Endocrinol Metab. 5. Yen SS, Tsai CC, Vandenberg G, Rebar R. Gonado- Am J Physiol. 2003;284:E521-E530. 1996;81:1512-1518. tropin dynamics in patients with gonadal dysgenesis: 12. Santoro N, Lasley B, McConnell D, et al. Body size 18. Grumbach MM, Kaplan SL. Puberty. In: Grum- a model for the study of gonadotropin regulation. and ethnicity are associated with menstrual cycle al- bach MM, Sizonenko PC, Aubert ML, eds. Control of J Clin Endocrinol Metab. 1972;35:897-904. terations in women in the early menopausal transi- the Onset of Puberty. Baltimore, Md: Williams & 6. Young JR, Jaffe RB. Strength-duration character- tion: the Study of Women’s health Across the Nation Wilkins; 1990:1-68. istics of estrogen effects on gonadotropin response to (SWAN) Daily Hormone Study. J Clin Endocrinol 19. Gold EB, Sternfeld B, Kelsey JL, et al. Relation of gonadotropin-releasing hormone in women, II: ef- Metab. 2004;89:2622-2631. demographic and lifestyle factors to symptoms in a fects of varying concentrations of estradiol. J Clin En- 13. Brown JR, Skurnick J, Sharma N, et al. Frequent multi-racial/ethnic population of women 40-55 years docrinol Metab. 1976;42:432-442. intermittent ovarian function in women with prema- of age. Am J Epidemiol. 2000;152:463-473. 7. Liu JH, Yen SSC. Induction of midcycle gonado- ture menopause: a longitudinal study. Endocrine. 1993; 20. North American Menopause Society Position tropin surge by ovarian steroids in women: a critical 1:467-474. Statement. Menopause. 2004;11:11-33. evaluation. J Clin Endocrinol Metab. 1983;57:797-802. 14. Santoro NS, Brown JR, Adel T, Skurnick JH. Char- 21. Chakraborty TR, Ng L, Gore AG. Age-related 8. Weiss G, Nachtigall LE, Ganguly M. Induction of acterization of reproductive hormonal dynamics in the changes in estrogen receptor beta in rat hypothala- an LH surge with estradiol benzoate. Obstet Gynecol. perimenopause. J Clin Endocrinol Metab. mus: a quantitative analysis. Endocrinology. 2003;144: 1976;47:415-418. 1996;81:1495-1501. 4164-4171.

Downloaded From: https://jamanetwork.com/ on 10/01/2021 LETTERS

ming sperm.”1 This description is supremely felicitous, as Rhesus Pieces the Veil Nebula is, in fact, the tattered remnants of a super- To the Editor: I read with interest the analysis of the sty- nova—an exploding star. listic aspects of the artwork of Wassily Kandinsky,1 but felt But there is a far deeper connection. The universe started that the geometric analysis regarding Untitled Improvisa- with little other than hydrogen and helium, and none of the tion III on the cover of the September 15, 2004, issue of JAMA heavy elements of our familiar world and ourselves. The sacrificed the whole at the expense of its parts, and in so heavier elements up through iron are produced inside mas- doing Kandinsky makes a monkey out of all of us. Just take sive stars much heavier than the sun, by thermonuclear fu- a second look and see if you don’t agree. sion during the stable portions of their lives. When such stars reach the end of their useable nuclear fuel, they become un- Harvey L. Edmonds, MD able to sustain their own weight, collapse, and rebound in [email protected] Quality Department a titanic explosion. This blast, occupying the last few sec- St Agnes Medical Center onds of the star’s existence, synthesizes the elements heavier Fresno, CA than iron and blows the entire star, aside from the core, into free space, where the heavy elements enrich the hydrogen 1. Southgate MT. The Cover. JAMA. 2004;292:1274. and helium of the pristine interstellar medium. The next gen- eration of stars and planetary systems born of the enriched gas thereby possesses the heavy elements required for the formation of solid planets and for life. Thus are we all, as CORRECTIONS astrophysicists and songwriters are wont to say, stardust. Data Error: In the Editorial titled “Stenting Small Coronary Arteries: Works in Progress” In this sense, the supernova actually is how “life begins.” published in the December 8, 2004, issue of the JOURNAL (2004;292:2777-2778), an incorrect number was published. On page 2777 at the bottom of the first col- The portion shown in the (much overexposed) photo- umn, the percentage of patients in the sirolimus stent group with diabetes melli- graph in the painting is called the Western Veil and is part tus should read 19% (not 9%).

of a larger complex called the Cygnus Loop, which is about Error in Table: In the Original Contribution entitled “Familial Risk of Lung Carci- 15000 years old, 2500 light years away, 4 times the appar- noma in the Icelandic Population” published in the December 22/29, 2004, issue ent diameter of the full moon, but very faint. Examined at of THE JOURNAL (2004;292:2977-2983), there was an error in Table 4. On page 2981, the second and fourth column headings, “⌬ RR of Lung Carcinoma – ⌬ RR leisure through a large telescope, under dark and transpar- of Smoking” should have read “RR of Lung Carcinoma – RR of Smoking.” ent skies, the Veil Nebula is a complex, subtle, and sublime Funding Omissions: In the Original Contribution titled “Menopause and Hypo- sight. Excellent photographs of the Veil Nebula can be found thalamic-Pituitary Sensitivity to Estrogen” published in the December 22/29, 2004, at the Web site of the National Optical Astronomy Obser- issue of the JOURNAL (2004;292:2991-2996), the funding statement was incom- vatories.2 plete. The paragraph should read: Howard L. Ritter, Jr, MD Funding/Support: The Study of Women’s Health Across the Nation (SWAN) was funded by the National Institute on Aging, the National Institute of Nursing Re- [email protected] search, and the NIH Office of Research on Women’s Health. The Toledo Clinic Toledo, OH In addition, the NIH Program Office paragraph was incomplete. It should read:

1. Southgate MT. The Cover. JAMA. 2004;292:1012. NIH Program Office: National Institute on Aging, Bethesda, Md: Sherry Sherman, 2. Veil Nebula image gallery. National Optical Astronomy Observatories Web site. 1994-present; Marcia Ory, 1994-2001; National Institute of Nursing Research, Available at: http://www.noao.edu/image_gallery/html/im0852.html. Accessi- Bethesda, Md: Yvonne Bryan, 2004-present; Janice Phillips, 2002-2004; Carole bility verified November 30, 2004. Hudgings, 1997-2002.

Downloaded From: https://jamanetwork.com/ on 10/01/2021