Dietary quality, as measured by the Alternative Healthy Eating Index for Pregnancy (AHEI-P), in couples planning their first pregnancy

Abstract

Objective:

Dietary quality (DQ), as assessed by the Alternative Healthy Eating Index for Pregnancy (AHEI-P), and conception and pregnancy outcomes were evaluated.

Design:

In this prospective cohort study on couples planning their first pregnancy, pregnancy was detected by pregnancy test (hCG≥20 mIU/ml). Cox proportional hazards regression was used to assess the relationship between AHEI-P score and clinical pregnancy, live birth and pregnancy loss. Models were adjusted for female age and energy intake (kcal) (Model 1), with additional adjustment for female BMI (Model 2), male education and smoking status (Model 3), and male DQ (Model 4).

Setting:

Participants were recruited from three cities in the Northeast region of the United States.

Subjects:

Participants were healthy, nulliparous couples (females (n = 132) and males (n = 131; one male did not enroll).

Results:

There were 80 clinical pregnancies of which, 69 resulted in live births, and 11 were pregnancy losses. Mean (± SD) AHEI-P was 71.0 ± 13.7 for females (mean age = 29.8 ± 3.2 y). Of those who achieved pregnancy, those in the highest tertile of AHEI-P had the greatest proportion of clinical pregnancies; however this association was not statistically significant (P=0.41). When the time it took to conceive was considered, females with the highest AHEI-P scores were 20% and 14% more likely to achieve clinical pregnancy (model 1; HR 1.20; 95% CI, 0.66–2.17) and live birth (model 1; HR 1.14; 95% CI, 0.59–2.20), respectively. Likelihood of achieving clinical pregnancy and live birth increased when the fully adjusted model, including male AHEI-P score, was examined (clinical pregnancy model 4; HR 1.55; 95% CI, 0.71–3.39 and live birth model 4; HR 1.36; 95% CI, 0.59–3.13).

Conclusion:

This study is the first to examine AHEI-P score and achievement of clinical pregnancy. DQ was not significantly related to pregnancy outcomes, even after adjustments for covariates.

INTRODUCTION

The role of dietary intake in conception is a topic of emerging interest. While eating a varied diet is a key part of overall health, certain food groups and nutrients have been shown to be beneficial in reproductive health. Recent studies have suggested that nutritional factors, including fruits, vegetables, and antioxidants, such as vitamins C and E and selenium may be beneficial for fertility; while other dietary components, such as trans fatty acids, protein, alcohol, and caffeine have been associated with detrimental effects on fertility ^(1–3). Interestingly, male diet has been shown to impact fertility as well ^(4–7). Diets higher in carbohydrate, fiber, folate, vitamin C, and lycopene ⁽⁸⁾ as well as and higher in fruits and vegetables ⁽⁹⁾ have been shown to be associated with higher sperm quality. Conversely, male intake of saturated fat ⁽¹⁰⁾, trans fat ⁽¹¹⁾ and alcohol ⁽¹²⁾ have been shown to have detrimental effects on reproductive outcomes. While a substantial body of research has focused on individual nutrients important during the preconception period, less is known about dietary patterns during this time.

Increasingly dietary patterns, rather than just individual nutrients or foods, are being assessed to evaluate the relationship between dietary intake and health outcomes. Conceptually, dietary patterns may better represent these relationships as the methodology captures the synergistic relationship that nutrients consumed in foods may have with each other ⁽¹³⁾. Dietary patterns can be evaluated using indices, or a summary score, that captures specific components that are associated with outcomes of interest. One such index is the Alternative Healthy Eating Index (AHEI).

The AHEI is an a priori dietary index that is based on the Healthy Eating Index, which was developed by the US Department of Agriculture ⁽¹⁴⁾. The AHEI is a measure of diet quality that focuses on foods and macronutrients, including assessment of unsaturated fats, associated with decreased chronic disease risk ^{(15, 16)}. To make the AHEI suitable for use in a pregnant population, Poon and colleagues modified the score to create the Alternative Healthy Eating Index for Pregnancy (AHEI-P), by excluding the alcohol component and including components for nutrients important for pregnancy (i.e., calcium, folate, and iron) ⁽¹⁷⁾. The AHEI-P has been used to determine the relationship between maternal diet in the third trimester and birth weight and early infant growth ⁽¹⁷⁾. Diet quality during the preconception period has not been previously assessed using the AHEI-P.

The Lifestyle and Fertility (ISIS) study, named after the Egyptian goddess of fertility, was designed and conducted specifically to prospectively evaluate the impact of measures of nutritional status on achieving pregnancy in couples planning their first pregnancy. The goal of this analysis was to determine the association between dietary quality as assessed by the AHEI-P and achieving pregnancy in the ISIS study. Less is known about dietary quality and male fertility, therefore a secondary goal of this study was to consider the potential role of diet quality for the male partner.

METHODS

Study participants

Study participants were recruited to be in the ISIS study, a multi-site, prospective cohort study of healthy, nulliparous couples with no known infertility conditions, who were planning their first pregnancy. Participants were recruited using a variety of methods, including referrals from medical providers in both primary care and obstetrics and gynecology practices, posting of brochures, posters and fliers in public spaces in and around surrounding areas of the sites, advertisements in local newspapers, and postcard mailings to potential participants.

Of the 802 couples who went through preliminary screening, including a baseline pregnancy test to ensure that female subjects were not pregnant at the start of the study, 181 were eligible. Couples were ineligible if there was history of recognized conception, history of infertility, or if the female had polycystic ovary syndrome. One hundred and thirty-two couples passed further screening and were enrolled between May 2008-June 2012 after attending a baseline visit at one of three clinical sites in the Northeast region (Boston, Massachusetts; State College, Pennsylvania; and Lebannon, New Hampshire). Females were between 18 and 39 years of age. All but one male partner was enrolled. Therefore, the study sample consisted of 132 females and 131 males. The protocol was approved by the participating institutions’ institutional review boards and all participants provided written informed consent. This trial was registered with Clinicaltrials.gov (ClinicalTrials.gov Identifier: NCT00642590).

Protocol

Lifestyle factors, dietary intakes, and biochemical measures were assessed among the couples. The couples were followed from the start of the study protocol until they achieved a clinically confirmed pregnancy, completed six menstrual cycles of attempted conception, or were lost to follow-up. Pregnancy was first determined by positive pregnancy test (hCG ≥ 20 mIU/ml; AIM MidStream OTC Professional, Craig Medical Distribution, Inc.) and confirmed clinically by each female participant’s physician either using a urine test, blood test and/or ultrasound. Those who conceived a clinical pregnancy were followed through the delivery or loss of the pregnancy.

Dietary Intake

Dietary intakes were assessed at baseline using a series of three (two weekday and one weekend day), unannounced 24-hour dietary recalls collected via telephone by trained interviewers of The Pennsylvania State University Diet Assessment Center (University Park, PA). Couples were individually contacted and asked to recall all foods and beverages consumed the previous day. Participants were provided a poster of 2-dimensional models to improve accuracy of portion estimation (2-D Food Portion Visual, Nutrition Consulting Enterprise, Framingham, MA). A multiple-pass technique was utilized to facilitate recall as well as standardize the process ⁽¹⁸⁾. Quality control procedures included range checks for energy (kcal) and selected vitamins, in addition to two questions are answered at the end of each food intake interview. One question asks participants if their intake was “typical”, “more than usual”, or “less than usual”. The second question is the interviewer’s assessment of the reliability of the participant’s report of their intake. For example, data might be coded as “unreliable” if they were unable to recall one or more meals or were unreliable for other reasons. In this study, we had several days of intake marked as “unreliable” for language barriers, but the intakes were in a normal range (e.g., > 500 kcal and < 5000 kcal). No participants in our sample reported implausible intakes (e.g., > 500 or < 5000 kcal). To reflect the marketplace throughout the study, dietary intake data were collected using Nutrition Data System for Research (NDSR) software versions 2008, 2009, 2010 and 2011, developed by the Nutrition Coordinating Center (NCC), University of Minnesota, Minneapolis, MN ⁽¹⁹⁾. Final calculations were completed using NDSR version 2012. The NDSR time-related database updates analytic data while maintaining nutrient profiles reflective of the marketplace at the time data were collected.

The AHEI-P is a dietary quality index based on a 130-point scale with 0–10 points awarded for optimal intake of six food groups (vegetables; whole fruit; whole grains; sugar-sweetened beverages; nuts and legumes; red/processed meats) and seven nutrient-based categories (trans fat; long-chain fats; polyunsaturated fatty acids; sodium; calcium; folate; and iron) with a higher score indicative of a higher quality diet ⁽¹⁷⁾. To calculate the score for healthy components (vegetables; whole fruit; whole grains; nuts and legumes; long-chain fats; polyunsaturated fatty acids; calcium; folate; and iron), intake of each food group or nutrient was divided by the criterion for maximum points and multiplied by 10. For less healthy components (sugar-sweetened beverages; red/processed meats; trans fat; and sodium), intake was divided by the criterion for maximum points, subtracted by 1, and then multiplied by 10. This calculation allows for higher points to be assigned for lower intake of the less healthy components. Nutrient intake was calculated from ‘food and beverages only’ (i.e., supplements were not included). Average intakes from baseline dietary recall data were used to calculate AHEI-P in both the males and females.

Covariates

Baseline demographic characteristics including age, race/ethnicity, education, household income, smoking status, and age at menarche were obtained through self-report by the participants at enrollment. Female participants were also asked if they had regular menstrual cycles at the time of the baseline visit (regular versus irregular), and their amount of weekly aerobic physical activity (<30 minutes; 30–59 minutes; 1–2 hours; ≥3 hours). Dietary covariates (e.g., energy, caffeine and alcohol) were estimated from the average of the dietary recalls. In addition, participants were weighed (within 0.1 kg) without shoes and wearing only light clothing on a calibrated, digital scale (SECA, Chino, CA). Standing height was measured without shoes (within 0.5 cm) on a wall-mounted stadiometer. All measurements were taken twice and averaged. If the first two measurements disagreed (>0.5 cm for height or >0.2 kg for weight), a third measurement was taken and the two closest were averaged and recorded. Body mass index (BMI) was calculated for each participant [weight (kg)/height (m)²].

Pregnancy Outcomes

Each confirmed clinical pregnancy was followed to determine the outcome, including date of loss or delivery of a live birth. Live birth was defined as a gasp, heart beat or sign of life at birth. Pregnancy loss included spontaneous or therapeutic abortions.

Statistical analyses

The final analytical sample included 263 participants (females = 132; males = 131). In addition, two males did not provide dietary data and therefore were excluded from any analyses that used male dietary data. AHEI-P scores were analyzed as both sex-specific tertiles and continuous variables (5-point increments). T-tests were used to examine statistical differences between females and males for AHEI-P components and total AHEI-P score (Table 1). Analysis of variance (continuous data) and chi-square and Fisher’s Exact test (categorical data) analyses were used to examine baseline maternal demographic, socioeconomic and health-related characteristics by AHEI-P-tertile (Table 2). Chi-square analysis was used to analyze the association between clinical pregnancy and AHEI-P tertile (Table 3).

Table 1.

Alternative Healthy Eating Index for Pregnancy (AHEI-P)⁸ scoring method and mean AHEI-P scores (± SD) for participants of the Lifestyle and Fertility (ISIS) Study^a

Component	Criterion for minimum score (0)	Criterion for maximum score (10)	Mean score ± SD (all)	Mean score ± SD for those females who achieved clinical pregnancy (n = 80)	Mean score ± SD for females who achieved live birth (n = 69)	Mean score ± SD for females who achieved clinical pregnancy but no live birth (n = 11)	Mean score ± SD (females)	Mean score ± SD (males)	P-valueb
Vegetablesc, servings/d	0	≥ 5	5.5 ± 2.7	5.5 ± 2.8	5.3 ± 2.8	6.4 ± 2.4	5.6 ± 2.7	5.5 ± 2.7	0.82
Whole fruit d, servings/d	0	≥ 4	3.0 ± 2.5	3.2 ± 2.4	3.2 ± 2.2	3.7 ± 3.3	3.4 ± 2.4	2.6 ± 2.5	0.02
Whole grainse, g/d	0	75 (females) 90 (males)	4.2 ± 2.7	4.5 ± 2.7	4.5 ± 2.8	5.0 ± 2.3	4.3 ± 2.6	4.0 ± 2.9	0.49
Sugar-Sweetened beverages f, servings/d	≥ 1	0	4.4 ± 4.2	5.5 ± 3.9	5.3 ± 3.9	6.4 ± 3.8	4.9 ± 4.0	3.9 ± 4.4	0.05
Nuts and legumes g, servings/d	0	≥ 1	5.7 ± 4.2	6.1 ± 4.0	6.1 ± 4.0	6.5 ± 3.9	5.8 ± 4.1	5.5 ± 4.3	0.59
Red/processed meats h, servings/d	≥ 1.5	0	5.6 ± 3.6	6.6 ± 3.1	6.6 ± 3.1	6.8 ± 2.9	6.6 ± 3.1	4.4 ± 3.7	<0.0001
trans fat i, % of energy	≥ 4	≤ 0.5	7.1 ± 1.8	7.4 ± 1.6	7.5 ± 1.6	7.1 ± 1.6	7.3 ± 1.7	6.8 ± 1.8	0.03
Long-chain (n = 3) fats (EPA + DHA)j, mg/d	0	250	3.2 ± 3.4	1.8 ± 2.8	1.9 ± 2.9	1.5 ± 1.8	2.9 ± 3.2	3.6 ± 3.6	0.06
PUFAk, % of energy	≤ 2	≥ 10	6.6 ± 1.9	6.5 ± 1.8	6.6 ± 1.8	5.8 ± 1.6	6.6 ± 1.7	6.7 ± 2.0	0.57
Sodiuml, mg/d	Highest decile 6287.6 (females) 6900.1 (males)	Lowest decile 1739.6 (females) 2110.1 (males)	5.5 ± 2.1	5.9 ± 1.8	6.0 ± 1.8	5.7 ± 1.8	5.9 ± 1.9	5.2 ± 2.1	0.004
Calciumm, mg/d	0	≥ 1200	7.0 ± 2.1	6.9 ± 2.0	6.8 ± 1.9	7.4 ± 2.3	6.8 ± 2.0	7.3 ± 2.2	0.04
Folaten, µg/d	0	≥ 600	7.3 ± 2.1	6.9 ± 2.2	6.8 ± 2.3	7.3 ± 2.0	6.9 ± 2.2	7.8 ± 2.0	0.0004
Irono, mg/d	0	≥ 27	5.8 ± 1.9	5.1 ± 1.7	5.1 ± 1.7	5.7 ± 2.1	5.2 ± 1.8	6.3 ± 1.9	<0.0001
AHEI-P	--	--	69.8 ± 14.7	72.1 ± 13.9	71.6 ± 14.1	75.1 ± 12.7	71.0 ± 13.7	68.6 ± 15.6	0.19

^a132 females; 129 males (one male partner did not enroll in the study; two other males did not provide complete dietary data); SD = standard deviation

^bP-value for comparison between males and females using t-tests.

^cIntake of at least 5 servings of vegetables per day was considered optimal. One serving is 0.5 cup of vegetables or 1 cup of leafy green vegetables.

^dIntake of ≥ 4 servings of fruit per day was considered optimal. One serving is 1 medium piece of fruit or 0.5 cup of berries.

^eIntake of 75 grams of whole grains per day (~5 servings per day) for females and 90 grams per day (~6 servings per day) for males was considered optimal. One serving of 100% whole grain product contains about 16 g of whole grains (per dry weight).

^fIntake of ≥ 1 serving of sugar-sweetened beverage per day was considered to be the least optimal. Fruit juice, including 100% fruit juice, was included in this category. One serving is 8 oz.

^gIntake of ≥1 serving of nuts and legumes per day was considered optimal. Meat alternatives such as tofu, tempeh, soy nuts, and vegetable burgers were included in this category. One serving is 1 oz. of nuts or 1 tablespoon of peanut butter.

^hIntake of < 1 serving of red or processed meat per month was considered optimal, with an upper limit of ≥ 1.5 servings per day. One serving is 4 oz. of unprocessed meat or 1.5 oz. of processed meat.

ⁱIntake of ≥ 4% of total energy intake from trans fat was considered least optimal.

^jEPA = eicosapentaenoic acid; DHA = docosahexaenoic acid; Intake of 250 mg per day of EPA and DHA (~2 4-oz. servings of fish per week) was considered optimal.

^kPUFA = polyunsaturated fatty acid; Intake of ≥ 10% of total energy intake from PUFA was considered optimal. PUFA does not include EPA or DHA intake.

^lCutoff ranges for sodium were based on the deciles of sodium distribution by sex. Values in the lowest decile (1739.6 mg per day for females and 2110.1 mg per day for males) was considered optimal.

^mIntake of ≥ 1200 mg of calcium per day was considered optimal.

ⁿIntake of ≥ 600 μg of folate per day was considered optimal.

^oIntake of ≥ 27 mg of iron per day was considered optimal.

Table 2.

Baseline characteristics of participants of the Lifestyle and Fertility (ISIS) Study by female Alternative Healthy Eating Index for Pregnancy (AHEI-P)⁸ tertile scores (females n = 132; males n = 131)

AHEI-P score tertiles (females)	T1 (42.5–65.0)	T2 (65.2–78.1)	T3 (78.6–102.0)	P-valuea	Overall
Female age, years, mean ± SD	29.2 ± 3.4	29.6 ± 3.2	30.6 ± 3.0	0.10	29.8 ± 3.2
Male age, years, mean ± SD	30.7 ± 4.7	31.1 ± 4.3	31.8 ± 4.3	0.54	31.2 ± 4.4
Female race/ethnicity, n (%)				0.40
Non-Hispanic white	34 (25.8)	30 (22.7)	36 (27.3)		100 (75.8)
Non-Hispanic black	4 (3.0)	1 (0.8)	2 (1.5)		7 (5.3)
Hispanic	1 (0.8)	3 (2.3)	1 (0.8)		5 (3.8)
Asian	5 (3.8)	8 (6.1)	3 (2.3)		16 (12.1)
Other/unknown	0 (0)	2 (1.5)	2 (1.5)		4 (3.0)
Male race/ethnicity, n (%)b				0.39
Non-Hispanic white	37 (28.5)	30 (23.1)	38 (29.2)		105 (80.8)
Non-Hispanic black	2 (1.5)	2 (1.5)	1 (0.8)		5 (3.9)
Hispanic	0 (0)	0 (0)	1 (0.8)		1 (0.8)
Asian	3 (2.3)	8 (6.2)	3 (2.3)		14 (10.8)
Other/unknown	1 (0.8)	3 (2.3)	1 (0.8)		5 (3.9)
Female education, n (%)				0.001
Less than college degree	11 (8.3)	3 (2.3)	1 (0.8)		15 (11.4)
College degree	20 (15.2)	18 (13.6)	14 (10.6)		52 (39.4)
Graduate degree	13 (9.9)	23 (17.4)	29 (22.0)		65 (49.2)
Male education, n (%)				0.001
Less than college degree	17 (13.0)	5 (3.8)	3 (2.3)		25 (19.1)
College degree	13 (9.9)	20 (15.3)	18 (13.7)		51 (38.9)
Graduate degree	13 (9.2)	19 (14.5)	23 (17.6)		55 (42.0)
Annual household income, n (%)c				0.82
< $60,000	14 (10.9)	12 (9.3)	9 (7.0)		35 (27.1)
$60,000 to $99,999	13 (10.1)	13 (10.1)	15 (11.6)		41 (31.8)
≥$100,000	16 (12.4)	19 (14.7)	18 (14.0)		53 (41.1)
Female smoking status, n (%)				0.67
Never	33 (25.0)	36 (27.3)	35 (26.5)		104 (78.8)
Past use	10 (7.6)	8 (6.1)	9 (6.8)		27 (20.5)
Current use	1 (0.8)	0 (0)	0 (0)		1 (0.8)
Male smoking status, n (%)				0.01
Never	24 (18.5)	27 (20.8)	20 (15.4)		71 (54.6)
Past use	11 (8.5)	16 (12.3)	23 (17.7)		50 (38.5)
Current use	7 (5.4)	1 (0.8)	1 (0.8)		9 (6.9)
Female alcohol intake, g/d, mean ± SD	4.6 ± 9.6	7.3 ± 11.2	5.6 ± 9.3	0.46	5.9 ± 10.1
Male alcohol intake, g/d, mean ± SD	13. 1 ± 20.3	9.8 ± 15.4	4.8 ± 8.9	0.05	9.2 ± 15.8
Female caffeine intake, mg/d, mean ± SD	61.5 ± 66.4	83.9 ± 71.9	98.6 ± 91.1	0.08	81.3 ± 78.1
Female aerobic activity				0.15
< 30 min. per week	11 (8.3)	6 (2.3)	3 (2.3)		20 (15.2)
30–59 min. per week	12 (9.1)	10 (7.6)	9 (6.8)		31 (23.5)
1–2 hours per week	8 (6.1)	8 (6.1)	13 (9.9)		31 (23.5)
≥ 3 hours per week	11 (8.3)	20 (15.2)	19 (14.4)		50 (37.9)
Regular menstrual cyclesd				1.00
No	3 (2.3)	3 (2.3)	3 (2.3)		9 (7.0)
Yes	39 (30.2)	40 (31.0)	41 (31.8)		120 (93.0)
Age at menarche, years, mean ± SD	12.8 ± 1.4	13.0 ± 1.4	12.7 ± 1.5	0.71	12.8 ± 1.4
Female pre-pregnancy BMI, kg/m2, mean ± SD	26.8 ± 6.5	23.0 ± 4.3	23.1 ± 4.1	<0.001	24.3 ± 5.4
Male BMI, kg/m2, mean ± SD	29.5 ± 5.0	26.1 ± 4.1	27.5 ± 7.3	0.02	27.7 ± 5.8
Female total energy, kcal/d, mean ± SD	1643.2 ± 490.3	1694.8 ± 431.3	1841.0 ± 497.5	0.13	1726.3 ± 477.8
Male total energy, kcal/d, mean ± SDe	2016.5 ± 659.8	2202.6 ± 607.4	2281.4 ± 578.6	0.13	2168.3 ± 620.9
Female AHEI-P score, mean ± SD (range)	55.5 ± 6.2 (42.5–65.0)	71.0 ± 4.1 (65.2–78.1)	86.4 ± 5.6 (78.6–102.0)	<0.0001	72.0 ± 13.8
Male AHEI-P score, mean ± SD (range)e	51.5 ± 6.5 (35.3–60.8)	67.6 ± 4.1 (61.2–75.6)	86.6 ± 7.8 (76.8–111.1)	<0.0001	68.7 ± 15.6

^aP-value for comparison across AHEI-P tertiles using analysis of variance (continuous variables) and chi-square and Fisher’s Exact test (categorical variables).

^bOne participant was missing ethnicity.

^cThree couples who reported unknown incomes were excluded from analysis.

^dThree females were missing information on regularity of menstrual cycles at baseline.

^eTwo male participants did not have complete dietary data and therefore, n = 129 for total energy and AHEI-P component scores and total score.

Table 3.

Clinical pregnancy status by Alternative Healthy Eating Index for Pregnancy (AHEI-P)⁸ score tertile for females of the Lifestyle and Fertility (ISIS) Study (n = 132)^a

		AHEI-P tertile (range)
Achieved clinical pregnancy, % (n)	Low (42.5–65.0)	Medium (65.2–78.1)	High (78.6–102.0)
No	40.9 (18)	45.5 (20)	31.8 (14)
Yes	59.1 (26)	54.6 (24)	68.2 (30)

^aP-value = 0.41 determined by chi-square analysis

To account for those couples who took longer to conceive, time to conception was considered using Cox proportional hazards regression models (PROC PHREG in the SAS). This method captures measurement of an event from a defined starting point to the outcome(s) of interest (e.g., clinical pregnancy). To complete this analysis, a ‘time to event’ variable (i.e., time to conception), which takes into account varying start and ‘censoring’ times, was constructed for each participant. The Cox model has advantages, namely, it accounts for the presumed reduction in the probability of conception with each subsequent menstrual cycle that goes by, by building it into the baseline hazard function. Ideally, time in the Cox model would be measured as the number of menstrual cycles since baseline. While cycle length varies between females and even varies within-person, an average cycle length, often 30 days, is used to represent the sample; however, in the absence of this data, time measured as ‘days’ was used in these analyses.

Time to conception of a clinical pregnancy was defined as the length of time in ‘days’ from the date the couple began attempting conception (day one of the first menstrual cycle after the baseline visit) to the date of the last menstrual period plus 14 days. Cox proportional hazards regression models were also used to examine ‘time to conception of a clinical pregnancy resulting in a live birth’ and ‘time to conception of a clinical pregnancy resulting in a pregnancy loss’ as events (Table 4). Those who conceived prior to the start of their first cycle were censored at the time of their baseline visit (n = 12). For those who did not conceive a clinical pregnancy and had follow-up to determine outcome, time was censored at the end of the six menstrual cycles. Those without follow-up (i.e., staff attempted to contact them and could not reach them) were censored at the date of last contact (n =5, of which 3 were at baseline visit). Those who were censored at their baseline visit (n = 15) were not included in the time to conception analyses.

Table 4.

Cox proportional hazards regression analysis for time to pregnancy as a function of Alternative Healthy Eating Index for Pregnancy (AHEI-P)⁸ score (tertile) and selected covariates for ISIS participants (n = 117 couples). Hazard Ratio (HR) with 95% Confidence Intervals (95% CI) for time to pregnancy outcomes^a

	HR (95% CI)	P-value
Clinical pregnancy (n = 80)
Model 1b
AHEI-P medium	1.23 (0.67–2.27)	0.50
AHEI-P high	1.20 (0.66–2.17)	0.56
Model 2c
AHEI-P medium	1.34 (0.70–2.53)	0.38
AHEI-P high	1.25 (0.66–2.35)	0.49
Model 3d
AHEI-P medium	1.50 (0.72–3.14)	0.28
AHEI-P high	1.43 (0.68–2.95)	0.35
Model 4e
AHEI-P medium	1.64 (0.78–3.47)	0.19
AHEI-P high	1.55 (0.71–3.39)	0.27
Clinical pregnancy that resulted in live birth (n = 69)
Model 1b
AHEI-P medium	1.35 (0.70–2.59)	0.37
AHEI-P high	1.14 (0.59–2.20)	0.70
Model 2c
AHEI-P medium	1.46 (0.74–2.90)	0.28
AHEI-P high	1.19 (0.59–2.38)	0.63
Model 3d
AHEI-P medium	1.48 (0.68–3.21)	0.32
AHEI-P high	1.23 (0.56–2.69)	0.60
Model 4e
AHEI-P medium	1.63 (0.74–3.59)	0.22
AHEI-P high	1.36 (0.59–3.13)	0.46
Pregnancy loss (n = 11)
Model 1b
AHEI-P medium	1.02 (0.16–6.56)	0.99
AHEI-P high	0.65 (0.14–3.12)	0.59
Model 2c
AHEI-P medium	0.99 (0.15–6.50)	0.99
AHEI-P high	0.55 (0.10–3.02)	0.49
Model 3d
AHEI-P medium	2.72 (0.26–28.92)	0.41
AHEI-P high	1.25 (0.13–12.32)	0.85
Model 4e
AHEI-P medium	2.31 (0.21–25.89)	0.50
AHEI-P high	0.93 (0.09–10.22)	0.95

^aThe sample size does not include the n = 15 who were censored at baseline visit. All covariates are for the female unless otherwise indicated. The reference group is the lowest tertile of AHEI-P. P-values determined by Cox proportional hazards regression.

^bAdjusted for age and energy intake

^cAdjusted for age, energy intake, and BMI

^dAdjusted for age, energy intake, BMI, male education, and male smoking status

^eAdjusted for age, energy intake, BMI, male education, male smoking status, and male AHEI-P score

Models were statistically adjusted for age and energy (kcal) (Model 1) with additional adjustment for female BMI (Model 2). Model 3 included all adjustments from Model 2, in addition to male education and smoking status. Partner’s education, in this case the male partner, is functioning as a marker for both health behaviors and economic status. Partner’s education has been significantly associated with self-assessed health ⁽²⁰⁾. Lastly, male dietary quality, as measured by AHEI-P score (Model 4), was included as a covariate. In a series of sensitivity analyses, we evaluated whether either caffeine intake or alcohol consumption were important contributors in multivariate models. Finally, we examined male AHEI-P score as the exposure of interest (Supplementary table). All models are reported with their respective estimates and 95% confidence intervals.

Assumptions of the Cox proportional hazards regression model were tested (e.g., examination of survival plots) and met. Statistical significance was set at P<0.05. All data analyses were performed using SAS 9.4 software (SAS Institute Inc., Cary, NC, USA) ⁽²¹⁾.

RESULTS

Baseline mean (± SD) AHEI-P scores are presented in Table 1. Males and females were statistically significantly different for mean AHEI-P component scores of whole fruit; red/processed meat; trans fat; sodium; calcium; folate; and iron (P<0.05); however, total AHEI-P score was not significantly different between males and females (P=0.19). AHEI-P mean (± SD) tertile scores for females were 55.5 ± 6.2; 71.0 ± 4.1; and 86.4 ± 5.6 for T1 (42.5–65.0), T2 (65.2–78.1), and T3 (78.6–102.0), respectively. When examined by AHEI-P tertile score, females were not statistically significantly different for baseline characteristics across tertiles, with the exception of education and pre-pregnancy (baseline) BMI. Females with lower AHEI-P scores tended to have higher pre-pregnancy BMI. When examined by female AHEI-P tertile score, males were significantly different for education, smoking status, BMI and AHEI-P score (Table 2). Male partners of females with lower AHEI-P scores tended to have lower educational status and higher BMIs.

There were 80 clinical pregnancies of which, 69 resulted in live births, and 11 were pregnancy losses reported in this study. Of those who achieved clinical pregnancy, those in the highest tertile of AHEI-P had the greatest proportion of clinical pregnancies; however this association was not found to be significant (P=0.41) (Table 3).

When time to conception was considered as an outcome, using Cox proportional hazards regression modeling on a sample of 117 couples, females with the highest AHEI-P scores were 20% and 14% more likely to achieve clinical pregnancy (model 1; HR 1.20; 95% CI, 0.66–2.17) and live birth (model 1; HR 1.14; 95% CI, 0.59–2.20), respectively. Likelihood of achieving clinical pregnancy and live birth increased when the fully adjusted model, including male AHEI-P score, was examined (clinical pregnancy model 4; HR 1.55; 95% CI, 0.71–3.39 and live birth model 4; HR 1.36; 95% CI, 0.59–3.13) (Table 4). Additionally, the association between male AHEI-P score and achieving clinical pregnancy; live birth; and pregnancy loss were examined. Results from the sensitivity analysis examining male AHEI-P as the exposure variable and the above pregnancy outcomes followed the same pattern as the results using female AHEI-P score (i.e., also non-significant) (Supplementary table).

DISCUSSION

This study is the first to examine dietary quality, using the AHEI-P score and achievement of clinical pregnancy. In this sample of couples planning their first pregnancy, dietary quality, as measured by AHEI-P, females with greater adherence to the AHEI-P dietary pattern were more likely and more quickly able to achieve clinical pregnancy. Females who more closely followed the AHEI-P dietary pattern were also less likely to experience pregnancy loss. It would be prudent of us to acknowledge the wide and non-significant confidence intervals for these results (Table 4), but given the limited power, these analyses should be repeated with a larger sample size.

While there is a breadth of studies on nutritional factors and fertility, this discussion focuses on examination of the literature on dietary patterns (i.e., a priori dietary quality scores or a posteriori dietary patterns), which may better reflect overall dietary intake more comprehensively than examination of single nutrients ⁽¹³⁾. While sparse, these studies on likelihood of pregnancy and pre-conception dietary quality are most relevant to the current study. A study by Gaskins and colleagues examined the association between pre-pregnancy diet, using dietary indices, and pregnancy loss in women from the Nurses’ Health Study II ⁽²²⁾. Dietary patterns, examined using a priori methods, were not significantly associated with pregnancy loss. It is worth noting that the present study included only nulliparous women, whereas the study by Gaskins et al did not have this same criteria. Conversely, two other studies did find statistically significant relationships between pre-conception dietary patterns and likelihood of achieving pregnancy; however, results may not be directly comparable as couples in these studies used couples undergoing assisted reproductive technology ^{(23, 24)}.

While this study is unique in that it is the first to examine AHEI-P score and achieving clinical pregnancy, studies have utilized other methods to characterize dietary patterns in relation to pregnancy outcomes. Chavarro et al described the “fertility diet” pattern, which was characterized by a high intakes of monounsaturated to trans fat, high-fat dairy, iron, higher frequency of multivitamin use and lower animal protein, glycemic load, and low-fat dairy ⁽²⁵⁾. Females with greater adherence to the “fertility diet” was associated with a 69% lower risk of ovulatory disorder infertility and this relationship was not modified by differences in age, parity, or BMI. Although a direct comparison between the AHEI-P and the “fertility diet” score is difficult, results from this study are similar to those of Charvarro and colleagues, in that those who achieved a clinical pregnancy had higher diet quality (i.e., higher mean AHEI-P scores (±SD) (72.1 ± 13.9; n =80)) compared to those who did not achieve clinical pregnancy (69.3 ± 13.4; n = 52) (P <0.001; data not shown). Similarities between the AHEI-P and the “fertility diet” score can be seen in that both award higher points to diets that are rich in vegetables, healthy fats, and iron, and low in animal protein/red/processed meats.

In regards to pregnancy loss, Maconochie et al examined associations between biological, behavioral, and lifestyle risk factors, including nutrition, and found that fresh fruit and vegetable intake on most days decreased the odds of having a miscarriage by half ⁽²⁶⁾. Additionally, consumption of dairy products, chocolate, and eating fish or white meat twice weekly was also suggestive of protective effects against miscarriage. Interestingly, both increasing frequency and average weekly intake of alcohol was found to be associated with statistically significant increased risk of miscarriage ⁽²⁶⁾. Similarly, in other study conducted in Italy, that examined pregnancy loss, alcohol consumption before pregnancy was also associated with increased risk of spontaneous abortion. In this same study, consumption of green vegetables, fruit, milk, cheese, fish and eggs, were also found to be protective ⁽²⁷⁾. It is noteworthy that the AHEI-P score used in the current study does not include alcohol or caffeine as components of the score. We evaluated these variables in our analyses; including them in our multivariate models did not appreciably change our results.

The current study is distinct in that it is the first to examine AHEI-P score and pregnancy outcomes using pre-pregnancy diet in nulliparous couples. Three other studies have utilized the AHEI-P to investigate dietary quality related to pregnancy outcomes; however the outcomes have been related to infant birth weight and did not examine the likelihood of achieving pregnancy and therefore are not addressed in this paper ^{(17, 28, 29)}.

Results of this study should be interpreted in the context of its strengths and limitations. This is the first study to examine dietary quality by AHEI-P in a sample of nulliparous couples planning pregnancy. This study is also unique in that it accounts for male dietary quality in terms of pregnancy outcomes. It is worth noting that the sample size was modest, particularly for the analysis that examined those who experienced pregnancy loss (n= 11) and may have been too small to detect significance. Therefore, interpretation of these findings must be made with caution. However, this study is exploratory in its use of the AHEI-P with male dietary data. There were statistically significant differences in some components of the AHEI-P between males and females; however, there were no differences in total AHEI-P between males and females. Thus, the AHEI-P may not capture all dietary components that may be relevant to the likelihood of pregnancy. The AHEI-P was developed for females and it is unknown how relevant the AHEI-P is for men. The AHEI-P may not adequately capture components of the diet that may affect male fertility (i.e., sperm count and quality), like fish intake ⁽³⁰⁾ or caffeine ⁽³¹⁾. Nevertheless, it is interesting that the relationship between maternal dietary quality and likelihood of achieving pregnancy increased once paternal dietary quality was included in the model (Table 4: Models 3 and 4). In Table 4, Model 4 adjusts for all the variables from Model 3, plus the addition of male AHEI-P score. The hazard ratio increased from 1.50 to 1.64 (AHEI-P medium) and from 1.43 to 1.55 (AHEI-P high) when comparing Model 3 to Model 4. While these results are not statistically significant, the magnitude of the hazard ratio increases. While less is known about the role that paternal nutrition plays in reproduction, results suggest that higher male dietary quality may be related to increased chances of achieving pregnancy. Dietary intakes of couples can be correlated; however in this sample, correlations between the components of the AHEI-P were moderate. For example, the Spearman correlation for total AHEI-P score between males and females was 0.52 (P<0.001).

CONCLUSIONS

In conclusion, results suggest that higher pre-pregnancy dietary quality, as measured by AHEI-P, was associated with increased likelihood of achieving pregnancy, including pregnancy that resulted in live birth. Adherence to the AHEI-P dietary pattern was also associated with a decreased risk of pregnancy loss. The role of pre-pregnancy diet and fertility is an area that deserves continued research. Future studies should address these research questions with adequate power. While the link between nutrition before pregnancy and pregnancy outcomes has been studied, it is possible that there are dietary components that are important for increasing the likelihood of getting pregnant that are not captured in the AHEI-P. Additionally, the AHEI-P has not been assessed for validity and reliability. Future studies should consider the development and evaluation of a dietary quality score that more accurately measures the likelihood of achieving pregnancy. Determining components of a dietary pattern that supports achieving conception is a high priority for future research. Because diet is a modifiable risk factor, examining preconception diet to identify nutrients important in achieving conception may be a potential approach to use with females planning pregnancy.