Fruit and vegetable consumption, pesticide residue intake from consumption of fruits and vegetables, and risk of uterine fibroids

Abstract

Objective:

To examine the association between the consumption of fruits and vegetables and pesticide residue intake from consumption of fruits and vegetables and risk of ultrasound or hysterectomy-confirmed fibroids. Only a few studies have evaluated the association of fruit and vegetable intake with uterine fibroids with inconsistent results. No studies have examined pesticide exposure through fruits and vegetables with fibroid risk.

Design:

Prospective cohort study. Cox proportional hazards models were used to calculate hazard ratios (HR) and 95% confidence intervals (CI).

Setting:

Not applicable.

Patient(s):

81,782 premenopausal participants from the Nurses’ Health Study II cohort were followed from 1991-2009 for the fruits and vegetable analysis, and 49,927 participants were followed from 1999-2009 for the pesticide residue burden analysis. Diet was assessed every four years with a food frequency questionnaire. Fruits and vegetables were classified into high- or low-pesticide-residues using a validated method based on surveillance data from the U.S. Department of Agriculture.

Intervention(s):

Not applicable

Main outcome measure(s):

Cases of ultrasound or hysterectomy-confirmed fibroids were identified from self-reports to validated questionnaires.

Results:

From 1991-2009, 9,706 incident cases of ultrasound or hysterectomy-confirmed fibroids were reported, and 4,195 incident cases were identified from 1999-2009. No association was observed between total fruit and vegetable consumption and uterine fibroid risk. Participants who consumed the highest intake of total fruits (≥4/day) were 10% less likely to develop uterine fibroids compared to participants who consumed <1/day (95% CI=0.80-1.01; p_trend=0.03). No associations were observed with any other fruit or vegetable groups. An inverse association was observed between intake of high-pesticide-residue fruits and vegetables and fibroid risk (HR for 5^th vs 1^st quintile=0.87; 95% CI=0.77-0.99; p_trend=0.04) while no association with low-pesticide-residue fruits and vegetables was observed (HR for 5^th vs 1^st quintile=1.08; 95% CI=0.95-1.23; p_trend=0.26).

Conclusion:

Our findings suggest that pesticide residues on fruits and vegetables are not associated with a higher risk of uterine fibroids. Further, our results suggest that intake of fruits may be associated with a lower risk of fibroids. Future research in this area should focus on dietary exposures across the life course as well as assessment of class-specific pesticides.

Introduction

Uterine fibroids (leiomyomas) are hormone-dependent, noncancerous growths of the uterus and are the most common gynecologic condition in the U.S. among reproductive-aged women (1, 2). Fibroids are the leading cause of hysterectomies in the U.S (1) with symptoms that include abnormal bleeding, pelvic pain, infertility, and adverse pregnancy outcomes (1-3). Annually, the direct cost of fibroids which includes surgeries, hospital admissions, outpatient visits, and prescribed medications range from 4.1 to 9.4 billion dollars and the costs of nonsurgical management per patient range from $5,563 to $8,665 (4). Despite the high prevalence and substantial morbidity, the etiology of fibroids is not fully understood.

Current dietary guidelines indicate the importance of consuming fruits and vegetables as part of a healthy diet to reduce risk of chronic diseases (5). Fruits and vegetables contain antioxidants and phytochemicals that may lower risk of fibroids via apoptosis or hormone-dependent pathways (6, 7). Conversely, contaminants (e.g., pesticides) in food products could increase fibroid risk (8, 9). Environmental changes can manifest as gross structural changes within the female reproductive tract (10-12). Pesticide exposure can result in changes of hormonal signaling, which includes gynecological disorders such a fibroids (13). Currently, few studies have examined the relation between fruits and vegetables and fibroid risk with most prior studies being limited by a case-control design (14), which may be impacted by differential dietary recall between fibroid cases and controls (15-18). To our knowledge, only one prospective study has examined the associations between fruit and vegetable intake and fibroid risk (17). The prospective Black Women’s Health Study (BWHS) observed that fruit intake was inversely associated with uterine fibroids; specifically, citrus fruit intake while no association was observed with vegetable intake (17).

One of the explanations for inconsistency across prior studies could be related to pesticide residues on fruits and vegetables which would offset potential health benefits. Greater than 90% of the U.S. population has measurable concentrations of pesticides or their metabolites in their urine or blood (19, 20) and for the general population conventionally produced fruits and vegetables are a main source of pesticide exposure (21, 22). However, whether pesticide residue exposure through fruit and vegetable intake impacts fibroid risk has not been examined. Therefore, the purpose of this study was to examine the association between the consumption of fruits and vegetables and pesticide residue intake from consumption of fruits and vegetables and risk of ultrasound or hysterectomy-confirmed fibroids.

Methods

Study Population

The Nurses’ Health Study II (NHSII) is an ongoing prospective cohort established in 1989. At baseline, 116,429 female registered nurses aged 25–42 years completed a questionnaire that collected information on demographic and lifestyle factors, anthropometric variables, and disease history. Follow-up questionnaires are sent biennially to update information on exposures and disease status and food frequency questionnaires (FFQ) have been completed every four years. Additional study details have been provided elsewhere (23). Response rates have been at least 90% throughout follow-up cycles. Implied consent was assumed upon return of the completed questionnaire. This study protocol was approved by the Institutional Review Boards of the Brigham and Women’s Hospital and Harvard T.H. Chan School of Public Health.

Analytic Populations

Fruits and Vegetables.

Follow-up for the analyses on fruits and vegetables began in 1991 when 95,233 participants completed the 1991 FFQ and concluded in 2009, the last year fibroids were assessed on the biennial questionnaire. We excluded participants for the following reasons: more than 70 blank FFQ responses in 1991 (n=15), fibroid diagnosis prior to 1991 (n=5,715), fibroids not ultrasound or hysterectomy confirmed (n=145), unknown date of fibroid diagnosis (n=759), loss to follow-up (n=643), postmenopausal (n= 477), hysterectomy prior to 1991 (n=4,806), and prior cancer diagnosis (n=891). After these exclusions, 81,782 premenopausal participants remained in the analytic sample.

Pesticide residue burden.

For the pesticide residue burden analyses we used the 1999 FFQ as baseline (n=84,938) because the USDA Pesticide Data Program (described below) expanded its surveillance data starting in 1996. Participants were followed from 1999 until 2009, which was the last year that uterine fibroids were assessed on the biennial questionnaires. We excluded participants for the following reasons: more than 70 blank FFQ responses in 1999 (n=288), deceased prior to 1999 (n=4,978), fibroid diagnosis prior to 1999 (n=10,479), fibroids not ultrasound or hysterectomy confirmed (n=2,029), unknown date of fibroid diagnosis (n=615), loss to follow-up (n=232), postmenopausal (n=6,976), hysterectomy prior to 1999 (n=7,391), and prior cancer diagnosis (n=2,023). After these exclusions, 49,927 premenopausal participants with PRBS dietary information remained in the analytic sample.

Uterine Fibroids Assessment

Starting in 1993, participants were asked on each biennial questionnaire if they ever received a physician diagnosis of uterine fibroids and, if so, the date of diagnosis and whether the diagnosis was confirmed by pelvic exam, ultrasound or hysterectomy. Cases were defined based on self-reported ultrasound or hysterectomy confirmation. Participants who reported fibroids not confirmed by ultrasound or hysterectomy (pelvic exam only) did not contribute person time to the study period but were allowed to re-enter the analysis if confirmed by ultrasound or hysterectomy in the future. In 2009, uterine fibroid incidence was assessed for the last time on the biennial questionnaire; during this time the youngest participant was aged 45 years.

In a previous validation study, a subset of newly diagnosed cases confirmed by ultrasound or hysterectomy (n=243; 100 White and 143 African-American) were mailed a questionnaire on symptoms and a review of their medical records was requested (24). Of the 216 who responded (89%), 6% denied the diagnosis and 34% confirmed the diagnosis but did not release their medical records. Among the cases in which medical records could be obtained, 93% were confirmed (24). The proportion diagnosed by hysterectomy, myomectomy, or ultrasound did not differ between those who did and did not give permission for medical record release. The proportion confirmed by medical records did not differ comparing White (94%) and African-American (92%) participants (24).

Dietary Assessment

A validated FFQ was used to assess diet every four years (1991,1995, 1999, 2003 and 2007). Specifically, participants were asked to select one of nine responses (“never or less than once per month” to “6 or more times a day”) for more than 130 serving-size specific food items about their average consumption over the previous year. The reproducibility and validity of the FFQ has been previously evaluated, indicating moderately strong correlations on average for its accuracy of individual food item consumption and overall dietary patterns(25, 26). The FFQ has been shown to provide valid estimates of fruit, vegetable, and nutrient intake with deattenuated correlation coefficients for fruits and vegetables between the FFQ and 1-week diet records ranging from 0.16 for yellow squash to 0.80 for apples (27). The coefficients for most fruits and vegetables were above 0.40. The FFQ data were combined with the below described pesticide residue burden score method to estimate dietary pesticide intake from fruit and vegetable consumption. Fruit and vegetable items were reported individually and fruit and vegetable items were summed to create total fruit (not including juices), total vegetable (not including potatoes), and total fruit juice intake. Fruit and vegetable subgroups were also examined including: citrus fruit, rosacea fruit, greens vegetables, orange vegetables, cruciferous vegetables, tomatoes, and legumes.

Pesticide Residue Burden Scores

The USDA Pesticide Data Program (PDP) monitors the presence of approximately 450 pesticides in the U.S. food supply by conducting annual samples, tests and reports (28, 29). The pesticide residue sampling data are collected from the agricultural commodities of 10 participating states, reflect the preparation and consumption habits of the general population and focus on measuring pesticides registered for agricultural use in the U.S. or that may be present on food commodities that arrive to the U.S. from other countries (29). The PDP pesticide monitoring data were used in the development of the validated pesticide residue burden score (PRBS) for use in epidemiology studies aimed at examining dietary pesticide residue intake from fruits and vegetables (30, 31). The PRBS method averages annual pesticide residue monitoring data across 4-year periods to correlate with the years of the FFQ data in this study (1999, 2003 and 2007); for example, PDP pesticide residue monitoring data were averaged from 1996-1999 for the average PRBS scores used in the 1999 FFQ (32). These averaged PDP data were used to tertile rank fruits and vegetables included in the FFQ based on three contamination measures: % of samples with any detectable pesticides; % of samples with pesticides > tolerance level; and % of samples with ≥ 3 types of pesticides. The tertile ranks were assigned a score value (lowest tertile=0; middle tertile=1; upper tertile=2) and then summed across all three measures for each fruit and vegetable to determine the final PRBS score. Fruits and vegetables were categorized into two pesticides status groups based on their final PRBS score, high pesticide residue (PRBS=4-6) or low pesticide residue (PRBS=0-3). A separate category, called undetermined, was created for fruits and vegetables items with no available PDP data in a specific surveillance period.

The PRBS score has been previously validated, with correlation between PRBS assessment and presence of urinary biomarkers of pesticide exposure in two separate cohorts. In the first validation study among 1,918 adult participants in the National Health and Nutrition Examination Survey (NHANES), urinary concentrations of non-specific organophosphate biomarkers were higher for participants in the highest quintile of high-pesticide residue fruit and vegetable intake than participants in the lowest quintile (33). In addition, urinary concentrations of non-specific organophosphate biomarkers were not significantly higher for participants in the highest quintile of low-pesticide residue fruit and vegetable intake than participants in the lowest quintile. In a second validation study of the PRBS among 90 men within the Environment and Reproductive Health Study, two urine samples per man were analyzed for seven commonly used pesticide biomarkers, including organophosphate and pyrethroids insecticides and 2,4 dichlorophenoxyacetic acid (30). Urinary concentrations of pesticide biomarkers were 21% higher (95% CI=2-44%) for each additional serving/day of high pesticide residue fruits and vegetables, and 10% lower (95% CI=1-18%) for each one serving/day increase in low pesticide residue fruits and vegetables, demonstrating the biological relevance of this questionnaire-based assessment (30).

Statistical Analysis

In the primary analysis, we examined the effect of fruits and vegetable intake (total fruits and vegetables and individual groups) and fibroid risk. Participants contributed person-time from 1991 until self-report of ultrasound or hysterectomy-confirmed uterine fibroid diagnosis, diagnosis of any cancer (except non-melanoma skin cancer), death, loss to follow-up, hysterectomy, menopause, or return of the 2009 questionnaire (the last year fibroids were assessed on the biennial questionnaire), whichever occurred first. As pesticide surveillance data was less well-characterized in the 1990s (described above), participant follow-up for this analysis began in 1999 and ended in 2009. Participants were classified according to quintiles of high pesticide residue fruit and vegetable intake, and low pesticide residue fruit and vegetable intake. In all analyses, Cox proportional hazards regression models with age and questionnaire period as the time scale were utilized to estimate hazard ratios (HR), and 95% confidence intervals (CI) using the lowest category as the reference. Total caloric intake was included in both age- and covariate-adjusted models. Covariate-adjusted models included potential confounders chosen for their a priori association to uterine fibroids or dietary factors: total caloric intake, age at menarche, parity, length of a menstrual cycle, oral contraceptive use, body mass index (BMI), and cigarette smoking. To adjust for health care utilization (which could influence receiving a fibroid diagnosis) we adjusted for recent gynecologic/breast exam. Tests for linear trend of the exposures of interest were performed by assigning the median value of each category to all participants in that group. In sensitivity analysis, we examined the association between fruit and vegetable intake during the same time period that the PRBS scores were examined (1999-2009). All statistical analyses were performed using SAS, version 9.4 (SAS Institute Inc.).

Results

Participant characteristics of total fruits and vegetables and low and high pesticide residue fruits and vegetables intakes are presented in Table 1. Participants with the greatest intakes of fruits and vegetables were less likely to be current smokers and were more likely to have never used oral contraceptives than those with the lowest intakes. Those in the highest quintile of both high and low PRBS were less likely to have a BMI>30 and to be current smokers than those in the lowest quintile. Intakes of high pesticide residue and low pesticide residue fruits/vegetables were positively correlated with each other (r_Spearman=0.67).

Table 1.

Distribution of participant characteristics by intakes of total fruits and vegetables, low pesticide residue fruits and vegetables, and high pesticide residue fruits and vegetables, Nurses’ Health Study

	Total fruit and vegetable intake		High pesticide residue fruit and vegetable intake		Low pesticide residue fruit and vegetable intake
	Quintile 1 (n=15, 331)	Quintile 5 (n=19,800)	Quintile 1 (n=10,035)	Quintile 5 (n=10,024)	Quintile 1 (n=10,597)	Quintile 5 (n=19,800)
Age, yearsa	35.09 (4.7)	36.17 (4.5)	42.53 (4.34)	43.14 (4.3)	42.65 (4.3)	43.09(4.3)
Total calories	1453 (476)	2132 (538)	1604 (444)	2047 (467)	1552 (416)	2118 (462)
Nulliparous, %	30	28	20	18	21	17
BMI, kg/m2
<19, %	6	5	3	3	3	3
19-20.4, %	14	12	7	9	8	9
20.5-21.9, %	18	18	13	16	13	16
22-24.9, %	26	26	24	27	25	26
25-29.9, %	17	19	24	24	24	23
30+, %	14	12	24	16	22	17
Cycle length
<26 days, %	11	11	11	10	11	11
26-31 days, %	66	66	65	65	65	65
32-50 days, %	17	17	18	19	18	18
≥ 51 or irregular, %	5	5	6	6	6	6
Age at menarche
< 12, %	22	26	21	24	22	24
12, %	30	30	30	30	30	30
13, %	28	26	28	27	27	27
≥ 14, %	19	17	20	19	20	20
Recent gynecological exam, %	80	83	85	88	85	87
OC use
Never, %	15	18	13	17	14	16
Past, %	72	70	75	72	74	73
Current, %	13	11	12	12	12	11
Smoking status
Never, %	65	66	67	68	68	69
Past, %	18	25	21	27	22	25
Current, %	16	10	12	5	10	6

Note: Values are means (SD) for continuous variables and percentages for categorical variables and are standardized to the age distribution of the study population. Columns may not add to 100% due to missing responses. Values of polytomous variables may not sum to 100% due to rounding.

^aValue is not age adjusted

During 18 years of follow-up (1991-2009), a total of 81,782 premenopausal participants contributed to 953,086 person-years and a total of 9,706 incident cases of ultrasound or hysterectomy confirmed fibroids were reported. No association was observed between total fruit and vegetable intake and fibroid risk with a HR of 0.99 (95% CI=0.92-1.06; p_trend=0.57) for an intake of ≥6 servings/day compared to ≤2 servings/day. An inverse association was observed between total fruit intake and risk of uterine fibroids. Those consuming ≥4 servings/day of total fruits had a 10% lower risk of uterine fibroids compared to participants who consuming <1 serving/day (95% CI= 0.80-1.01; p_trend=0.03). No associations were observed between any of the individual fruit and vegetable categories and fibroid risk (Table 2).

Table 2.

Hazard Ratios and 95% Confidence Intervals for fibroids by intake of fruit and vegetables among premenopausal women in the Nurses’ Health Study II, 1991 through 2009

Intake in servings per day or per week	Cases	Person-years	Age-Adjusted HRa (95% CI)	Multivariable- Adjusted HRb (95% CI)
Total Fruits and Vegetables
≤2/day	1,588	161,894	1.00 (ref)	1.00 (ref)
3/day	2,139	212,103	1.00 (0.94-1.07)	0.99 (0.93-1.06)
4/day	2,135	204,702	1.03 (0.96-1.10)	1.01 (0.94-1.08)
5/day	1,515	152,703	0.98 (0.91-1.05)	0.95 (0.88-1.02)
≥6/day	2,329	221,685	1.04 (0.97-1.11)	0.99 (0.92-1.06)
ptrendc			0.42	0.57
Total Vegetables
≤2/day	3,610	367,279	1.00 (ref)	1.00 (ref)
3/day	2,838	275,296	1.02 (0.97-1.07)	1.01 (0.96-1.06)
4/day	1,700	164,485	1.02 (0.96-1.08)	0.99 (0.93-1.05)
5/day	847	79,654	1.05 (0.97-1.13)	1.01 (0.93-1.09)
≥6/day	711	66,372	1.08 (0.99-1.18)	1.02 (0.93-1.11)
ptrendc			0.07	0.76
Green Vegetablesd
≤1/week	976	112,053	1.00 (ref)	1.00 (ref)
2-6/week	5,805	563,775	1.08 (1.01-1.15)	1.06 (0.99-1.14)
1/day	1,255	121,812	1.08 (1.00-1.18)	1.05 (0.96-1.14)
>1/day	1,670	155,447	1.11 (1.03-1.21)	1.06 (0.98-1.15)
ptrendc			0.04	0.47
Orange Vegetablese
<1/week	1,078	112,200	1.00 (ref)	1.00 (ref)
1/week	1,285	133,942	0.97 (0.89-1.05)	0.97 (0.90-1.05)
2-6/week	6,162	586,838	0.98 (0.92-1.05)	0.99 (0.92-1.06)
≥1/day	1,181	120,106	0.94 (0.86-1.02)	0.93 (0.86-1.02)
ptrendc			0.18	0.10
Cruciferous Vegetablesf
<1/week	808	91,650	1.00 (ref)	1.00 (ref)
1/week	1,127	119,006	1.03 (0.94-1.13)	1.03 (0.94-1.13)
2-6/week	6,785	648,348	1.08 (1.00-1.16)	1.07 (0.99-1.15)
≥1/day	986	94,082	1.11 (1.01-1.22)	1.05 (0.96-1.16)
ptrendc			0.06	0.52
Tomatoes
<1/week	368	39,805	1.00 (ref)	1.00 (ref)
1/week	870	92,681	0.99 (0.87-1.12)	1.00 (0.89-1.13)
2-6/week	6,898	670,693	1.01 (0.91-1.12)	1.03 (0.92-1.14)
≥1/day	1,570	149,908	1.05 (0.93-1.17)	1.04 (0.93-1.17)
ptrendc			0.16	0.44
Legumesg
<1/week	880	92,432	1.00 (ref)	1.00 (ref)
1/week	1,145	115,867	0.98 (0.90-1.07)	0.99 (0.90-1.08)
2-6/week	6,873	667,062	0.98 (0.91-1.06)	0.99 (0.92-1.06)
≥1/day	808	77,725	1.03 (0.93-1.13)	1.02 (0.92-1.12)
ptrendc			0.40	0.59
Other Vegetablesh
<=1/week	426	40,562	1.00 (ref)	1.00 (ref)
2-6/week	6,175	613,561	0.93 (0.84-1.03)	0.93 (0.84-1.03)
1/day	1,169	113,088	0.94 (0.84-1.06)	0.93 (0.83-1.04)
>1/day	1,936	185,875	0.97 (0.87-1.08)	0.94 (0.84-1.05)
ptrendc			0.25	1.00
Total Fruits
<1/day	3,599	356,263	1.00 (ref)	1.00 (ref)
1/day	1,244	114,913	1.05 (0.98-1.12)	1.04 (0.97-1.11)
2/day	3,544	348,730	0.98 (0.93-1.03)	0.96 (0.92-1.01)
3/day	1,004	98,605	0.99 (0.92-1.07)	0.97 (0.90-1.04)
≥4/day	315	34,576	0.92 (0.82-1.04)	0.90 (0.80-1.01)
ptrendc			0.17	0.03
Citrus Fruiti
<1/week	1,190	121,872	1.00 (ref)	1.00 (ref)
1/week	1,054	106,013	1.01 (0.93-1.10)	1.01 (0.93-1.10)
2-6/week	5,074	484,322	1.01 (0.94-1.07)	1.01 (0.95-1.08)
≥1/day	2,388	240,879	0.98 (0.91-1.05)	0.99 (0.92-1.06)
ptrendc			0.25	0.45
Citrus Fruitsi not including juices
<1/week	3,897	199,231	1.00 (ref)	1.00 (ref)
1/week	1,969	100,747	0.96 (0.91-1.02)	0.96 (0.91-1.02)
2-6/week	3,543	165,302	1.01 (0.96-1.06)	0.99 (0.95-1.04)
≥1/day	297	15,266	0.97 (0.85-1.09)	0.93 (0.83-1.05)
ptrendc			0.96	0.43
Rosaceae Fruitj
<1/week	533	51,915	1.00 (ref)	1.00 (ref)
1/week	646	65,400	0.95 (0.84-1.06)	0.95 (0.85-1.07)
2-6/week	6,037	579,550	0.97 (0.89-1.07)	0.97 (0.89-1.06)
≥1/day	2,490	256,221	0.94 (0.85-1.03)	0.93 (0.85-1.03)
ptrendc			0.15	0.12
Fruit Juice Only
<1/week	1,570	153,916	1.00 (ref)	1.00 (ref)
1/week	906	91,564	0.99 (0.91-1.07)	1.00 (0.92-1.09)
2-6/week	4,440	418,299	1.03 (0.97-1.09)	1.06 (0.99-1.12)
≥1/day	2,790	289,308	0.99 (0.93-1.06)	1.04 (0.97-1.11)
ptrendc			0.53	0.44

CI, confidence interval; HR, hazard ratio

^aAdjusted for age (continuous) and total calories (continuous)

^bAdjusted for age at menarche (<10, 10, 11, 12, 13, 14, 15, 16,>16 y), length of menstrual cycle between ages 18-22 y (<26, 26-31, 32-50, ≥51 d), parity (nulliparous, 1, 2, 3, ≥4 pregnancies lasting >6 mo), body mass index (<19, 19-20.4, 20.5-21.9, 22-24.9, 25-29.9, ≥30 kg/m²), recent gynecologic exam (no, yes), total calories (continuous), OC use (never, past, current) and smoking history (never, past, current)

^cDetermined using category medians

^dGreen vegetables include kale, spinach, and head, romaine or leaf lettuce

^eOrange vegetables include carrots, orange winter squash and yams/sweet potatoes

^fCruciferous vegetables include broccoli, cauliflower, cabbage and Brussel sprouts

^gLegumes include peas, lima beans, lentils and tofu/soy.

^hOther vegetables include those not classified as green, orange, cruciferous, or legumes

ⁱCitrus fruits include oranges, orange juice, grapefruit and grapefruit juice

^jRosaceae fruits include prunes, prune juice, apples, apple juice, apple sauce, strawberries and peaches

Table 3 shows the classification of individual fruits and vegetables into high and low pesticide residue groups. In 2007, 18 fruits/vegetables were classified as low PRBS and 12 were classified as high PRBS. The PRBS classification stayed consistent across follow-up for some fruits and vegetables (e.g., fresh apples or pears, strawberries, and raw spinach), but changed for others. Specifically, romaine lettuce, iceberg or head lettuce, onion, salsa moved from low PRBS group in earlier questionnaire cycles to high PRBS group in 2007, while cantaloupe, cooked spinach, carrots moved from high to low PRBS over time. Brussel sprouts, mixed vegetables, avocado, prune juice, other juice, and apricots did not have a classifiable PRBS status for any of the FFQ cycles included in the PRBS analysis.

Table 3.

Fruit and Vegetable Pesticide Residue Burden Score (PRBS) Status Group at study mid-follow-up 2007^a

2007 FFQ Fruit and Vegetable Item	PRBS Status Group (Low/High)b
Apple juice or cider	Low
Bananas	Low
Cantaloupe	Low
Grapefruit	Low
Oranges	Low
Orange juice	Low
Prunes	Low
Tomatoes	Low
Eggplant, zucchini or other summer squash	Low
Broccoli	Low
Spinach, cooked	Low
Potatoes	Low
Peas	Low
Yams or sweet potatoes	Low
Orange Winter Squash	Low
Carrots	Low
Cauliflower	Low
Tofu, soybeans or other soy protein	Low
String beans	High
Kale, mustard green or chard	High
Spinach, raw	High
Romaine lettuce	High
Iceberg or head lettuce	High
Celery	High
Green, yellow or red peppers	High
Fresh apples or pears	High
Blueberries	High
Strawberries	High
Peaches or plums	High
Raisins or grapes	High

^aPesticide residue data from the Pesticide Data Program was averaged from 2004-2007.

^bFruit and vegetable items with undetermined pesticide residue status in 2007 were corn, onion, tomato juice/tomato sauce, salsa, cabbage or coleslaw, Brussel sprouts, mixed vegetables, avocado, prune juice, other juice, apricots. Some fruit and vegetable items were classified in the low PRBS group in the previous questionnaire cycles (romaine lettuce, iceberg or head lettuce, onion, salsa), while other fruits and vegetables were classified in the high PRBS group (cantaloupe, cooked spinach, carrots). The following fruit and vegetable items varied in PRBS grouping in the two previous questionnaire cycles: tomatoes, yams or sweet potatoes, raisins or grapes, peaches or plums, green/yellow/red peppers.

In the PRBS analyses, during 10 years of follow-up (1999-2009), a total of 49,927 premenopausal participants contributed to 335,225 person-years and a total of 4,195 incident cases of fibroids were reported. Higher intake of fruits and vegetables with a high PRBS was associated with a lower risk of fibroids. Participants in the 5^th quintile of high PRBS score had a 13% lower risk of fibroids compared to those in the 1^st quintile (95% CI=0.77-0.99; p_trend=0.04) (Table 4). There was no difference in fibroids risk across quintiles of low pesticide residue fruits and vegetables (HR for 5^th versus 1^st quintile=1.08; 95% CI=0.95-1.23; p_trend=0.26). When individual fruits and vegetables were examined to explore what was driving the lower risk in those with a high PRBS no individual fruits or vegetables explained this association. In sensitivity analysis, which restricted the fruit and vegetable analysis to the same time period (1999-2009) that the PRBS data was derived from, the results were not materially different.

Table 4.

Hazard ratios and 95% confidence intervals for fibroids according to pesticide residue burden score intake, Nurses’ Health Study II, 1999 through 2009

	Cases	Person-years	Age-Adjusted HR (95% CI)a	Multivariable- Adjusted HR (95% CI)b
High pesticide fruit and vegetable intake
Quintile 1	821	65,633	1.00 (ref)	1.00 (ref)
Quintile 2	838	67,096	0.97 (0.87-1.07)	0.97 (0.87-1.07)
Quintile 3	843	67,417	0.94 (0.85-1.05)	0.94 (0.85-1.05)
Quintile 4	882	67,320	0.97 (0.86-1.08)	0.97 (0.86-1.08)
Quintile 5	811	67,759	0.87 (0.76-0.98)	0.87 (0.77-0.99)
ptrendc			0.03	0.04
Low pesticide fruit and vegetable intake
Quintile 1	792	65,537	1.00 (ref)	1.00 (ref)
Quintile 2	825	66,970	1.03 (0.93-1.14)	1.03 (0.93-1.14)
Quintile 3	873	67,395	1.07 (0.96-1.20)	1.08 (0.97-1.20)
Quintile 4	857	67,319	1.07 (0.95-1.20)	1.08 (0.96-1.21)
Quintile 5	848	68,004	1.07 (0.94-1.21)	1.08 (0.95-1.23)
ptrendc			0.33	0.26

CI, confidence interval; HR, hazard ratio

^aAdjusted for age (continuous) and total calories (continuous)

^bAdjusted for age at menarche (<10, 10, 11, 12, 13, 14, 15, 16,>16 y), length of menstrual cycle between ages 18-22 y (<26, 26-31, 32-50, ≥51 d), parity (nulliparous, 1, 2, 3, ≥4 pregnancies lasting >6 mo), body mass index (<19, 19-20.4, 20.5-21.9, 22-24.9, 25-29.9, ≥30 kg/m²), recent gynecologic exam (no, yes), total calories (continuous), OC use (never, past, current) and smoking history (never, past, current)

^cDetermined using category medians

Discussion

In this prospective cohort, we observed no association with total fruit and vegetable intake and fibroid risk. However, a significant inverse association was observed for total fruit intake alone. We also observed that consumption of high pesticide residue fruits and vegetables was associated with a lower risk of uterine fibroids but no lower risk of fibroids with consumption of specific categories of fruits and vegetables was observed. The pattern seen in this study suggest pesticide residue exposure through fruits does not outweigh the small, but positive (i.e., healthy) benefits that fruit intake may have on fibroid risk.

Only a few studies have evaluated the association of fruit and vegetable intake with uterine fibroids. The Black Women’s Health Study (BWHS), the only prospective cohort to examine this association (n=22,583 premenopausal women, 6627 incident fibroid cases), observed an inverse association with total fruit and vegetable consumption and risk of fibroids (HR for 4+ servings/day vs <1/day=0.90; 95% CI=0.82-0.98; p_trend=0.03) (17). In addition, they observed a lower risk of fibroids with a higher intake of citrus fruit (≥3 servings/week compared with <1 month; HR=0.92; 95% CI=0.86-1.00; p_trend=0.01) (17). A clinic-based case-control study which included 73 women with fibroids and 210 women without fibroids, reported that total fruit and vegetable intake, assessed with a validated short-form FFQ, was associated with a lower odds of fibroids in premenopausal women (≥3 servings/week compared with never/<1 day/week; OR=0.4; 95% CI=0.2-0.9; p_trend=0.02) (15). Another hospital-based case-control study which assessed dietary intake in face to face interviews, among 600 cases and 600 controls, reported that women with uterine fibroids consumed broccoli, cabbage, tomatoes, and apples less frequently than those without fibroids (p<0.05) (16). In a case-control study based in Italy (843 cases and 1557 controls), women participants who consumed high levels of green vegetables and fresh fruit were less likely to have fibroids than those with low levels of consumption (OR=0.5, 95% CI=0.4-0.6 and OR=0.8, 95% CI=0.6-1.0, respectively) (18). Limitations of these case-control studies included lack of adjustment for total caloric intake as well as retrospectively collected diet data that were limited to examining diet at one time point most proximate to fibroid diagnosis. However, regardless of cohort or case-control study design, prior studies have generally observed lower risk of fibroids with fruit intake, which is consistent with our results, suggesting that the positive benefits of fruit consumption outweigh any potential pesticide exposure through fruit and vegetable intake.

Carotenoids are an abundant source of Provitamin A, precursors to vitamin A (34). Lycopene, a bright red carotenoid, is found in fruits such as tomatoes, watermelons, grapefruits, and papaya (35). In vitro cell culture studies have associated the retinoic acid pathway genes, derived from Vitamin A to expressions in uterine fibroids (36-38). Retinoic acid plays a role in cell proliferation, differentiation, and apoptosis (39). Additionally, retinoids have inhibited the growth of fibroids in vitro and in animal models (40). However, previously in this same cohort, we reported no association between carotenoid intake and fibroids risk (41) indicating that a different mechanism may explain the observed associations.

Current literature suggests that environmental agents, such as certain pesticides, may disrupt endocrine pathways by modifying hormone function or binding to estrogen or androgen receptors (3, 42) and such disruption may be associated with hormonally-dependent conditions (3), such as uterine fibroids. Data from a cohort of women undergoing laparoscopic surgery observed a significant positive association between dichlorodiphenyldichloroethylene (p,p’-DDE) measured in serum and fibroid diagnosis and significantly higher mean concentrations of p,p’-DDT, p,p’-DDE and four chlordane pesticides (cis-chlordane, cis-nonachlor, trans-nonachlor, and oxychlordane) measured from omental fat in women with fibroids compared to those without fibroids (3). This is consistent with evidence from both observational and laboratory studies that endocrine disrupting chemicals, which included pesticides, may contribute to the etiology of fibroids (13). Further, there is evidence that pesticide residues in fruits and vegetables impact the association between these foods and reproductive and other health outcomes (19, 32, 43, 44).

Prior observational studies have evaluated the association between organochlorine pesticides (OCPs) with uterine fibroids, however most have been retrospective in nature (3, 45, 46). The Study of Environment, Lifestyle, and Fibroids (SELF), which includes reproductive aged Black women residing in the Detroit, Michigan area, is the only prospective study to evaluate OCPs with the incidence of uterine fibroids. Among 1,693 participants, OCPs plasma concentrations were not associated with a higher risk of fibroids with a HR of 0.58 (95% CI=0.32-1.04) for the 90%+ percentile of total OCPs plasma concentration compared to the <50% percentile, with similar results for individual plasma OCPs (46). This study was similar to our results, as we also saw no elevated risk of fibroids among those with a higher PRBS score.

In the current study, pesticide exposure was not directly measured. The PRBS method has been previously validated with PRBS scores positively associated with ranked scores of all pesticide metabolites that were measured in urinary and serum concentrations among participants (n=3,679) of the National and Nutrition Examination Survey (NHANES) study (33). This provides evidence to support the PRBS method in providing a valid rank of pesticide exposure from fruits and vegetables in epidemiology studies in lieu of direct biomarker data exposure measurements. Moreover, PRBS-assessed high and low-pesticide residue fruits and vegetable intake is related to non-overlapping serum metabolomic patterns in women (47). However, the PRBS method is based upon the PDP monitoring data that changes from year to year and that may not be available for a particular fruit or vegetable for any given year. The averaging of the PDP data across 4-year increments are likely to account for this variation and potential gaps in the pesticide monitoring data. A more significant challenge is likely to be the variation in pesticide toxicology and half-lives that is not accounted for in the PRBS method; some pesticides metabolize quickly, while others can accumulate and reside in our bodies for several years. However, previous studies objectively and subjectively measured pesticide exposure and validated PRBS to rank participants by pesticide residue exposure (20, 21). Additionally, measurement error may exist in calculating the PRBS score. As described above, the PRBS score is based on the summation of tertile rank values corresponding to three different measures (i.e. % of samples with any detectable pesticides; % of samples with pesticides > tolerance level; and % of samples with ≥ 3 types of pesticides), and although each measure represents a different type of exposure, they are all considered to contribute equal weight to the final score (48). The impact of this measurement error would have been expected to push our results toward the null, since any misclassification would be non-differential with respect to the outcome. While the PDP consists of samples of fruits and vegetables from select U.S. states to measure the distribution of pesticide residue exposure (33), agricultural practices are relatively consistent across the U.S. and most of the general population consumes non-local produce, resulting in a score that is appropriate to apply on a national level.

Limitations of our study should be considered. Although we utilized validated FFQs (24), self-reported questionnaires are subject to error which would most likely result in non-differential misclassification. Despite the potential non-differential misclassification, we observed an association for total fruit intake, which could indicate that the true associations may be stronger than observed. In addition, we relied on self-report of fibroids and there may have been fibroids that were not diagnosed among those classified as not having a fibroid diagnosis. However, by utilizing a case definition of fibroids that was restricted to cases confirmed by ultrasound or hysterectomy we are confident that participants reporting a fibroid diagnosis are true cases and represent those most clinically relevant. Finally, residual or unmeasured confounding may be a possibility; however, were able to adjust for a variety of confounders.

To our knowledge this is the first study to examine the association between high- and low-pesticide-residue fruit and vegetable intake and risk of fibroids, and only the second prospective study to examine the association between fruit and vegetable intake and fibroid risk. Strengths of this study include 18 years of follow-up and dietary assessments at multiple timepoints. In addition, the data present in the USDA Pesticide Data Program of high- and low-pesticide-residue fruit and vegetable intake aligned with the FFQ data, which allowed us the unique opportunity to examine pesticide residue burden and fibroids risk in a large, prospective study. A limitation of our study is that it consisted of predominantly White participants, and we were not able to examine associations by race. Future studies that examine the PRBS score in Black participants who are disproportionately impacted by fibroids may provide additional insight.

In conclusion, our findings of a lower risk of fibroids with higher intake of fruit and vegetables with high pesticide residue and a lower risk with a higher intake of total fruits indicate that the benefits of fruit and vegetable consumption may outweigh the potential impact of pesticide residue exposure. Future studies are needed to confirm if the nutrient properties associated with fruits specifically (regardless of pesticide residue) are particularly protective against fibroids. Further, given the age range of our study population, additional studies examining an exposure window more proximal to fibroid initiation among younger participants and assessing class specific pesticides are needed.

Data availability:

Further information including the procedures to obtain and access data from the Nurses’ Health Studies is described at https://www.nurseshealthstudy.org/researchers (contact nhsaccess@channing.harvard.edu) and https://sites.sph.harvard.edu/hpfs/for-collaborators/