Hum. Reprod. Advance Access originally published online on May 26, 2006
Human Reproduction 2006 21(9):2450-2454; doi:10.1093/humrep/del191
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Duration not severity of the climacteric syndrome predicts resumption of hormone therapy after discontinuation: a prospective cohort study
Department of Obstetrics and Gynecology, Hadassah Medical Center, Hebrew University Hospital, Jerusalem, Israel
1 To whom correspondence should be addressed at: Department of Obstetrics and Gynecology, Hadassah Hebrew University Hospital, Mount Scopus, POB 24035, il-91240 Jerusalem, Israel. E-mail: kochman{at}hadassah.org.il
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
BACKGROUND: Predictive factors of women who are unable to quit prolonged hormonal therapy (HT) are largely unknown. We sought to identify predictors for the resumption of HT after the discontinuation of treatment. METHODS: A cohort prospective study was conducted allocating menopausal women treated with HT for over 3 years. Menopausal symptoms were monitored periodically after HT cessation by the Greene climacteric scale. RESULTS: Eighty-two women participated in the study. Age, the age of menopause, BMI, HT duration, the type of regimen, reasons cited to discontinue HT and the method of discontinuation did not differ between the subjects who successfully discontinued HT and those who failed to quit HT. Only the prevalence of vasomotor symptoms when HT was first prescribed significantly differed between the groups (P = 0.03). Comparable maximal Greene score was recorded in both groups. Over time, the subjects who returned to HT had higher Greene score [Hazard ratio 1.25, confidence interval (CI) 95% (1–1.07)] and significantly higher vasomotor score [Hazard score 1.22, CI 95% (1.02–1.46)]. CONCLUSIONS: The history of hot flashes and the duration of menopausal symptoms upon HT discontinuation predict the resumption of HT. Thus, the return to HT is expected in individuals who are intolerant of prolonged climacteric syndrome.
Key words: discontinuation/hormone therapy/hot flashes/menopause/predictors
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Several studies and especially Women’s Health Initiative (WHI) Study have shown that risks associated with hormonal therapy (HT) may outweigh the benefits (Hulley et al., 2002
; Writing Group for the Women’s Health Initiative Investigators, 2002; Shumaker et al., 2003). Increased risks for coronary artery disease, venous thromboembolic events, stroke, breast cancer and dementia were associated with HT use. These findings created an atmosphere that encouraged patients and their physicians to quit HT (Li et al., 2000; Ettinger et al., 2003) and led the North American Menopause Society (NAMS) panellists to recommend the use of the lowest effective dose of HT for the shortest time. Short-term HT was defined as 3–5 years, whereas long-term therapy was considered longer than 3–5 years (North American Menopause Society, 2002). Except for this vague recommendation for short-term use of HT, scarce literature identifying the characteristics of women who would experience difficulties in discontinuing HT could be found, leaving medical health providers unready to deal with these women as well as give evidence-based counselling. To identify the characteristics of individuals who resume treatment following a trial to discontinue HT, we prospectively followed a cohort of subjects aged 48–73 years, who had used HT regularly for at least 3 years and opted to stop therapy. We hypothesized that the severity of the climacteric symptoms upon cessation of HT would play an important role on the decision of women to resume HT.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Subjects
From May 2001 to April 2003, 91 healthy post-menopausal women opting to quit HT were enrolled into the study. An informed consent was obtained from all subjects. All women were treated with either combined estrogen/progestagen treatment or estrogen-alone therapy for a period longer than 3 years. None had received any medication or OTC supplementation that could affect the evaluation. The following parameters were recorded at the entrance to the study: age, the age of menopause, BMI, reasons to start therapy, the type of regimen, HT duration, reasons cited for HT discontinuation and the method of HT discontinuation. Women with the following conditions were excluded from the trial: vaginal bleeding of undiagnosed cause, abnormal PAP smear, finding of breast cancer, known or suspected estrogen-dependent tumours, fibroids, alcoholism, smoking, severe liver or kidney disorders, endometrial hyperplasia, severe uncontrolled hypertension, active ischaemic heart disease, evidence of acute thrombosis and infectious diseases.
Experimental design
This was a prospective clinical cohort study conducted in a menopause clinic setting at a university hospital and affiliated clinics. Study-eligible women were monitored regarding symptoms and compliance by the use of a standard climacteric scale (the Greene Scale; Greene, 1998). This scale distinguishes 21 different symptoms clustered into four subclasses: 11 psychological symptoms (subdivided in six anxiety symptoms and five depression symptoms), 7 somatic symptoms (e.g. headaches, muscle and joint pains), 2 vasomotor symptoms (hot flashes and night sweats) and a sexual symptom (loss of sexual interest). Each symptom score ranges from 0 (‘not at all’) to 3 (‘quite a bit’), compiling a Greene score range of 0–63. The questionnaire of the Greene climacteric scale was completed at 1, 3, 6, 9 and 12 months by the caring physician at the patient’s visits and by a telephone questionnaire. The maximal Greene score was used to represent the severity of symptoms, and the relation between the maximal and the last scores corresponds to the duration of symptoms.
Additionally, unscored complaints of vaginal dryness and weight gain of 1 kg per month were recorded at each time point. The telephone questionnaire was obtained by a single non-physician, highly trained interviewer.
This was a secondary outcome analysis of a previous prospective non-blinded study that investigated the recurrence of climacteric symptoms following two methods of discontinuation of HT (Haimov-Kochman et al., in press). In that study, women were randomized into two groups either stopping HT abruptly or gradually. The assumptions for the determination of the sample size are detailed elsewhere (Haimov-Kochman et al., in press).
Statistical methods
We used Chi-square or Wilcoxon test, as appropriate, to compare the characteristics of those who successfully stopped with those who resumed therapy. To develop a predictive model to identify women who were likely to fail an attempt to stop HT, we used variables that were associated with unsuccessful cessation in a univariate analysis. Variables found to be associated with unsuccessful stopping at P < 0.2 as well as historical risk factors were entered into a extended Cox hazard model, including Greene, vasomotor and mood scores as time-dependent variables. All statistical analyses were performed with SAS 8.2 (SAS Institute, Cary, NC, USA).
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Cohort characteristics
Eighty-two white women aged 48–73 years (mean age ± standard deviation, 56.5 ± 5.15 years) completed the full follow-up period of the study and were available for further analysis. The mean age of menopause was 47.5 ± 5.5 years with a mean HT duration of 8.25 ± 4 years. The mean BMI of the patients’ sample was 25.5 ± 4.3 (Table I). Around a quarter of the cohort (27.5%) was treated by estrogen-alone therapy following hysterectomy. Over half (53.9%) of the subjects in the study cited hot flashes as the reason to embark on HT, when first prescribed. The severity of hot flashes was subjectively defined by each interviewee. Other reasons to begin HT were sleep disturbances (28%), mood deterioration (19.5%), sexual dysfunction (9.8%), osteoporosis (6.1%), surgical (18.3%) and early menopause (6.1%) (Table I). Reasons to quit HT included the fear of breast cancer (38.3%), prolonged HT over 5 years (35.8%) and beyond 10 years (33.3%) (Table I). Fifty subjects (54.9%) stopped HT abruptly [according to randomization method described earlier (Haimov-Kochman et al., in press)].
|
Baseline characteristics of unsuccessful HT quitters
Of the 82 women who completed the study, 36 subjects (43.9%) returned to HT during 1 year (group 1), whereas 46 women (56.1%) stopped HT successfully (group 2). Most unsuccessful HT quitters returned to therapy within 3 months (28% within the first month and additional 33% within 2–3 months). None has returned to HT after 9 months (Figure 1). No differences were noted in individual characteristics between the two groups regarding age, the age of menopause, BMI and HT duration (Table I). The type of HT regimen, e.g. combined estrogen/progestagen or estrogen alone, did not differ between the groups. Likewise, the method of HT discontinuation either abruptly or over 6 months did not vary between the groups. Reasons cited to discontinue HT were similar between the groups (Table I). Of all the reasons mentioned to start HT, only the prevalence of vasomotor symptomatology significantly differed between the groups (P = 0.03) (Table I).
|
Characteristics of unsuccessful quitters upon HT discontinuation
Upon HT discontinuation, the climacteric symptomatology was monitored during 12 months follow-up period using both unscored and scored systems. The prevalence of the unscored complaints of vaginal dryness and weight gain did not differ between successful and unsuccessful HT
quitters (Table II).
|
align="char"After stopping HT, comparable mean maximal Greene scores (recorded during the follow-up period) were found in the patients who resumed HT and successfully quit (Figure 2). In subjects who returned to HT during the year of follow-up, the maximal Greene score was close to their last score before returning to HT (Figure 2). In this group, the mean last Greene score and specifically the mean last mood and vasomotor subscores (reported at the end of follow-up) were significantly higher than those in successful quitters (Figure 2). Successful quitters had significantly lower last Greene, mood and vasomotor scores than the recorded maximal scores (Figure 2).
|
The difference between the maximal and the last scores suggested change of the scores over time and led to performing Hazard model with time-dependent variables. The covariates entered into two separate models were the history of hot flashes, s/p oophorectomy, sleep disturbances, weight gain, vaginal dryness and the time-dependent variables: either Greene score or vasomotor and mood subscores. Over time, the subjects who returned to HT had higher Greene score [Hazard ratio 1.25, confidence interval (CI) 95% (1–1.07)] and significantly higher vasomotor score [Hazard score 1.22, CI 95% (1.02–1.46)] than their successful HT quitter counterparts.
Assessment of risk factors for unsuccessful quitters
We attempted to identify a set of variables that are easily ascertained in the clinic to identify women at risk to be unsuccessful HT quitters. We initially used all variables associated with unsuccessful discontinuation of HT in a univariate analysis that can be easily assessed before the attempt to stop HT. The history of hot flashes at baseline when HT was initially prescribed was found to be a major predictor for difficulty in stopping HT. Then, the development of climacteric symptomatology upon HT discontinuation was entered into the analysis as Greene score above the median. The probability to stay off HT was depicted over time (Figure 3). After a year, the probability to quit HT with these two specific risk factors was 0.32, with only one risk factor 0.60–0.64 and without risk factors 0.78 (P = 0.028) (Figure 3).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This article provides pioneering data on a spectrum of symptoms reported by a homogeneous sample of relatively healthy, young post-menopausal women after HT discontinuation. These findings offer preliminary insights into the anticipated symptom repertoire following HT cessation that may predict the resumption of HT.
Among the subset of women who tried to stop HT in this study, two of five resumed therapy. An equivalent high rate of return to HT (26%) has been previously reported in a large survey of the Kaiser Foundation Health Plan (Grady et al., 2003). Lately, in another large retrospective survey, about half of the women who tried to stop HT following the WHI publication eventually restarted therapy (Rolnick et al., 2005). Hence, return to HT may be expected in a significant portion of women who were on short-term HT following the NAMS recommendation. In view of the high failure rate to quit HT, high-risk patients for unsuccessful trial should be further characterized.
In this study, a major predictive factor for the resumption of HT after withdrawal was the history of hot flashes when HT was first prescribed. This finding is in agreement with the data of Grady et al. (2003) who reported that those who started HT for reasons other than health promotion were more likely to fail in their attempts to stop therapy. Another recent large cross-sectional survey indicated that more than half of the women with vasomotor symptoms at baseline reported these symptoms after discontinuing HT use (Ockene et al., 2005). And, vice versa, about 90% of HT users who reported vasomotor symptoms after the discontinuation of HT had also experienced them at baseline. In that study, 
80% of women with moderate to severe vasomotor symptoms at baseline returned to HT after HT cessation in contrast to 50% who had mild or no vasomotor symptoms at baseline (Ockene et al., 2005).
We have also shown that the following baseline parameters—age, the age of menopause, BMI, HT duration, the type of HT regimen and the method of HT discontinuation—were neither decreasing nor enhancing the chance of successful HT cessation. Therefore, advanced age, prolonged HT use and gradual tapering off HT, which could intuitively be considered as promoting successful cessation, were not shown to be of help to quit HT. Even the fear of breast cancer as a reason of stopping HT did not serve in this study as a predictor of successful cessation. These findings corroborate with others who reported that age, BMI, perceived risk of breast cancer and the method of stopping were not associated with the likelihood of resuming HT (Grady et al., 2003).
To our knowledge, this is the first investigation of the climacteric syndrome following HT cessation over time, focusing on its onset and duration. Our data showed that most unsuccessful HT quitters (61%) returned quickly within 3 months to HT and that none returned to HT beyond the 9th month. These findings suggest a time frame within which the risk to resume HT is highest and point out that after 9 months HT quitters are no more at risk to resume therapy.
In this study, we evaluated the menopausal symptoms repertoire by the Greene climacteric score, which provided us with a tool to quantify the many aspects of the climacteric syndrome by its subscores. The maximal score was used to represent the severity of symptoms and the relation between the maximal and the last scores represents the duration of symptoms. Comparable maximal Greene score values were found in both successful HT quitters and those who returned to HT. Moreover, the maximal values of mood, somatic, vasomotor and sex subscores were similar in both groups, indicating similar severity of the climacteric syndrome (Figure 2). During the year of follow-up, the maximal Greene score of the subjects who returned to HT was equal to their last score. Conversely, the last Greene score of successful quitters (reported at the end of 12 months follow-up) was significantly lower than the recorded maximal Greene score. Using the Cox hazard model, we found that over time, the subjects who returned to HT had higher Greene scores and significantly higher vasomotor scores. Altogether, our data testify that withdrawal symptoms in both successful and unsuccessful HT quitters were of the same severity, but the duration of climacteric symptoms in the group who resumed HT was longer. It has been shown in a large survey that the strongest predictor of resuming HT use was the development of troublesome withdrawal symptoms (Grady et al., 2003). Women who reported troublesome symptoms after quitting HT were nearly nine times more likely to resume HT. In this regard, vasomotor and non-vasomotor symptoms were equally likely to result in the resumption of HT use (Grady et al., 2003). A recent large cross-sectional survey reported that the most frequent reason to re-embark on HT after discontinuation was symptom management (Ockene et al., 2005). Likewise, the return of vasomotor symptoms upon HT discontinuation developed in 25% of quitters in another recent study, some of whom required treatment for symptom relief (Ness et al., 2006). However, climacteric symptoms also occurred in about half (47.5%) of the successful quitters in the retrospective survey of Grady et al. (2003) and were tolerated better by successful HT quitters. Our data suggest that troublesome withdrawal symptomatology in equivalent severity happens in both successful and unsuccessful HT quitters, but the duration of symptoms differs, leading to reduced tolerance of unsuccessful HT quitters. The discordance from the results of Grady et al. can be explained by Grady’s retrospective study design and by a recollection bias of the patients who resumed HT. It is still unclear why the climacteric syndrome is tolerated by some but not by others. We suggest here longer duration of the climacteric syndrome upon HT discontinuation as a possible explanation for reduced tolerance of the subjects returning to HT. Other explanations for better tolerance of climacteric symptoms that have not been investigated in our study include (i) the presence of support systems—spouse, family or friends, (ii) patient’s knowledge in regard to the climacterium as well as HT benefits and risks and (iii) biological factors such as higher serum/tissue estrogen levels and a wider hypothalamic thermoregulatory range (Freedman, 2001).
In this study, we identified two major predictors for HT resumption after an attempt to stop therapy—history of hot flashes when HT was initially prescribed and the reappearing climacteric syndrome after HT discontinuation independently increased the probability to return to HT by 2-fold. Having both risk factors augmented the probability to return to HT by 2.4-fold. Our findings are timely in the light of the current recommendations to limit HT primarily to the treatment of severe climacteric symptoms for the shortest duration possible. We argue that vasomotor symptomatology when HT was first prescribed predicts difficulty in stopping HT and suggest that intolerance to the ensuing climacteric syndrome depends more on its duration rather than on its severity.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Ettinger B, Grady D, Tosteson AN, Pressman A, Macer JL. (2003) Effect of the Women’s Health Initiative on women’s decisions to discontinue postmenopausal hormone therapy. Obstet Gynecol 102:1225–1232.
Freedman RR. (2001) Physiology of hot flashes. Am J Hum Biol 13:453–464.[CrossRef][ISI][Medline]
Grady D, Ettinger B, Tosteson AN, Pressman A, Macer JL. (2003) Predictors of difficulty when discontinuing postmenopausal hormone therapy. Obstet Gynecol 102:1233–1239.
Greene JG. (1998) Constructing a standard climacteric scale. Maturitas 29:25–31.[CrossRef][ISI][Medline]
Haimov-Kochman R, Barak-Glantz E, Arbel R, Leefsma M, Brzezinski A, Milwidsky A, Hochner-Celnikier D. Gradual discontinuation of hormone therapy does not prevent the reappearance of climacteric symptoms. A randomized prospective study. Menopause. (in press).
Hulley S, Furberg C, Barrett-Connor E, Cauley J, Grady D, Haskell W, Knopp R, Lowery M, Satterfield S, Schrott H, et al. (2002) Noncardiovascular disease outcomes during 6.8 years of hormone therapy. Heart and estrogen/progestin replacement study follow-up (HERS II). JAMA 288:58–66.
Li C, Samsioe G, Lidfelt J, Nerbrand C, Agardh CD. Women’s Health in Lund Area (WHILA) Study. (2000) Important factors for use of hormone replacement therapy: a population-based study of Swedish women. The Women’s Health Lund Area (WHILA) Study. Menopause 7:273–281.[ISI][Medline]
Ness J, Aronow WS, Beck G. (2006) Menopausal symptoms after cessation of hormone replacement therapy. Maturitas 53:356–361.[CrossRef][ISI][Medline]
North American Menopause Society. (2002) NAMS Advisory Panel on Report from the Postmenopausal Hormone Therapy. Menopause, October(http://www.menopause.org/news.html#advisory).
Ockene JK, Barad DH, Cochrane BB, Larson JC, Gass M, Wassertheil-Smoller S, Manson JE, Barnabei VM, Lane DS, Brzyski RG, et al. (2005) Symptom experience after discontinuing use of estrogen plus progestin. JAMA 294:183–193.
Rolnick SJ, Kopher RA, DeFor TA, Kelley ME. (2005) Hormone use and patient concerns after the findings of the Women’s Health Initiative. Menopause 12:399–404.[CrossRef][ISI][Medline]
Shumaker SA, Legault C, Rapp SR, Thal L, Wallace RB, Ockene JK, Hendrix SL, Jones BN III, Assaf AR, Jackson RD, et al. (2003) Estrogen plus progestin and the incidence of dementia and mild cognitive impairment in postmenopausal women: The Women’s Health Initiative Memory Study: a randomized controlled trial. JAMA 289:2651–2662.
Writing Group for the Women’s Health Initiative Investigators. (2002) Risks and benefits of estrogen plus progestin in healthy postmenopausal women. Principal results from the Women’s Health Initiative randomized controlled trial. JAMA 288:321–333.
Submitted on March 4, 2006; resubmitted on April 21, 2006; accepted on May 3, 2006.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||