Salivary Progesterone Levels Before Menarche: A Prospective Study of Adolescent Girls

Abstract

Objective: Our objective was to determine the frequency of ovulatory salivary progesterone levels before menarche in healthy girls.

Design: We recruited a prospective cohort of midpubertal, premenarcheal girls. Participants collected weekly saliva samples until first menses or for 12 months. Menstrual cycles were considered to have ovulatory salivary progesterone levels if values were greater than 0.100 ng/ml (0.318 nmol/liter) 0–14 d before menarche.

Settings: Participants collected saliva samples weekly at home and attended monthly clinic visits.

Patients: Patients included 63 premenarcheal females, age 9–15 yr, with body mass index higher than the fifth percentile for age and Tanner stage III or greater for both pubic hair and breast development.

Main Outcome Measures: Frequency of ovulatory levels of salivary progesterone before menarche was assessed.

Results: Fifty-five girls completed the study, and 43 experienced menarche. Of the 42 girls who reached menarche and provided a sample within 14 d of menarche, five (12%) had ovulatory progesterone levels. The mean anovulatory salivary progesterone level was 0.041 ng/ml (0.130 nmol/liter; range, 0.012–0.078 ng/ml, 0.038–0.248 nmol/liter), and the mean ovulatory level was 0.194 ng/ml (0.617 nmol/liter; range, 0.125–0.343 ng/ml, 0.398–1.09 nmol/liter). The levels of estrogen, testosterone, and 17-hydroxyprogesterone were higher and the mean BMI was lower in those with ovulatory progesterone levels vs. those with anovulatory levels.

Conclusion: A significant proportion of girls displayed ovulatory levels of progesterone before menarche. More research is needed to clarify the hormonal events that occur in the perimenarcheal time period.

Twelve percent of pubertal girls had ovulatory levels of salivary progesterone prior to menarche, as well as higher levels of estradiol, testosterone, and 17-hydrogyprogesterone.

Over the last decade, several studies have examined the age of onset of pubertal development and menarche in the United States and other developed countries to assess whether factors such as race and body weight might explain earlier onset of maturation (1,2,3,4,5). An increased understanding of the biochemical events that trigger female sexual development, and specifically the initiation of ovulatory menstrual cycles, is important in understanding the origins of conditions such as dysfunctional uterine bleeding, polycystic ovary syndrome, and breast cancer (6,7). Several investigators have studied the frequency of ovulation in the first 2 yr after menarche, and two studies have examined hormonal changes using urinary samples before and after menarche (8,9,10,11,12). However, the frequency of elevated progesterone levels that seem to imply ovulatory events preceding the first menstrual period is still not well understood, largely because of the limited feasibility of longitudinal studies requiring frequent sampling in young adolescents. We sought to determine the frequency of ovulatory levels of progesterone immediately before menarche by using salivary hormone levels.

Patients and Methods

Premenarcheal females, age 9–15 yr old, were recruited through advertisements placed in local publications, hospital-based outpatient pediatric clinics, and local primary care offices. Participants were screened for the following exclusion criteria: presence of previous menstrual bleeding, anomalous reproductive or sexual development, and intensive (elite) athletic activity. A study enrollment visit included a medical history and a brief physical examination, including Tanner staging of breast development and pubic hair. Height, weight, and vital signs were measured and body mass index (BMI) calculated. Subjects with BMI above the fifth percentile for age and Tanner stage III or above for both pubic hair and breast development met study inclusion criteria and were instructed on how to complete the study protocol. The study was approved by the Children’s Hospital Boston Committee on Clinical Investigation. The parents of all participants gave informed consent, and the participants gave assent.

Participants collected weekly saliva samples, using 2-ml Eppendorf tubes, Parafilm M to induce salivation, and nonreactive plastic straws. Subjects were instructed not to collect samples within 30 min of eating or chewing gum. Dated samples were stored in the home freezer and transported to the clinic using a Styrofoam cooler and ice packs. Participants used a diary to record the dates samples were collected and presence of vaginal bleeding. They remained in the study for 12 months or until first menses, whichever came first.

Commercially available salivary ELISA kits were used to measure concentrations of estradiol, 17-hydroxyprogesterone (17-OHP), testosterone, and progesterone (ALPCO Diagnostics, Salem, NH). The saliva samples were stored at −70 C. Samples were thawed in a room-temperature water bath, vortexed at high speed for 30 sec, and centrifuged (10,000 × g) for 3 min. Aliquots of 25 μl were used for 17-OHP and progesterone, 50 μl for testosterone, and 100 μl for estradiol measurements, which were processed according to the manufacturer’s instructions. Coefficients of variation for intraassay and interassay precision were reported by the manufacturer or derived from manufacturer’s data as follows: 17β-estradiol (5.9 and 3.4%), 17-OHP (6.2 and 5.6%), testosterone (6.0 and 8.7%), and progesterone (9.4 and 10.6%).

Premenarcheal cycles were determined to have ovulatory progesterone levels if the value was from 0.100–0.500 ng/ml (0.318–1.590 nmol/liter, ALPCO Diagnostics’ expected normal luteal-phase range) 0–14 d before menarche. This is a conservative cutoff for a normal (i.e. ovulatory) luteal salivary progesterone level; Ishikawa et al. (13) suggested a cutoff of 0.059 ng/ml, or 0.189 nmol/liter. Lipson et al. (14) reported an average midluteal salivary progesterone for women aged 18–24 with regular (presumably ovulatory) cycles of 0.093 ng/ml, or 0.296 nmol/liter, although they cautioned against using this as a standard for determination of fertility given the variability of their assay. For girls with multiple progesterone values within the selected range, the highest value was included in the analysis.

To compare characteristics between the girls with ovulatory and anovulatory progesterone levels, we used Fisher’s exact test for categorical variables and the Student’s t test for continuous measures, corroborated for skewed samples by the Wilcoxon two-sample test. All computations were performed with SAS version 9.1 (SAS Institute, Cary, NC).

Results

Sixty-three girls met the inclusion criteria and enrolled in the study. At enrollment, the mean age was 12.0 yr, mean BMI was 21.1 kg/m², and the majority of the patients were Tanner stage III or IV for breast and pubic hair development. Fifty-five girls completed the protocol, 43 with menarche and 12 without. The remaining eight girls dropped out or were lost to follow-up. At menarche, the mean age was 12.6 ± 1.2 yr. Of the 43 girls who reached menarche, 42 had a salivary progesterone value within 14 d of menarche. Five girls (12%) were found to have an ovulatory level in the premenarcheal cycle. There was no significant difference between the ovulatory and the anovulatory level progesterone groups in the median number of days the analyzed sample was taken before menarche or in the distribution of age, race, or Tanner Stage. The levels of estradiol, testosterone, and 17-OHP were significantly higher in the group with ovulatory progesterone levels, and the mean BMI of the ovulatory progesterone group was lower, as shown in Table 1. Two girls in the anovulatory level group had salivary progesterone levels at or above the cutoff of 0.100 ng/ml more than 14 d before menarche, as shown in Fig. 1.

Table 1.

Characteristics of 42 participants who provided salivary progesterone within 14 d before menarche, comparing those with ovulatory with those with anovulatory levels of progesterone

	Ovulatory (n = 5)	Anovulatory (n = 37)	P
Median (quartile 1–3)
Age at menarche, yr	14.4 (12.3–14.5)	12.3 (11.8–13.3)	0.21a
BMI at menarche, kg/m2	19.2 (17.2–19.5)	22.0 (19.1–23.8)	0.04a
Sample time, d before menarche	4 (2–4)	7 (3–10)	0.08a
Salivary progesterone, ng/mlc	0.17 (0.14–0.19)	0.04 (0.03–0.05)
Salivary estradiol, pg/mld	0.85 (0.46–6.38)	0.32 (0.18–0.61)	<0.05a
Salivary testosterone, pg/mld	43 (40–52)	11 (6–19)	<0.005a
Salivary 17-OHP, pg/mld	75 (64–81)	20 (10–37)	<0.001a
n (%)
Race
African-American	1 (20)	8 (22)	0.99b
Caucasian	4 (80)	26 (70)
Mixed	0 (0)	2 (5)
Unknown	0 (0)	1 (3)
Tanner stage, breasts
III	5 (100)	24 (65)	0.38b
IV	0 (0)	12 (32)
V	0 (0)	1 (3)
Tanner stage, pubic hair
III	2 (40)	23 (62)	0.38b
IV	3 (60)	14 (38)

^aComparison of distribution in ovulatory vs. anovulatory by Wilcoxon two-sample test. 

^bComparison by Fisher’s exact test. 

^cMaximum within 14 d preceding menarche; ovulatory cycle defined by this value ≥0.1 ng/ml. 

^dMeasured on same date as maximum progesterone. 

Figure 1.

Salivary progesterone levels before menarche for seven participants who had at least one value above the ovulatory level threshold (≥0.10 ng/ml). Shapes represent individual participants. Five participants depicted in red had at least one value of 0.10 ng/ml or higher in the 14 d immediately preceding menarche. Participants depicted in black reached the threshold only at times earlier than 14 d before menarche.

Discussion

Our study identified a significant proportion of girls, 12%, with evidence of salivary levels of progesterone elevated enough to suggest the presence of ovulation before menarche. We did not include ultrasound confirmation of ovulation as part of this small study, but notably, this subset of girls also had higher levels of estradiol, testosterone, and 17-OHP. The proportion of girls with premenarcheal ovulatory levels of progesterone is similar to data collected on the frequency of ovulation in the first year after menarche (8,10,15,16,17). However, much less is known about the frequency of ovulation before menarche. In a recent study of daily urinary hormone collections of six girls ages 11–13 who provided sufficient data for analysis, Zhang and colleagues (11) found LH surges and small-amplitude progesterone increments [mean 1.33 μg/mg creatinine (Cr)] before menarche compared with progesterone levels of 3.95 μg/mg Cr in the 6–12 months after menarche and 22.18 μg/mg Cr in adult controls. Two girls were found to have LH surges in the adult range (26.7 and 21.8 mU/mg Cr) with estrogen and progesterone rises before menarche; two others were documented to have LH surges without progesterone rises. Notably, after menarche, the luteal phase increased from 2–4 to 11–12 d. In an earlier study, Legro and colleagues (18) had collected urine every 6 months; they found a mean weight at menarche of 47.2 kg (BMI of 19.3 kg/m²) with rises in LH, testosterone, and estradiol before menarche (progesterone levels were not measured). We used a larger sample size than Zhang and colleagues (11) but sampled hormone levels much more frequently than Legro and colleagues (18), so our data provide a more detailed viewpoint on the immediate perimenarchal period. The documented ovulatory levels of progesterone occurring before menarche also provide an explanation regarding occasional reported cases of premenarcheal pregnancy (19). A better understanding of whether particular girls do in fact ovulate before menarche may fill knowledge gaps regarding what constitutes normal reproductive development and possibly risk for certain diseases. We also found that age and race were not significantly different between those with ovulatory and anovulatory progesterone levels; contrary to our expectations, the mean BMI of the ovulatory progesterone group was lower than that of participants in the anovulatory progesterone group. These findings are intriguing and are consistent with the observation that normal-weight American girls may experience rapid maturation of the reproductive endocrine axes independent of body composition (11,18).

Our study offers a unique window on the immediate perimenarcheal period using a relatively large, nutritionally replete, and racially diverse sample that is more representative of the current U.S. population than older studies. There are several limitations to our study. The samples were not obtained during the same day of the hypothesized luteal phase or at the same time of day. It is well known that salivary hormone assays display greater variability than urine or serum assays, although their noninvasiveness and ease of use contributed to the unique study design and ability to obtain frequent samples over a long period of time (20). Although not part of this study, it would have been valuable to track the subsequent cycles to determine whether ovulatory progesterone levels before menarche predicted onset of more regular or ovulatory cycles over the next year. We used the salivary progesterone level of 0.100 ng/ml (0.318 nmol/liter) as our threshold to determine ovulatory progesterone levels, as suggested by the manufacturer of the commercially available assay we used. A range of luteal salivary progesterone values have been published to date, and others have suggested lower cutoff levels (13,14). It is possible that a larger proportion of first cycles were ovulatory and that the elevated progesterone level was missed. Our classification system based on the cutoff value of 0.1 ng/ml, however, seemed to differentiate the two groups because we also found that there were significant differences in the estrogen, testosterone, and 17-OHP levels.

We also used the usual 14-d luteal phase cutoff to decide which cycles would be classified as potentially ovulatory levels. However, two girls in our sample had high salivary progesterone levels earlier than 14 d before menarche, as shown in Fig. 1. We do not have a good explanation for these elevations in progesterone without subsequent withdrawal bleeding other than that they could potentially be consistent with the Zhang and colleagues (11) study that found that girls may have small-amplitude progesterone increments before menarche without withdrawal bleeding. Use of ultrasound might have revealed additional evidence of follicular development and would be useful to pursue in future studies.

In summary, this study demonstrates that the first menstrual cycle was preceded by an ovulatory level of salivary progesterone in 12% of healthy Tanner III and above pubertal girls in our sample. Further studies of changes in serum hormone levels, potentially accompanied by ultrasound monitoring, would shed additional light on the incidence of ovulatory cycles before the first cycle and during the first year of menses.