Table 2.
Mean or Median Values of Candidate Predictor Variables at Initial Measurement and at Each Time Interval of Follow-upa
| Participant Characteristics and Biomarker Level | Time at Which Observations Were Made |
||||
|---|---|---|---|---|---|
| Referenceb | ≥2 y Before FMP | Between 2 and 1 y Before FMP | Between 1 y Before FMP and FMP | Between FMP and 1 y After FMP | |
| Sample size | 554 | 1390 | 376 | 324 | 424 |
| Age, y | 46.3 (2.4) | 48.4 (2.4) | 50.8 (2.3) | 51.8 (2.2) | 52.8 (2.3) |
| Initial FSH level, IU/L | 15.4 (11.7, 22.1) | ||||
| Initial E2 level, pg/mL | 56.7 (32.8, 81.3) | ||||
| FSH ratio | 1.19 (0.81–1.84) | 1.75 (1.03–2.98) | 2.54 (1.39–4.53) | 5.03 (3.03–7.74) | |
| E2 ratio | 0.86 (0.47–1.46) | 0.78 (0.40–1.69) | 0.68 (0.36–1.57) | 0.41 (0.21–0.83) | |
| Log(FSH ratio) | 0.20 (0.70) | 0.52 (0.88) | 0.90 (0.87) | 1.57 (0.75) | |
| Log(E2 ratio) | −0.15 (0.97) | −0.14 (1.22) | −0.28 (1.20) | −0.85 (1.10) | |
| MT stage | |||||
| Premenopause | 332 (64.2%) | 322 (23.2%) | 28 (7.5%) | 4 (1.2%) | 1 (0.2%) |
| Early perimenopause | 185 (35.8%) | 1031 (74.2%) | 267 (71.0%) | 154 (47.5%) | 35 (8.3%) |
| Late perimenopause | 37 (2.7%) | 81 (21.5%) | 166 (51.2%) | 388 (91.5%) | |
| In-windowc | 554 (100%) | 1362 (98.0%) | 306 (81.4%) | 172 (53.1%) | 44 (10.4%) |
| Smoking | 66 (12.8%) | 154 (11.2%) | 32 (8.6%) | 44 (13.8%) | 39 (9.2%) |
| BMI, kg/m2 | |||||
| <22 | 125 (24.5%) | 310 (22.7%) | 82 (22.1%) | 52 (16.3%) | 78 (18.8%) |
| 22 ≤ BMI < 25 | 130 (25.4%) | 327 (23.9%) | 89 (24.0%) | 79 (24.7%) | 94 (22.7%) |
| 25 ≤ BMI < 30 | 115 (22.5%) | 354 (25.9%) | 91 (24.5%) | 80 (25.0%) | 100 (24.2%) |
| ≥30 | 141 (27.6%) | 375 (27.5%) | 109 (29.4%) | 109 (34.1%) | 142 (34.3%) |
| Initial NTx, nM BCE/mM creatinine | 29.6 (22.1–38.3) | ||||
| NTx ratio | 1.02 (0.78–1.32) | 1.08 (0.81–1.48) | 1.10 (0.84–1.52) | 1.38 (0.98–1.95) | |
| Log(NTx ratio) | 0.01 (0.47) | 0.06 (0.49) | 0.11 (0.48) | 0.33 (0.47) | |
With the exception of NTx, values are from the 554 participants who had FSH and E2 levels assessed because this larger sample was used to derive the final models. Each woman contributed data for at least 2 observation times. The NTx values are from the subset of women (n = 472) who had all 3 candidate predictors (FSH, E2, and NTx) measured; mean or median values of the variables shown did not differ between the larger and the smaller analysis samples (data not shown). Means and SDs are shown for continuous variables; numbers and percentages are provided for categorical variables; for the ratios of predictors, medians and IQRs are given.
Reference value is the earliest value obtained at a time point greater than 2 years before the FMP. FSH, E2, and NTx ratios are computed as current value divided by the reference measurement.
In-window indicates whether the sample was collected between days 2 to 5 of the menstrual cycle and applies only to observations made during late perimenopause; in earlier MT stages, observations made outside this window were not used (see Materials and Methods for details).