Participants and study design

The BioCycle Study (20052007) was a prospective cohort study designed to evaluate the relationship between reproductive hormones and oxidative stress levels throughout the menstrual cycle [26, 27]. The participant cohort consisted of 259 regularly menstruating, healthy women between the ages of 1844years, recruited from western New York. Participants were recruited in a variety of ways, including: advertising in clinical practices and the University at Buffalo student health center, paid advertising in print media, radio and television interviews, notices sent via list serves, and flyers at the university and throughout the region. For those interested, an initial screening phone call was conducted, followed by a mailing and an in-person visit. Exclusion criteria included factors that may interfere with a normal menstrual cycle or vitamin levels, such as: use of oral contraceptives during the study period or in the previous three months; use of Depo provera, implant or IUD in previous twelve months; current use of vitamin, mineral, or herbal supplements; use of prescription medications; pregnancy or breast feeding in the previous six months; reported attempts to conceive in the previous six months; diagnosis of uterine abnormalities or chronic conditions, such as ovulatory disorders and premenstrual dysphoric disorder (PMDD); and a self-reported body mass index (BMI) of <18 or >35kg/m2 at screening [27]. Those eligible and interested after the screening visit were scheduled for a baseline enrollment visit 12weeks prior to the start of their next menses. Women were followed for one (n=9) or two (n=250) menstrual cycles (Additional file 1: Figure S1).

Women completed up to 8 clinic visits per cycle for up to 2 cycles. Study visits were scheduled using fertility monitors (Clearblue Easy Fertility Monitor; Inverness Medical, Waltham, Massachusetts) to coincide with critical phases of the menstrual cycle, including menstruation; mid- and late follicular phases; luteinizing hormone (LH) surge (predicted ovulation); and early, mid-, and late luteal phases [28]. At each of these visits, fasting blood samples were collected from participants from which the antioxidant vitamins and oxidative stress concentrations were measured. Participants were highly compliant with the study protocol; 94% of the women completed 7 or 8 clinic visits, and 100% completed at least 5 clinic visits per cycle.

The University at Buffalo Health Sciences Institutional Review Board (IRB) approved this study and served as the IRB designated by the National Institutes of Health under a reliance agreement. All participants provided written informed consent. Further details of the study design are described elsewhere [27].

Ascorbic acid (vitamin C), retinol (vitamin A), and - and - tocopherol (vitamin E) were measured in all blood samples taken from participants across a menstrual cycle and subsequently averaged to reflect the mean levels across a cycle. Total ascorbic acid was determined by the dinitrophenylhydrazine (DNPH) method. Samples for ascorbate analysis were stabilized immediately following phlebotomy and centrifugation by adding 0.5mL of heparin plasma to 2.0mL of 6% meta-phosphoric acid and centrifuging at 3000g for 10min. Clear supernatant was decanted and frozen at 80C for analysis. The absorbance of each DNPH derivatized sample was determined at 520nm on a Shimadzu 160U spectrophotometer (Shimadzu Scientific Instruments, Inc.). Across the study period, the coefficient of variation (CV) for this test reported by the laboratory was 10%.

Fat-soluble vitamins (including retinol, and vitamin E components: - and -tocopherol) were measured at the Kaleida Health Center (Buffalo, New York) simultaneously in serum using high performance liquid chromatography with photodiode array detection [29]. -tocopherol was also detected but was below the lower limit of quantification for our assay (0.28). The limits of detection were 0.0054 for retinol, 0.0768 for -tocopherol, and 0.1052 for -tocopherol. The CV for these tests across the study period were <6% for retinol and <2% for - and -tocopherol. Continuous monitoring of standard reference material 968c from the National Institute of Standards and Technology (NIST) and participation in the NIST Micronutrients Measurement Quality Assurance Program provided external checks on analytical accuracy.

Mean concentrations of antioxidants, including vitamin A, vitamin C, -tocopherol, and -tocopherol, were calculated per cycle and were used in all analyses. Overall median concentrations were also compared with levels reported previously by reproductive aged women (i.e., 2039years) in the 2012 National Health and Nutrition Examination Survey (NHANES) to assess the comparability of our results with those of a nationally representative population [30].

Plasma free F2-isoprostane, a breakdown product of ROS and a marker of oxidative stress, was measured with a gas chromatography-mass spectrometrybased method by the Molecular Epidemiology and Biomarker Research Laboratory (University of Minnesota, Minneapolis, Minnesota) (CV=9.4%).

Frequency and severity of 20 premenstrual symptoms was assessed through questionnaires completed at four time points of each menstrual cycle: menses, follicular phase, peri-ovulation, and luteal phase (Additional file 2: Figure S2). Participants recalled the occurrence and severity of symptoms in the prior week. The symptoms included in this assessment were: sadness, crying spells, anger, nervousness, insomnia, tension, abdominal bloating, cravings of chocolates, cravings of sweets, cravings of salty foods, cravings of other foods, breast tenderness, lower abdominal cramping, general aches, backache, headache, acne outbreaks, change in appetite, fatigue, and swelling of the hands/feet. The severity of each symptom was ranked by the participant on a scale of 0 to 3 (0=none, 1=mild, 2=moderate, 3=severe). The symptoms included in this questionnaire were adapted from validated surveysincluding the Daily Record of Severity of Problems (DRSP) and the Premenstrual Symptoms Screening Tool (PSST)but slightly modified, given that DRSP and the PSST were designed to identify patients with PMDD specifically (a population excluded in our study) [31,32,33,34].

We categorized severity as none/mild (reference group) or moderate/severe to estimate odds of having a moderate or severe symptom during the premenstrual week. We then calculated severity scores for groups of related symptoms by summing the severity score of symptoms (as reported in the premenstrual week) within each grouping to generate an overall score. The groupings were established based on clinical expertise and included: depression (sadness, crying spells, anger) and anxiety (nervousness, insomnia, tension); hydration (abdominal bloating) and cravings (chocolate cravings, sweets cravings, salty food cravings, other food cravings); pain (breast tenderness, lower abdominal cramping, general aches, backache, headache); and other (acne outbreak, change in appetite, fatigue, swelling of hands or feet).

Overall PMS severity was evaluated using four different approaches, which utilize information on all symptoms from the luteal and follicular phase questionnaires from each cycle: (1) 5 or more moderate or severe symptoms during the luteal phase; (2) 8 or more moderate or severe symptoms during the luteal phase; (3) 3 or more moderate or severe symptoms where the luteal phase score was 30% greater than the follicular phase and at least one symptom was psychological (referred to as PMS-1 in the tables and results); and (4) 5 or more moderate or severe symptoms where the luteal phase score was 30% greater than the follicular phase and at least one symptom was psychological (referred to as PMS-2 in the tables and results) [8, 35, 36]. When summing the number of moderate or severe symptoms for each cycle, each of the individual cravings symptoms were combined into a single variable. These criteria were based upon various defintions of PMSincluding those of the National Institute of Mental Health, [37] the American College of Obstetrics and Gynecology, [38] the American Psychiatric Association, [39] and the International Society for Premenstrual disorders (ISPMD) [40]which were further expanded and implemented in studies such as Gollenberg et al. [35] and Borenstein et al. [8] Of note, these approaches attempt to establish the necessary temporality between pre- and post-menstrual symptoms in line with traditional PMS definitions and diagnoses that assume resolution of symptoms within 12days of the onset of menses.

At study enrollment, a trained research assistant measured height and weight for the calculation of BMI using standardized protocols. Demographics such as age, race, education, smoking habits, reproductive history, and physical activity were also collected at baseline through self-reported questionnaires. Physical activity was assessed at baseline using the International Physical Activity Questionnaire, and estimated for high, moderate, and low levels of activity based upon accepted cutoffs [41]. Dietary information was obtained using 24-h recalls (up to 4 times per cycle) and analyzed using the Nutrition Data System for Research software (version 2005) developed by the Nutrition Coordinating Center of the University of Minnesota (Minneapolis, Minnesota). Cycle-averaged measures of total energy (kcal/day) and fiber (g/day) were used in these analyses, as we previously found these intakes do not vary significantly across the cycle [42]. All covariates assessed had at least a 95% response rate.

Demographic characteristics were compared between those with <5 versus 5 moderate or severe symptoms during the luteal phase of either study menstrual cycle, and between those with <8 versus 8 moderate or severe luteal phase symptoms in either cycle. Repeated measures ANOVA and McNemars tests were used for comparisons.

We estimated associations between mean antioxidant concentrations and F2-isoprostane concentrations from each menstrual cycle and odds of reporting a moderate/severe symptom during the premenstrual week for each cycle using generalized linear models. Next, we evaluated associations between antioxidant concentrations, F2-isoprostane concentrations, and scores for symptom severity within groups during the premenstrual week (e.g., depression, cravings, pain) using linear mixed models. We used generalized linear models to assess the association between mean antioxidant concentrations, F2-isoprostane concentrations, and overall PMS severity in each cycle using the four different classifications of PMS severity (5 or more moderate or severe symptoms, 8 or more moderate or severe symptoms, PMS-1 criterion, PMS-2 criterion). All models were adjusted for age, race, BMI, physical activity, smoking status, alcohol use, pain reliever use, and average total energy intake per cycle and accounted for repeated measures (i.e., multiple cycles per woman). Results were adjusted for multiple comparisons using the false discovery rate (FDR). An alpha of 0.05 was considered statistically significant. As antioxidants and oxidative stress measures have been shown to vary somewhat over the menstrual cycle [26, 43], we also evaluated associations between time-varying measures of antioxidants and oxidative stress, with time-varying symptoms as a sensitivity analysis. Splines were used to evaluate the assumption of linearity. We did not find evidence to suggest that linear modeling was inappropriate (e.g., quadratic or restricted cubic spline modeling did not help explain the associations in our population). All statistical analyses were calculated using SAS 9.4 (SAS Institute, Cary, North Carolina).

Read this article:
Serum antioxidant vitamin concentrations and oxidative stress markers associated with symptoms and severity of premenstrual syndrome: a prospective...

Comments are closed.