What this study adds
We believe that this article shows the ability of participants to successfully evaluate and collect biospecimens on a rigorous, biologically-determined schedule in their own homes with an acceptable degree of compliance. This proved vital to our study of the periconceptional environment, and it is extendable to many other study objectives.
The periconceptional and prebirth environment has come under increasing scrutiny in recent years, with broadening understanding regarding the ways in which exposures during these periods may shape both fertility and future health and development.1 A growing body of research is exploring relationships between various exposures and reproductive and fetal outcomes,2–6 but this research has produced as many questions as answers.
One difficulty of early exposure research is the transient nature of many exposures of interest.6,7 Endocrine disrupting chemicals, which have been implicated in a wide range of fertility issues and health outcomes for developing neonates, often possess biological half-lives measured in hours.8 This fact, combined with the limited state of current understanding regarding exposure windows of importance in the periconceptional and gestational periods, makes such relationships difficult to investigate and requires specialized study designs.2,9
Longitudinal cohort studies, beginning preconception, hold promise regarding periconceptional exposures and birth outcomes.2,10–12 The utility of such a design depends on sufficient exposure measurements, timed such that exposures during the narrow band of fertile time are fully assessed.10,12 The number of measurements required can become costly for investigators and unwieldy for subjects, leading to noncompliance and loss to follow-up.13,14 These hurdles can be offset by the use of at-home biospecimen collection protocols, in which subjects, having been trained in collection procedures and provided with sampling schedules and materials, collect specimens at home for later collection by study investigators.13–15
This article examines whether such a schema, used for the collection of urine from both members of pregnancy-seeking couples at carefully scheduled, participant-identified intervals surrounding days of fertility and ovulation, can be performed with high compliance.
The results we present were collected from within an observational, prospective cohort study designed to accomplish home-collected, individual-level biomonitoring among heterosexual couples during the sensitive windows of conception, implantation, and very early pregnancy. Study procedures have previously been described, including recruitment and informed consent.7 The study was approved by the University of Utah Institutional Review Board. Briefly, women were taught to recognize changes in cervical fluid that prospectively indicate the onset of the fertile window, and were trained to identify the estimated day of ovulation, which may be also be the estimated day of conception (EDO/C), during the menstrual cycle (based on the Peak Day method).16 These observations were recorded on a daily fertility chart that was turned in to the study staff at the end of the cycle (see Supplemental Digital Content 1; http://links.lww.com/EE/A46). Men and women both were asked to collect daily first-morning urine samples (first void upon waking) from the first day of fertile-quality cervical fluid throughout the fertile window until the EDO/C+2 days. Men were asked to discontinue collecting after EDO/C+2 days (the end of the biologically relevant time period for conception), and women were asked to continue to collect for the remainder of the menstrual cycle until the onset of the next menses or until she had a positive home pregnancy test (QuickVue, Quidel, San Diego, California), tested at EDO/C+18 days.
Women reported the dates of the onset of the fertile window and their EDO/C to the study staff as soon as identified and this was recorded as the real-time or participant-identified EDO/C and fertile window. Daily fertility charts were later independently reviewed by two expert reviewers who objectively identified the woman’s EDO/C and fertile window (investigator-identified sampling days). Premenstrual fluid occurring during the last 7 days of the cycle can often mimic cervical fluid,16,17 so expert reviewers excluded any fertile-type cervical fluid observations in the last 7 days before menses when determining the objective EDO/C.
Dates of collected urine specimens were compared to the daily fertility charts to determine which urine specimens were correctly collected on expected days, which were missed and not collected when they were expected, and which were collected on days that were not expected (either early/before the onset of the fertile window, late/after the fertile window [men] or during the first days of following cycle [women], and other/date cannot be determined). Specimens that were missed on the first day of the fertile window were noted separately from specimens missing on other days, because first-morning specimen collection may not have been possible if the woman first noticed fertile-quality cervical fluid later in the day. Additionally, a woman’s specimen that was collected on the first day of menses for the following cycle was considered expected and correctly collected for the previous cycle because she likely noticed the onset of menses later in the day after the sample was already collected. The proportions of correctly and incorrectly collected samples, according to both investigator-identified and participant-identified windows, were compared according to participant demographics and cycle to assess overall compliance.
As shown in Figure, 6,246 urine samples were submitted from 286 cycles collected by 170 couples. Forty samples were excluded for inadequate labeling leaving 6,206 samples available for analysis; 4,727 collected by women and 1,479 collected by men. Among the 170 couples and 286 cycles for which we have samples, two cycles belonging to one couple are not included in participant-identified assessments because, although urine was collected and a chart was turned in, no real-time participant-identified EDO/C was reported. Additionally, 10 cycles belonging to nine couples are not included in investigator-identified assessments because an objective EDO/C could not be identified for the following reasons: (1) no fertile-fluid occurred before the last 7 days of the cycle (n = 3); (2) fertile-fluid was not correctly recorded on the chart (n = 2); or (3) the chart is missing (n = 5), reducing the final count to 274 cycles. One hundred sixty-four individuals have cycles included in both participant- and investigator-identified assessments, five individuals are not included in participant-identified but are included in investigator-identified, and five are included in investigator-identified but not included in participant-identified, for a total of 169 individuals in each stratum, noting that the couples excluded from the investigator-identified results are not the same as those excluded from the participant-identified results.
As shown in Table 1, the expert reviewer and participant agreed with the exact date of the first day of the fertile window in 65% (n = 178) of cycles, and 87% for agreement ±3 days of the participant-identified date. The expert reviewer and participant agreed with the exact date of the EDO/C in 75% (n = 204) of cycles, and 93% for agreement ±3 days of the participant-identified date.
The average duration of the fertile window, as identified by the expert reviewer, was 6.9 days (SD = 4.0) and the mean was 6 days (data not shown). Nearly 17% of cycles were 3 days of length or less (2.9% 1 day, 4.0% 2 days, and 9.5% 3 days). The average participant identified fertile window was slightly shorter with an average of 5.2 days (SD = 2.3), median of 5 days, and 18.2% of cycles being 3 days of less.
Compliance among women
Using investigator-identified EDO/C dates and fertile windows (Table 2), 5,329 urine samples were expected and 4,546 were collected. Eighty-two percent of expected samples were correctly collected on expected days, 15% were not collected, and 3% of expected samples missed were to have been collected on the first day of the fertile window. A total of 5% of collected samples were collected on days that were not expected (i.e., samples collected early, late, or where dates could not be determined or appeared incorrect as written). The proportion of correctly collected samples did not vary by woman’s age, and varied slightly by income level (range 77%–92%) and education level (range 81%–90%). The majority of women (n = 116) had income ranging from $20,000 to $74,999 and correctly collected 80%–82% of expected specimens. Similarly, most women were college graduates (n = 117) and correctly collected 81% of expected specimens. Compliance was similar but slightly higher when examined based on the EDO/C identified and reported by the woman in real-time (participant-identified EDO/C dates and fertile windows). Five thousand two hundred nineteen specimens were expected and 4,650 were collected, of which 85% were correctly collected on expected days. Among all categories for age, income, and education, and by both participant- and investigator-identified assessments, 0%–6% of the collected samples were collected on days that were not expected and were either early, late, or collected on days that were not consistent with cycle dates.
The proportion of correctly collected specimens was slightly (3%) less from the first cycle of collection to the second in both participant- and investigator-identified assessments (Table 3).
Compliance among men
Men correctly collected 59% (1,421 out of the 2,049 expected samples) of samples when examined using investigator-identified sampling days, and 69% (1,468 out of the 1,763 expected samples) of samples when examined using participant-identified sampling days (Table 2). There was little variability in the proportion of correctly collected specimens when stratified by education, income, and age; with proportions ≥50% in all groups by both participant-identified and investigator-identified measurement. Men missed sample collection on the first day of the woman’s fertile window with proportions ranging from 9% to 12%, and frequently continued specimen collection past the EDO/C+2, with proportions of late samples ≥10% in each stratum. The proportion of correctly collected specimens was very similar in the first cycle and the second cycle, whereas the proportion of specimens collected late decreased from 15% to 9% (Table 3).
This study, like others in a growing body of research, demonstrates that a prospective pregnancy study can successfully gather intensively scheduled urine samples from both members of the couple.4,10,14,15,18 It further demonstrates that, in a study setting, women can be taught to understand biomarkers of ovulation and record their ovulatory cycles with a high degree of accuracy, building on past research with similar findings.16,18,19 Additionally, we demonstrate that women can apply their observations of ovulatory signs to trigger biospecimen collection at their own biologically relevant times.
Studies seeking to examine the periconceptional environment have several options for urine collection. Women or couples can be instructed to collect urine daily, which may minimize the risk of missing important fertility windows but can be resource-intensive;20–23 fixed calendar or menstrual cycle days can be used, simplifying collection instructions but not capturing all fertile days;4,20 and various methods to detect the hormonal surge leading up to ovulation can be employed, allowing exposure ascertainment to begin before ovulation occurs. Hand-held electronic monitors, which track luteinizing hormone and an estradiol metabolite, are the most common method used by studies to track ovulation; these have been shown effective at restricting urine collection to the fertile window,14,15 but may prove expensive in a large study. This study used women’s observations of their cervical mucus to detect impending ovulation, a method validated against hand-held urine monitors in previous studies and which is both conservative of study resources and accurate.16,22,24,25
In this study, women correctly collected 82% of samples on expected days, as identified by expert reviewers (investigator-identified); men correctly collected as expected by reviewers 59% of the time. This compares well with other studies requesting daily urine collection from women and cycle-timed collection of varying length from men. Bonde et al4 instructed women to collect daily first morning urine beginning on the first day of menstrual bleeding and continuing through day 10 for six cycles or until pregnancy, while their partners were asked to provide pre- and post-shift monthly spot urine samples (two total) during the same period; they reported a compliance rate of 84% for women and 59% for men. Ronnenberg et al23 instructed women to collect daily urine samples for 12 months or until achieving pregnancy; they had an 82% compliance rate–equivalent to compliance observed in this study. Wilcox et al21 asked women to collect daily first-morning urine for up to 6 months, or until becoming pregnant, when they were to collect until 8 weeks from their last menses; they reported 98% compliance and their partners were not asked to participate in urine collection. These studies instructed women to collect urine daily, without regard for cycle status and also achieved high compliance.20,21 It is possible that this is a routine more conducive to compliance since it is unnecessary to determine each day’s collection status, but it is highly resource intensive.
Several other studies collect spot urine samples, rather than daily samples, at time points related to the fertile window. The subjects of a study by Buck Louis et al14 used fertility monitors to estimate ovulation and study staff collected urine twice surrounding approximate fertility time-points; their compliance ranged from 77% to 100% for women and 94% to 100% for men. Luderer et al15 also used fertility monitors and had participating women collect a urine sample on the 10th day after the day of menses onset each month for two to three cycles; they achieved 94% compliance. These studies requesting spot urine samples demonstrate high compliance and are less resource intensive compared to studies requesting daily collection, but they may not be able to classify exposure to transient contaminants with high intrapersonal variability, such as BPA, with the same level of accuracy as studies receiving a week or more of samples.26
In this study, as in others, men’s compliance rates were markedly lower than women’s.4,14 Further, men were more likely than women to miss the sample on the first day of the fertile window and frequently continued to collect past the EDO/C+2. It is unclear whether this is due to lack of interest or a lack of intra-couple communication regarding the onset and end of the fertile window and appropriate sampling days. Overall, future studies including male partners may wish to explicitly discuss strategies to enhance compliance and communication with enrolled couples.
Our decision to restrict sample collection to the fertile window (men) and fertile window through EDO/C+18 days (women), and to train women to detect this period through observation of their own biomarkers, yielded compliance rates comparable with studies using fertility monitors to achieve the same end.14,15 This was a cost-effective decision for this study, compared to electronic monitors, which may cost hundreds of dollars for each monitor and at least $1 a day for test strips.27 Additionally, the training given in the study may deepen participants’ understanding of their own fertility in important, though less quantifiable, ways.28
Our prior validation work, and the work of others, has found that the actual date of ovulation occurs within three days of the estimated day of ovulation by the Peak Day method.22,24,25 Theoretically, for some cycles, women could miss collecting urine on the actual day of ovulation, but this should be rare, as the average duration of the estimated fertile window with this approach was 6.9 days.
This study was limited by the relatively small size and demographic homogeneity7 of the study population which may restrict generalizability. We recognize that couples planning pregnancy are substantially different than couples not planning pregnancy and that a high proportion of pregnancies in the United States are unplanned.29 However, within the population of couples willing to participate in a preconception cohort, this method can achieve high compliance with cost savings from not using fertility monitoring equipment or daily sampling not relative to the fertile window.
This study demonstrates the possibility of achieving high levels of compliance in tightly scheduled, home urine collection in the context of a preconception pregnancy cohort. It further demonstrates the utility of subject training in biomarker identification as a method of predicting the fertile window and the periconceptional time frame. Future studies may want to consider ways to improve the communication of fertile days within couples to maximize male participation.
Manuscript was prepared by Sarah Hatch Pollard, MSPH.
1. Gluckman PD, Hanson MA, Buklijas T. A conceptual framework for the developmental origins of health and disease.J Dev Orig Health Dis201016–18
2. Chapin RE, Robbins WA, Schieve LA, Sweeney AM, Tabacova SA, Tomashek KM. Off to a good start: the influence of pre- and periconceptional exposures, parental fertility
, and nutrition on children’s health.Environ Health Perspect200411269–78
3. Lenters V, Portengen L, Rignell-Hydbom A, et al. Prenatal phthalate, perfluoroalkyl acid, and organochlorine exposures and term birth weight in three birth cohorts: multi-pollutant models based on elastic net regression.Environ Health Perspect2016124365–372
4. Bonde JPE, Hjollund NHI, Jensen TK, et al. A follow-up study of environmental and biologic determinants of fertility
among 430 danish first-pregnancy planners: design and methods.Reprod Toxicol19981219–27
5. Barker DJ. Fetal origins of coronary heart disease.BMJ1995311171–174
6. Smarr MM, Grantz KL, Sundaram R, Maisog JM, Kannan K, Louis GM. Parental urinary biomarkers of preconception
exposure to bisphenol A and phthalates in relation to birth outcomes.Environ Health20151473
7. Porucznik CA, Cox KJ, Schliep KC, Wilkins DG, Stanford JB. The Home Observation of Periconceptional Exposures (HOPE) study, a prospective cohort: aims, design, recruitment and compliance.Environ Health20161567
8. WHOState of the Science of Endocrine Disrupting Chemicals–20122016Geneva, SwitzerlandWHO
9. Sly PD, Carpenter DO, Van den Berg M, et al. Health consequences of environmental exposures: causal thinking in global environmental epidemiology.Ann Glob Health2016823–9
10. Stanford JB, Brenner R, Fetterer D, Palmer L, Schoendorf KC; U.S. National Children’s StudyImpact of preconception
enrollment on birth enrollment and timing of exposure assessment in the initial vanguard cohort of the U.S. National Children’s Study.BMC Med Res Methodol20151575
11. Tingen C, Stanford JB, Dunson DB. Methodologic and statistical approaches to studying human fertility
and environmental exposure.Environ Health Perspect200411287–93
12. Buck GM, Lynch CD, Stanford JB, et al. Prospective pregnancy study designs for assessing reproductive and developmental toxicants.Environ Health Perspect200411279–86
13. Schultze A, Akmatov MK, Andrzejak M, et al. Comparison of stool collection on site versus at home in a population-based study: feasibility and participants’ preference in pretest 2 of the German National Cohort.Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz2014571264–1269
14. Buck Louis GM, Schisterman EF, Sweeney AM, et al. Designing prospective cohort studies for assessing reproductive and developmental toxicity during sensitive windows of human reproduction and development–the LIFE Study.Paediatr Perinat Epidemiol201125413–424
15. Luderer U, Christensen F, Johnson WO, et al. Associations between urinary biomarkers of polycyclic aromatic hydrocarbon exposure and reproductive function during menstrual cycles in women.Environ Int2017100110–120
16. Porucznik CA, Cox KJ, Schliep KC, Stanford JB. Pilot test and validation of the peak day method
of prospective determination of ovulation against a handheld urine hormone monitor.BMC Womens Health2014144
17. Hilgers TW. Hilgers TW. Introduction to the Creighton Model System.In: The Medical and Surgical Practice of NaProTechnology2004Omaha, NEPope Paul VI Institute Press43–56
18. Stanford JB, Smith KR, Varner MW. Impact of instruction in the Creighton model fertilitycare system on time to pregnancy in couples of proven fecundity: results of a randomised trial.Paediatr Perinat Epidemiol201428391–399
19. Ecochard R, Duterque O, Leiva R, Bouchard T, Vigil P. Self-identification of the clinical fertile window and the ovulation period.Fertil Steril20151031319.e3–1325.e3
20. Zinaman MJ, O’Connor J, Clegg ED, Selevan SG, Brown CC. Estimates of human fertility
and pregnancy loss.Fertil Steril199665503–509
21. Wilcox AJ, Weinberg CR, O’Connor JF, et al. Incidence of early loss of pregnancy.N Engl J Med1988319189–194
22. Fehring RJ, Raviele K, Schneider M. A comparison of the fertile phase as determined by the Clearplan Easy Fertility
Monitor and self-assessment of cervical mucus.Contraception2004699–14
23. Ronnenberg AG, Venners SA, Xu X, et al. Preconception
B-vitamin and homocysteine status, conception, and early pregnancy loss.Am J Epidemiol2007166304–312
24. Zinaman MJ. Using cervical mucus and other easily observed biomarkers to identify ovulation in prospective pregnancy trials.Paediatr Perinat Epidemiol200620suppl 126–29
25. Stanford JB. Revisiting the fertile window.Fertil Steril20151031152–1153
26. Cox KJ, Porucznik CA, Anderson DJ, et al. Exposure classification and temporal variability in urinary bisphenol a concentrations among couples in Utah–the HOPE Study.Environ Health Perspect2016124498–506
27. Brezina PR, Haberl E, Wallach E. At home testing: optimizing management for the infertility physician.Fertil Steril2011951867–1878
28. Vigil P, Blackwell LF, Cortés ME. The importance of fertility
awareness in the assessment of a woman’s health a review.Linacre Q201279426–450
29. Robbins CL, Zapata LB, Farr SL, et al; Centers for Disease Control and Prevention (CDC)Core state preconception
health indicators–pregnancy risk assessment monitoring system and behavioral risk factor surveillance system, 2009.MMWR Surveill Summ2014631–62