This Article Abstract Figures Only Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Baca-García, E. Articles by Saiz-Ruiz, J. Search for Related Content PubMed PubMed Citation Articles by Baca-García, E. Articles by Saiz-Ruiz, J. Related Collections Depression Sexual Medicine: Female Suicide or Suicidal Behavior

The Relationship Between Menstrual Cycle Phases and Suicide Attempts

From the Department of Psychiatry (E.B.-G., C.D.-S., J.S.-R.), Hospital Ramon y Cajal, Universidad de Alcalá, Madrid, Spain; and Eastern State Hospital (J.D.), Lexington, Kentucky.

Address reprint requests to: Jose de Leon, MD, Mental Health Research Center, Eastern State Hospital, 627 W. 4th St., Lexington, KY 40508. Email: jdeleon{at}pop.uky.edu.us

	ABSTRACT

TOP ABSTRACT INTRODUCTION COMPLETED SUICIDES SUICIDE ATTEMPTS SUICIDE THREATS CONCLUSIONS FROM THE LITERATURE... MATERIALS AND METHODS RESULTS DISCUSSION NOTES ACKNOWLEDGMENTS REFERENCES

 
OBJECTIVE: The validity of prior studies on the menstrual cycle and suicide attempts assumes that suicidal women accurately describe their cycles. The three objectives of this study were 1) to explore whether prior inconsistencies are due to the effects of sample selection and method of assessment of the menstrual cycle, 2) to assess the relationship between the menstrual cycle phase and suicide attempts, and 3) to establish the role of sexual hormones in suicide attempts.

METHODS: The original sample included 134 women who came to the emergency room of a general hospital after a suicide attempt. One hundred eight female blood donors were recruited as control subjects. The menstrual cycle was divided into follicular, midcycle, and luteal phases using two clinical methods and serum hormonal assessment. Dividing the follicular phase into menstrual and nonmenstrual phases was also considered.

RESULTS: Two of 11 previously used sampling methods produced a sample size similar to that of the hormonal assessment. values between the two clinical and the endocrinological methods were low (0.40–0.50). The number of suicide attempts during the follicular phase (particularly during the menstrual phase) was significantly higher than expected.

CONCLUSIONS: Despite the inability to control for other variables and limitations, the results of this study suggest that sample selection could introduce biases and that studies relating psychiatric symptomatology and menstrual cycle phases need to use hormonal determinations. New studies are needed to verify that suicide attempts are more frequent during the follicular phase (particularly during the menstrual phase).

Key Words: menstrual cycle • suicidal behavior • sexual hormones • hormonal assessment • agreement • gonadotropins

Abbreviations: FSH = follicle-stimulating hormone; LH = luteinizinghormone.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION COMPLETED SUICIDES SUICIDE ATTEMPTS SUICIDE THREATS CONCLUSIONS FROM THE LITERATURE... MATERIALS AND METHODS RESULTS DISCUSSION NOTES ACKNOWLEDGMENTS REFERENCES

 
In the past few years, there has been an interest in the relationship of the menstrual cycle with psychiatric symptomatology in general and with mood disorders in particular (1–4). Suicidal behavior may also be influenced by the menstrual cycle, but the results of the various studies are not consistent because some important underlying methodological issues have not been resolved (5, 6).

Following the organization of the excellent review of Weztel and McClure (5), the suicide studies can be classified into studies of completed suicide, attempted suicide (the scope of this article), and suicidal threats. Regardless of whether these behaviors are different phenomena (7), the methodological problems of studying them are very different.

	COMPLETED SUICIDES

TOP ABSTRACT INTRODUCTION COMPLETED SUICIDES SUICIDE ATTEMPTS SUICIDE THREATS CONCLUSIONS FROM THE LITERATURE... MATERIALS AND METHODS RESULTS DISCUSSION NOTES ACKNOWLEDGMENTS REFERENCES

 
Although the studies of completed suicides can accurately estimate the phase of the menstrual cycle by endometrial histology (8, 9), they lack reliable psychiatric assessments. Another problem with these studies is selection bias, because not all suicides are reported and diagnosed (10). This could explain why the two best methodological studies of completed suicides and menstrual cycle did not obtain the same results. McKinon et al. (8) point out a relationship between completed suicides and luteal phase; however, Vanezis (9) did not find any relationship between menstrual cycle and completed suicides.

	SUICIDE ATTEMPTS

TOP ABSTRACT INTRODUCTION COMPLETED SUICIDES SUICIDE ATTEMPTS SUICIDE THREATS CONCLUSIONS FROM THE LITERATURE... MATERIALS AND METHODS RESULTS DISCUSSION NOTES ACKNOWLEDGMENTS REFERENCES

 
In the literature, there is great controversy about the correct nomenclature of suicide (11). For this reason, it is important to explain the definition used in research. In fact, only 6 of the 20 studies reviewed in the last 40 years give a clear definition of the behavior analyzed (Table 1). The main methodological problems in the study of suicide attempts and menstrual cycle are sample selection and the assessment of menstrual cycle phases.

Assessment of menstrual cycle phases.
Timing is a crucial factor when the objective of the research is a state marker, such as the menstrual phase (13). Unfortunately, only 8 of 20 articles reported the interval between the suicidal behavior and assessment of the menstrual cycle. Twenty years ago, Pallis and Holding (14) pointed out that in studies of the relationship between the menstrual cycle and suicidal behavior, it is fundamental to accurately estimate the phases. Most of these studies have used patient report instead of hormonal assessment to estimate menstrual cycle phases (5). Even recent studies have continued to use patient report (6, 15–19). The validity of these studies rests on the assumption that women interviewed after a suicide attempt can provide accurate enough information to permit a clinical estimation of the phase of their menstrual cycle. However, many researchers interested in the menstrual cycle believe that untrained women may not provide reliable information about their menstrual cycles (20).

The agreement between the phases estimated by patient report and by hormonal assessment has never been studied in women who have attempted suicide, and this methodological problem could explain the inconsistency of the results in this area of research (6). Recently, results of a small study (not of suicidal behavior but of the relationship between psychosis and the menstrual cycle) suggested that 32 psychotic women were not significantly different in their ability to estimate their menstrual cycle phases from 26 nonpsychotic women (21).

The second problem, also related to the determination of the menstrual cycle by patient report, is the representativeness of the samples used for analysis. Most of the studies (Table 1) exclude a significant part of the sample. The ratio of patients excluded fluctuates between 60% (22) and 6% (23) (Table 1). Besides exclusions, an additional explanation for differences in studies of the menstrual cycle in suicidal patients is the bias produced by stress (15, 17).

Other common problems associated with the clinical methods are the artificial adjustment of cycle length to 28 days and not considering the fixed duration of the luteal phase (between 12 and 14 days). This could produce errors in classification because women with cycles longer than 28 days usually have a longer follicular phase, and these methods could classify this phase as the midcycle or even luteal phase. In addition, the classification used in some studies does not follow the traditional physiological phases (9). The progressive, widespread use of hormonal contraceptives reduces the possibility of studying the effect of the "natural" menstrual cycles in suicidal behavior.

	SUICIDE THREATS

TOP ABSTRACT INTRODUCTION COMPLETED SUICIDES SUICIDE ATTEMPTS SUICIDE THREATS CONCLUSIONS FROM THE LITERATURE... MATERIALS AND METHODS RESULTS DISCUSSION NOTES ACKNOWLEDGMENTS REFERENCES

 
There are many methodological problems associated with collecting research data on suicidal threats. The definition of this type of patient is the most problematic, and the bias in sample selection is the greatest. Two published studies described phone callers (24–26), but the validity and reliability of the patient information was not verified with information from any other source.

	CONCLUSIONS FROM THE LITERATURE REVIEW

TOP ABSTRACT INTRODUCTION COMPLETED SUICIDES SUICIDE ATTEMPTS SUICIDE THREATS CONCLUSIONS FROM THE LITERATURE... MATERIALS AND METHODS RESULTS DISCUSSION NOTES ACKNOWLEDGMENTS REFERENCES

 
In summary, despite these methodological differences, review of the literature from the past 40 years shows that 1) no relationship was identified between week of the menstrual cycle and suicidal behavior in six studies 15, 17, 18, 22, 27, 28), 2) suicide attempts were more frequent during the premenstrual or fourth week in five studies (29–33), 3) suicide attempts or completed suicides were more frequent during the menstrual or first week in five studies (6, 16, 23, 34, 35), and 4) suicide attempts were more frequent before and after menses (fourth and first week) or during the perimenstrual weeks in two studies (19, 36).

Authors of six studies concluded that a relationship between menstrual cycle and suicidal behavior does not exist (15, 17, 18, 22, 27, 28), and authors of three studies (31, 32, 36) believed that the menstrual cycle plays an important role in this behavior. However, most authors think that the menstrual cycle could be considered as another contributing factor in suicide (23).

Ekeberg et al. (17) and Luggin et al. (15) concluded that the menstrual cycle is affected by the stress associated with suicidal behavior; thus, the relationship between menstrual cycle and suicidal behavior may be explained by the effect of the suicidal behavior on the menstrual cycle. Following Frank’s hypothesis (37), some authors have speculated that the premenstrual phase is a trigger of acute psychiatric problems (36). Glass et al. (31) and Tonks et al. (32) proposed that the premenstrual phase is a period of increased vulnerability to suicidal attempts in susceptible women (eg, nonpsychotic women with a medical or gynecological history and poor sexual and marital adjustment), whereas in others the premenstrual phase has a protective effect.

Finally, some causal explanations have been proposed as the underlying phenomena responsible for the relationship between the menstrual cycle and suicide attempts. The most popular explanation is the hormonal effect. In fact, a recently published study supports the idea that the hypoestrogenic phases of the cycle are associated with suicide attempts (16). Lester (38) hypothesized that the effect of estrogen over dopamine can influence mood in women and therefore influence suicidal behavior. Other possible explanations are emotional instability linked with premenstrual syndrome (36), the fear of unexpected pregnancy after a delay of menstruation (28), and psychosocial gender factors (39, 40).

This study had three objectives. The first objective was to explore how the effects of two methodological issues (sample selection and assessment of menstrual cycle phases) may contribute to the inconsistent results found in the literature. The second objective was to assess the relationship between the different physiological phases of the menstrual cycle and suicide attempts. The third objective was to establish the role of sexual hormones in suicide attempts. The sample sizes for the different analyses varied according to the exclusion criteria.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION COMPLETED SUICIDES SUICIDE ATTEMPTS SUICIDE THREATS CONCLUSIONS FROM THE LITERATURE... MATERIALS AND METHODS RESULTS DISCUSSION NOTES ACKNOWLEDGMENTS REFERENCES

 
The Ramon y Cajal Hospital is a general hospital in the Spanish National Health System that provides medical coverage for 500,000 people in a catchment area of Madrid. All persons from this area who attempt suicide are seen in this hospital before they undergo any psychiatric treatment. Two hundred nineteen patients (134 women and 85 men) who came to the emergency room of this hospital from January 1996 to December of 1997 because of a suicide attempt, defined by the criteria proposed by O’Carroll et al. (11), were screened after informed consent was obtained and within 24 hours of the attempt. Eight (3.7%) of these patients were admitted to the intensive care unit, 8 (3.7%) were admitted to a medical ward, 91 (41.6%) were admitted to the psychiatry unit, and 111 (50.7%) were discharged after psychiatric evaluation.

One hundred eight female blood donors at the same hospital who did not have a history of suicidal behavior or psychiatric disorders were also recruited to serve as control subjects.

Sample Selection
The selection criteria used in 11 previous studies (Figure 1) were used to assess the menstrual cycle. Of the initial 134 female patients selected, 44 were excluded. The reasons for exclusion were menopause (29 patients), menopause and hormonal therapy (1 patient), hormonal levels outside reference values (3 patients), use of hormonal contraceptives (7 patients), and technical reasons (4 patients). Ninety premenopausal patients were included and further characterized on the basis of their hormonal assessment, which estimated their menstrual situation. Within this group of 90 patients, different criteria were used to determine the various percentages. Confidence intervals of these proportions (see Figure 1) were calculated using the computer program Power and Precision (41).

View larger version (10K): [in this window] [in a new window]   Fig. 1. Percentage of patients selected using different authors’ inclusion criteria to assess the menstrual cycle.

Of the 108 control subjects, 78 remained after exclusion criteria were considered. Thirty were excluded because of menopause (21 patients), irregular cycles (1 patient), and use of oral contraceptives (8 patients).

Patients and control subjects were asked about characteristics of their menstrual cycles in the last 6 months, including regularity, use of hormonal contraceptives, maternity status, cycle length (average and range), and menstrual phase length (average and range). They were also asked to identify their last date of bleeding onset (Figure 2). This information was used to estimate the menstrual cycle phase using two clinical methods.

View larger version (16K): [in this window] [in a new window]   Fig. 2. Frequency of reported suicide attempts and number of days from the onset of last menstrual period (N = 68).

The McIntosh et al. (42) method was developed in a prospective study to obtain equations to subsequently predict the most likely day of the LH surge, and hence the time of ovulation, solely from knowledge of the average length and variability of a women’s cycle. This method requires application of two formulas. The formula used to calculate the first day of midcycle is 0.770 x average days of cycle length - 7.685 - 1.96 x (0.432 x range of cycle length + 0.694). The formula used to calculate the last day of midcycle is 0.770 x average days of cycle length - 7.685 + 1.96 x (0.432 x range of cycle length + 0.694). These formulas were used to determine the midcycle; the other two phases could then be calculated by exclusion. A recent review (43) described the McIntosh et al. method as a calculation method based on average cycle length and provided formulas that allow one to predict the day of ovulation with 95% confidence. This method could be used in 68 of the 134 patients. Sixty-six patients were excluded because of a lack of information needed to apply the method and/or menopause. This method could also be used in 65 of the 108 control subjects. Forty-three control subjects were excluded because they did not provide the information required for this method and/or menopause.

A blood sample was collected from each subject in both groups to assess levels of sexual hormones (estradiol, progesterone, FSH, and LH) and was analyzed at the hospital laboratory by using Immulite assays (Diagnostic Products Corporation (DPC), Los Angeles, CA), which are chemiluminescent enzyme immunoassays (44). The three physiological phases were determined using the hormonal levels recommended by the manufacturer (45–48). The levels used for the follicular phase are as follows: FSH, 3.4 to 10 mIU/ml; LH, 1.6 to 8.3 mIU/ml; estradiol, 0 to 266 pg/ml; and progesterone, 0 to 1.5 ng/ml. The levels used for midcycle are as follows: FSH, 5.7 to 20 mIU/ml; LH, 15 to 62 mIU/ml; and estradiol, 118 to 355 pg/ml. The luteal phase was determined using progesterone levels >2.3 ng/ml. The final sample for hormonal analysis included 90 patients and 75 control subjects. Forty-four patients were excluded because of menopause (29 patients), menopause and hormonal therapy (1 patient), hormonal levels outside reference values (3 patients), use of oral contraceptives (7 patients), and technical reasons (4 patients). Thirty-three control subjects were excluded because of menopause (21 patients), irregular cycles (1 patient), use of hormonal contraceptives (8 patients), and hormonal levels outside reference values (3 patients).

values and 95% confidence intervals were used to measure the agreement between the two clinical methods using subject report and hormonal assessment. Weighted values were calculated but are not described because they were very similar.

Relationship Between Menstrual Cycle Phases and Suicide Attempts
Of the 134 patients, 90 could be assessed by the hormonal method (see exclusion above), and data from these patients were used to compare the frequency of suicide attempts during the three periods of the menstrual cycle established with the hormonal method against a theoretical distribution by means of ² test. This theoretical distribution was calculated by dividing the sample into groups based on the duration of the three physiological phases (follicular, days 1–11; midcycle, days 12–16; and luteal, days 17–28).

In addition, the menstrual cycle can be further divided into four phases by dividing the follicular phase into menstrual and nonmenstrual follicular phases. Therefore, the menstrual cycle was divided into menstrual, follicular, midcycle, and luteal phases for the 68 patients who could be classified according to the McIntosh et al. method (ie, they gave enough information to apply the formula) and the hormonal method. These 68 patients (see exclusions in Assessment of Menstrual Cycle Phases) were used to test the association between suicide attempts and menstruation. Seventeen of the 68 patients in this group were menstruating when they attempted suicide (Figure 2). The mean, mode, and median of the duration of menstruation for these patients were 4.24, 4, and 4 days, respectively, with a standard deviation of 1.45 days. Thus, the menstrual cycle in these 68 patients could be divided into a menstrual phase (days 1–4), follicular phase after menstruation (days 5–11), midcycle (days 12–16), and luteal phase (days 17–28). A ² test was performed using the same procedure described above but using these new intervals to calculate the theoretical distribution.

Role of Sexual Hormones in Suicide Attempts
The hormonal levels of 90 of the 134 patients and 75 of the 108 control subjects were compared for each phase of the cycle by means of analysis of variance for two factors (patient or control subject and menstrual cycle phases). Forty-four patients were excluded because of menopause (29 patients), menopause and hormonal therapy (1 patient), hormonal levels outside reference values (3 patients), use of hormonal contraceptives (7 patients), and technical reasons (4 patients). Thirty-three control subjects were excluded because of menopause (21 patients), irregular cycles (1 patient), use of hormonal contraceptives (8 patients), and hormonal levels outside reference values (3 patients).

	RESULTS

TOP ABSTRACT INTRODUCTION COMPLETED SUICIDES SUICIDE ATTEMPTS SUICIDE THREATS CONCLUSIONS FROM THE LITERATURE... MATERIALS AND METHODS RESULTS DISCUSSION NOTES ACKNOWLEDGMENTS REFERENCES

 
Sample Selection
Only the methods proposed by Forestié et al. (16) (88 patients), Glass et al. (31) (90 patients), and Tonks et al. (32) (79 patients) produced a sample size similar to that produced by hormonal assessment (90 patients, 69% of total patient sample). Figure 1 demonstrates that most of the confidence intervals of the sample sizes using clinical selection criteria do not include 69% (percentage of patients included with hormonal assessment). This lack of overlap means that there is a significant difference between most of the sample sizes obtained by using the different clinical selection criteria and the sample size using the hormonal method.

It is interesting that use of the Forestié et al., Glass et al., and Tonks et al. sampling methods led to inclusion of one, two, and one menopausal patients, respectively. These patients had FSH levels >20 mIU/ml.

Assessment of Menstrual Cycle Phases
Table 2 demonstrates the agreement between the clinical methods and the hormonal assessment. The frequency of suicidal patients correctly classified was 65% (42 of 65) for the Holding and Minkoff method (22) and 59% (40 of 68) for the McIntosh et al. method (42). The frequency was 69% for both methods (50 of 72 and 45 of 65) in control subjects. The respective values (and their 95% confidence intervals) between the two clinical methods and the endocrinological method were 0.40 (0.21–0.58) and 0.32 (0.14–0.50) for patients and 0.48 (0.31–0.65) and 0.48 (0.31–0.65) for control subjects. The value between both clinical methods of assessment was 0.92 (0.84–1.00) in patients and 1.00 in control subjects. Therefore, use of one or the other clinical method did not appear to influence agreement with the hormonal assessment.

View larger version (17K): [in this window] [in a new window]   Fig. 3. Distribution (%) of suicide attempts during the menstrual cycle (three phases) (N = 90).

View larger version (17K): [in this window] [in a new window]   Fig. 4. Distribution (%) of suicide attempts during the menstrual cycle (four phases) (N = 68).

	DISCUSSION

TOP ABSTRACT INTRODUCTION COMPLETED SUICIDES SUICIDE ATTEMPTS SUICIDE THREATS CONCLUSIONS FROM THE LITERATURE... MATERIALS AND METHODS RESULTS DISCUSSION NOTES ACKNOWLEDGMENTS REFERENCES

 
Sample Selection
As Holding and Minkoff (22) reported, admission policy and methodology determine the sample size and sample characteristics in all studies in this field. Our data support the idea that the apparent contradiction between studies could reflect bias in sample selection due to exclusion criteria related to menstrual cycle determination. Moreover, in five studies, the sample was selected by method of suicide attempt (eg, self-poisoning). This is an important source of bias because the method of suicide attempt is determined by sociocultural context and availability (49).

Assessment of Menstrual Cycle Phases
Considering the hormonal method as the gold standard, the two clinical methods using patient report did not seem to be reliable enough to estimate menstrual cycle phases. Although the values between the two clinical methods and the hormonal method for both patients and control subjects were highly significant (p < .001), they were both in the moderate range, which reflects that only approximately two-thirds of the samples were in agreement (50).

Thompson et al. (21) provided original data, making it possible to calculate the percentage of agreement in psychotic women (78%, 25 of 32) and female control subjects (81%, 21 of 26) as well as the respective values and 95% confidence intervals (0.61 (0.39–0.82) and 0.64 (0.38–0.89), respectively). Using analysis of variance, Thompson et al. did not find a significant difference between agreement in the menstrual cycle phases, so they concluded that the report of psychotic women was "extremely accurate" or as accurate as the report of nonpsychotic women. Similarly, in our study, the value for control subjects was slightly better than that for patients. In addition, the difference in agreement between control subjects and patients was not statistically significant because the 95% confidence intervals of values for patients and control subjects overlapped. Therefore, in our sample, patients were not significantly worse than control subjects in providing accurate information about their menstrual cycles. Although our results are similar to Thompson et al.’s results, our interpretation is different. There were no differences between patients and control subjects, but inaccurate classifications of menstrual cycle phase were produced in both groups by both clinical methods when menstrual cycle data provided by the women were used to calculate cycle phases. Our results support the idea that the reliability of memory of menstrual events is low in untrained women (20). A possible reason for the slightly lower reliability of patients is that 6 patients were mildly confused after self-poisoning and 25 patients (26 using the McIntosh’s method) seemed to be reluctant to disclose personal information.

In conclusion, our finding of an error rate of approximately one third in the clinical estimations suggests that subject report, in both patients and untrained control subjects, may introduce significant error measurement in studies relating psychiatric symptomatology and the menstrual cycle. Subject report may be an appropriate method for gross initial estimation in pilot studies, but more rigorous and definitive studies need to use hormonal assessments to establish menstrual cycle phases. In addition, these types of methods could produce additional sampling biases. Reviewers of studies of the menstrual cycle in psychiatry should pay attention to the serious limitations of most studies because patient report was used to establish the phases. Unless patients are highly educated or are trained to provide a better estimation of their menstrual cycle, the studies may be biased. Collins (51) noted that retrospective analysis of the data suggests that the failure rates of methods could be low when used by motivated women of reproductive age.

Relationship Between Menstrual Cycle Phases and Suicide Attempts
Our results showing that suicide attempts are more frequent during the follicular phase and to a greater extent during the menstrual period are consistent with results of the latest studies (6, 16, 35). This result must be carefully interpreted because temporal relationship is not an indicator of causality (52, 53). Despite this, the menstrual cycle could be considered a state marker of suicidal behavior.

Role of Sexual Hormone in Suicide Attempts
Another interesting result is that hormonal levels in patients with suicidal behavior were not different from those of control subjects. One might speculate that the hormonal profile during the follicular phase (characterized by low gonadotrophin levels) could induce suicidal behavior in predisposed women (31, 32, 54). Along this line of thinking, impulsivity has been proposed as a predisposing factor to suicidal behavior (55). Serotoninergic function is also linked to suicidal behavior (56). Finally, impulsivity and serotoninergic function change across the menstrual cycle (57–59). In addition, serotoninergic function is lower when the levels of gonadal hormones are low, so it could be hypothesized that impulsive women are more likely to attempt suicide during the cycle phase with the lowest levels of gonadal hormones (perimenstrual phase).

Other Variables
Because of the small size of the midcycle group of patients who attempted suicide, there was not enough power to control for the relationship between sociodemographic data or clinical variables, such as mood disorders, and menstrual cycle. However, an overview of frequency distribution (Table 3) indicates a more serious profile of suicidal behavior during the follicular phase than during the midcycle and luteal phases.

With respect to premenstrual syndrome and suicidality, we did not find a significant association between menstrual cycle phase and antecedents of premenstrual syndrome in the past year (² = 0.81, df = 3, p = .84). The rate of premenstrual syndrome (22%) was higher than the estimated prevalence in the US population, 8% to 13% (60), but not as high as the 30.5% described in a study of the Spanish general population (61).

Limitations
One limitation of our study is the lack of a measure of the influence of stress on the menstrual cycle to determine whether it could be responsible for the link between menstrual cycle and suicidal behavior (15, 17, 27). There is evidence that stress can increase the length of the menstrual cycle (62). However, we think that the impact of stress on the menstrual cycle in our sample was small for two reasons. First, most of the patients had hormonal profiles in the normal range, and second, only 7.5% (5 of 68 patients) had longer than expected cycles. It cannot be completely ruled out that early termination of the luteal phase may have confounded the results of the study. It is not likely that this had a major confounding effect because 1) it is an infrequent phenomenon and 2) the stress associated with the suicide attempt was too short in most cases to have an immediate effect on sexual hormones levels. Early termination of the luteal phase seems to be an infrequent phenomena, at least in normal women (5–6% of cycles) (63), but has not been studied in women under stress. All patients were assessed within the first 24 hours after the attempt, and 78% of the attempts were impulsive (<3 hours considering suicide) (Table 3). Therefore, in 78% of subjects, the duration of stress was <27 hours. This duration is probably not long enough to induce menses or abruptly change the progesterone levels so that early termination of the luteal phase occurs. For this analysis, one of the inclusion criteria was regular menstrual cycles before the suicide attempt. Because all patients with a history of irregularities in their cycles or abnormal values were excluded, it is likely that women prone to be sensitive to stress were also excluded. However, it cannot completely be ruled out that in a small number of patients, stress may contribute to early termination of the luteal phase.

It is possible that psychotropic medications may influence the menstrual cycle in women taking them. However, this is not likely because 1) women with prior menstrual irregularities were excluded from the analysis and 2) antipsychotics, the psychotropics consistently related to changes in the menstrual cycle, were taken by 11% of the sample. The most frequently taken psychotropics, benzodiazepines (56% of patients) and antidepressants (50% of patients), are not believed to have important and consistent influences on the menstrual cycle.

Another possible limitation of this study is the validity of the use of a single hormone sample to determine menstrual cycle phase because gonadotrophins have a pulsate diurnal pattern of secretion. In this sense, during the midcycle phase, the hormonal FSH and LH levels are higher than the highest levels during the other phases (64). The rise of progesterone is a signal that ovulation has occurred and, in this sense, confirms the luteal phase (65). Moreover, some authors have defended the idea that a single serum progesterone measurement is enough to confirm ovulation (66). Grinsted et al. (67) report that the serum estradiol peak has a positive predictive value of 83% and a negative predictive value of 100% when used to predict ovulation. The use of a single serum hormonal determination has an advantage: It is well tolerated by women (68). Repeated hormonal assessment would have been better to establish menstrual cycle phases. However, we were unable to convince patients to call us after the suicide attempt to get information on the onset of menses, so it would have been highly improbable that we could have convinced them to come back for repeated hormonal assessments.

New studies with hormonal determination are needed to test the hypothetical relationship between impulsiveness, menstrual cycle, and serotoninergic function.

	ACKNOWLEDGMENTS

TOP ABSTRACT INTRODUCTION COMPLETED SUICIDES SUICIDE ATTEMPTS SUICIDE THREATS CONCLUSIONS FROM THE LITERATURE... MATERIALS AND METHODS RESULTS DISCUSSION NOTES ACKNOWLEDGMENTS REFERENCES

 
We acknowledge Dr. Villar Palasi and the staff of the emergency service for their cooperation in the completion of this study, the blood donors who served as control subjects, and Lori Dinsmore, BS, and Debra Browne, BS, who helped edit the manuscript. This study was partially supported by the Hospital Ramon y Cajal, which provided hormonal assessments at no cost, and by a National Alliance for Research on Schizophrenia and Depression (NARSAD) Young Investigator Award to Dr. Baca-García.

	NOTES

TOP ABSTRACT INTRODUCTION COMPLETED SUICIDES SUICIDE ATTEMPTS SUICIDE THREATS CONCLUSIONS FROM THE LITERATURE... MATERIALS AND METHODS RESULTS DISCUSSION NOTES ACKNOWLEDGMENTS REFERENCES

 
Preliminary results of this study, using a reduced sample size, were presented in abstract form at the 28th Annual Meeting of the International Society of Psychoneuroendocrinology, July 26–30, 1997, San Francisco, California.

Received for publication March 9, 1998.

Revision received July 21, 1999.

	REFERENCES

TOP ABSTRACT INTRODUCTION COMPLETED SUICIDES SUICIDE ATTEMPTS SUICIDE THREATS CONCLUSIONS FROM THE LITERATURE... MATERIALS AND METHODS RESULTS DISCUSSION NOTES ACKNOWLEDGMENTS REFERENCES

 

1. Halbreich U, Endicott J, Nee J. Premenstrual depressive changes. Arch Gen Psychiatry 1983; 40: 535–42.[Abstract/Free Full Text]
2. Abramowitz ES, Baker AH, Fleischer SF. Onset of depressive psychiatric crisis around the menstrual cycle. Am J Psychiatry 1992; 139: 465–78.
3. Leibenluft E, Fiero PL, Rubinow DR. Effects of the menstrual cycle on dependent variables in mood disorder research. Arch Gen Psychiatry 1994; 51: 761–81.[Abstract/Free Full Text]
4. Seeman M. Psychopathology in women and men: focus on female hormones. Am J Psychiatry 1997; 154: 1641–7.[Abstract/Free Full Text]
5. Wetzel RD, McClure JN Jr. Suicide and the menstrual cycle: a review. Compr Psychiatry 1972; 13: 369–74.[Medline]
6. Baca-García E, Sánchez González A, González Diaz-Corralero P, González Garcia I, de Leon J. Menstrual cycle and profiles of suicidal behavior. Acta Psychiatr Scand 1998; 97: 33–5.
7. Diekstra RFW. The epidemiology of suicide and parasuicide. Acta Psychiatr Scand Suppl 1993; 371: 9–20.[Medline]
8. McKinon LC, McKinon PC, Thompsom AD. Lethal hazards of luteal phase of the menstrual cycle. BMJ 1959; 1: 1015–7.
9. Vanezis P. Deaths in women of reproductive age and relationship with menstrual cycle phase: an autopsy study of cases reported to the coroner. J Forensic Sci 1990; 47: 39–57.
10. Leads from the MMWR: operational criteria for determining suicide. JAMA 1989;261:360, 366.
11. O’Carroll PW, Berman AL, Maris RW, Moscicki EK, Tanney BL, Silverman M. Beyond the Tower of Babel: a nomenclature for suicidology. Suicide Life Threat Behav 1996; 26: 227–35.
12. Diekstra RFW. Parasuicide: is it a distinct phenomenon? In: Botsis AJ, Soldatos CR, Stefanis CN, editors. Suicide biopsychosocial approaches. Amsterdam: Elsevier Science; 1997. p. 177–85.
13. Yehuda R, Southwick SM, Ostroff RB, Mason JW, Giller E Jr. Neuroendocrinological aspects of suicidal behavior. Neurol Clin 1988; 6: 83–102.[Medline]
14. Pallis DJ, Holding TA. The menstrual cycle and suicidal intent. J Biosoc Sci 1976; 8: 27–33.[Medline]
15. Luggin R, Bernsted B, Petersson B, Jacobsen AT. Acute psychiatric admission related to the menstrual cycle. Acta Psychiatr Scand 1984; 69: 461–5.[Medline]
16. Forestié V, De Lignières B, Roudot-Thoraval F, Fulli-Lemaire I, Cremniter D, Nahoul K. Suicide attempts in hypo-estrogenic phases of the menstrual cycle. Lancet 1986; 2: 1357–60.[Medline]
17. Ekeberg O, Jacobsen D, Sorum Y, Aass G. Self-poisoning and the menstrual cycle. Acta Psychiatr Scand 1986; 73: 239–41.[Medline]
18. Targum SD, Caputo KP, Ball SK. Menstrual cycle phase and psychiatric admissions. J Affect Disord 1991; 22: 49–53.[Medline]
19. Gisselmann A, Ait Ameur A, Pinoit JM, Francois I, Trapet P. Tentatives de suicide et cycle menstrual: etude epidemiologiques. Ann Med Psychol Paris 1996; 154: 136–9.[Medline]
20. Shorttridge LA. Assessment of menstrual variations in working populations. Rep Toxicol 1988; 2: 171–6.[Medline]
21. Thompson K, de Castella A, Kulkarni J. Psychotic women can estimate cycle phase as well as controls. Schizophr Res 1997; 23: 185–6.[Medline]
22. Holding TA, Minkoff K. Parasuicide and the menstrual cycle. J Psychosom Res 1973; 17: 365–8.[Medline]
23. Trautman EC. The suicidal fit. Arch Gen Psychiatry 1961; 5: 98–105.
24. Wetzel RD, McClure JN Jr, Reich T. Premenstrual symptoms in self-referrals to a suicide prevention service. Br J Psychiatry 1971; 119: 525–6.[Free Full Text]
25. Wetzel RD, Reich T, McClure JN Jr. Phase of the menstrual cycle and self-referrals to a suicide prevention service. Br J Psychiatry 1971; 119: 523–4.[Free Full Text]
26. Mandell AJ, Mandell MP. Suicide and the menstrual cycle. JAMA 1967; 200: 792–3.[Abstract/Free Full Text]
27. Buckle RC, Linnane J, McConachy . Attempted suicide presenting at the Alfred Hospital, Melbourne. Med J Aust 1965; 1: 754–8.[Medline]
28. Birtchnell J, Floyd S. Attempted suicide and the menstrual cycle: a negative conclusion. J Psychosom Res 1974; 18: 361–9.[Medline]
29. Zacco M, Pacilio V, Piliego N, Jannone D. Il suicidio in rapporto al ciclo menstruale. Medicina Psicosomatica 1960; 5: 187–208.
30. Jannone D. Suicidio e tentato suicidio nei rapporti col ciclo menstruale. Policlinico (Prat) 1962; 69: 1144–50.
31. Glass GS, Heninger GR, Lansky M, Talan K. Psychiatric emergency related to the menstrual cycle. Am J Psychiatry 1971; 128: 705–11.[Abstract/Free Full Text]
32. Tonks CM, Rack PH, Rose MJ. Attempted suicide and menstrual cycle. J Psychosom Res 1968; 11: 319–23.[Medline]
33. Janowsky DS, Gorney R, Mandell AJ. The menstrual cycle: psychiatric and ovarian-adrenocortical hormone correlates: case study and literature review. Arch Gen Psychiatry 1967; 17: 459–69.[Abstract/Free Full Text]
34. Thin RNT. Premenstrual symptoms in women who attempt suicide. J R Army Med Corps 1968; 114: 136–9.
35. Bourgeois M, Pedariosse AM, Degeilh B, Dartigues JF. Incidence des periode et syndrome pre ou peri-menstruale sur le comportement suicidaire: Enquete sur 127 femmes ayant fait une tentative de suicide. Ann Med Psychol 1987; 145: 429–38.
36. Dalton K. Menstruation and acute psychiatric illness. BMJ 1959; 17: 148–69.
37. Frank R. The hormonal causes of premenstrual tension. Arch Gen Psychiatry 1931; 26: 1053–7.
38. Lester D. Suicide and the menstrual cycle. Med Hypotheses 1990; 31: 197–9.[Medline]
39. Bebbington PE. Sex and depression. Psychol Med 1998; 28: 1–8.[Medline]
40. Canetto SC, Sakinofsky I. The gender paradox in suicide. Suicide Life Threat Behav 1998; 28: 1–23.[Medline]
41. Borenstein M, Rothstein H, Cohen J. Power and Precision. Teaneck (NJ): Biostat; 1997.
42. McIntosh JE, Matthews CD, Crocker JM, Broom TJ, Cox LW. Predicting the luteinizing hormone surge: relationship between the duration of the follicular and luteal phases and the length of the human menstrual cycle. Fertil Steril 1980; 34: 125–30.[Medline]
43. Martinez AR, Zinaman MJ, Jennings VH, Lamprecht VM. Prediction and detection of the fertile period: the markers. Int J Fertil Menopausal Stud 1995; 40: 139–55.[Medline]
44. Babson AL. The IMMULITE® automated immunoassay system. J Clin Immunoassay 1991; 14: 83–8.
45. Diagnostic Products Corp. Immulite FSH. Los Angeles: Diagnostic Products Corp.; 1997.
46. Diagnostic Products Corp. Immulite LH. Los Angeles: Diagnostic Products Corp.; 1997.
47. Diagnostic Products Corp. Immulite Progesterone. Los Angeles: Diagnostic Products Corp.; 1997.
48. Diagnostic Products Corp. Immulite Estradiol. Los Angeles: Diagnostic Products Corp.; 1997.
49. van Praag HM. Biological suicide research: outcome and limitations. Biol Psychiatry 1986; 21: 1305–23.[Medline]
50. Landis JR, Kock GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33: 159–74.[Medline]
51. Collins WD. The evolution of reference methods to monitor ovulation. Am J Obstet Gynecol 1991; 165: 1994–6.[Medline]
52. Clare AW. Hormones, behavior and the menstrual cycle. J Psychosom Res 1985; 29: 225–33.[Medline]
53. Magos A, Studd J. Suicide attempts and the menstrual cycle. Lancet 1987; 1: 217–8.
54. Bancroft J. The menstrual cycle and the well being of women. Soc Sci Med 1995; 41: 785–91.
55. Mann JJ. The neurobiology of suicide. Nat Med 1998; 4: 25–30.[Medline]
56. Mann JJ, McBride PA, Brown RP, Linnoila M, Leon AC, De Meo M, Mieczkowski T, Myers JE, Stanley M. Relationship between central and peripheral serotonin indexes in depressed and suicidal psychiatric inpatients. Arch Gen Psychiatry 1992; 49: 442–6.[Abstract/Free Full Text]
57. Malone KM, Corbitt E, Li Shushua, Mann JJ. Prolactin response to fenfluramine and suicide attempt lethality in major depression. Br J Psychiatry 1996; 168: 324–9.[Abstract/Free Full Text]
58. Rubinow DR, Schmidt PJ, Roca CA. Estrogen-serotonin interactions: implications for affective regulation. Biol Psychiatry 1998; 44: 839–50.[Medline]
59. Howard R, Gifford M, Lumsden J. Changes in an electrocortical measure of impulsivity during the menstrual cycle. Pers Individual Differences 1988; 9: 917–8.
60. Kouri EM, Halbreich U. State and trait serotoninergic abnormalities in women with dysphoric premenstrual syndromes. Psychopharmacol Bull 1997; 33: 767–70.[Medline]
61. De la Gándara Martín JJ, De Diego Herrero E. Trastorno disfórico premenstrual: un estudio epidemiológico. Actas Luso Esp Neurol Psiquiatr Cienc Afines 1996; 24: 111–7.[Medline]
62. Harlow SD, Matanoski GM. The association between weight, physical activity, and stress and variation in the length of the menstrual cycle. Am J Epidemiol 1991; 133: 39–49.
63. Lenton EA, Landgren BM, Sexton L. Normal variation in the length of the luteal phase of the menstrual cycle: identification of the short luteal phase. Br J Obstet Gynaecol 1984; 91: 685–9.[Medline]
64. Carr BR. Disorders of the ovary and female reproductive tract. In: Wilson JD, Foster DW, editors. Williams’ textbook of endocrinology. 8th ed. Philadelphia: WB Saunders; 1992. p. 733–98.
65. Demetriou JA. Progesterona. In: Pesce AJ, Kaplan LA, editors. Química clínica métodos. Buenos Aires: Editorial Médica Panamericana SA; 1990. p. 266–71.
66. Adekunle AO, Matson C, Collins WP, Whitehead MI. Single serum progesterone measurement in the mid-luteal phase as an index of ovulation. S Afr J Med Sci 1987; 16: 157–62.
67. Grinsted J, Jacobsen JD, Grinsted L, Schantz A, Stenfoss HH, Nielsen SP. Prediction of ovulation. Fertil Steril 1989; 52: 388–93.[Medline]
68. Wright DM, Kesner JS, Schrader SM, Chin NW, Wells VE, Krieg EF Jr. Methods of monitoring menstrual function in field studies: attitudes of working women. Reprod Toxicol 1992; 6: 401–9.[Medline]

This article has been cited by other articles:

				B. Mullen and J. M. Smyth Immigrant Suicide Rates as a Function of Ethnophaulisms: Hate Speech Predicts Death Psychosom Med, May 1, 2004; 66(3): 343 - 348. [Abstract] [Full Text] [PDF]

				E. Baca-Garcia, C. Diaz-Sastre, A. Ceverino, J. Saiz-Ruiz, F. J. Diaz, and J. de Leon Association Between the Menses and Suicide Attempts: A Replication Study Psychosom Med, March 1, 2003; 65(2): 237 - 244. [Abstract] [Full Text] [PDF]

				E. Baca-Garcia, C. Diaz-Sastre, J. Saiz-Ruiz, and J. de Leon Influence of Psychiatric Diagnoses on the Relationship Between Suicide Attempts and the Menstrual Cycle Psychosom Med, May 1, 2001; 63(3): 509 - 510. [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Baca-García, E. Articles by Saiz-Ruiz, J. Search for Related Content PubMed PubMed Citation Articles by Baca-García, E. Articles by Saiz-Ruiz, J. Related Collections Depression Sexual Medicine: Female Suicide or Suicidal Behavior