Systemic Effects of Prenatal Carotenoid Supplementation in the Mother and her Child: The Lutein and Zeaxanthin in Pregnancy (L-ZIP) Randomized Trial —Report Number 1

Abstract

Background

Adding carotenoids, particularly lutein (L) and zeaxanthin (Z), to prenatal micronutrient formulations has been promoted to enhance infant visual and neural development and to maintain maternal health. Although these claims are biologically plausible, they are not yet supported by a compelling prospective trial.

Objective

We investigated the effect of prenatal carotenoid supplementation on biomarkers of maternal and infant systemic carotenoid status.

Methods

We randomly assigned 47 first trimester pregnant subjects by 1:1 allocation to receive standard-of-care prenatal vitamins plus a 10 mg L and 2 mg Z softgel (the Carotenoid group) or standard-of-care prenatal vitamins with a placebo softgel (the Control group) for 6–8 mo. Maternal carotenoid concentrations in the serum and skin at the end of each trimester and postpartum were measured with HPLC and resonance Raman spectroscopy, respectively. Infants’ systemic carotenoid status was assessed using similar techniques but optimized for infants. Repeated measures and paired t-tests were determined, and a P value < 0.05 was considered statistically significant.

Results

After supplementation, there was a statistically significant increase in maternal serum L + Z concentrations, serum total carotenoid concentrations, and skin carotenoid status (P < 0.001 for all) in the Carotenoid group relative to the Control group at all study time points. Similarly, infants whose mothers were in the Carotenoid group had a significant 5-fold increase in cord blood L + Z concentrations, over a 3-fold increase in cord blood total carotenoids, and a 38% increase in skin carotenoids compared with the Control group (P < 0.0001 for all). In addition, there was a strong positive, statistically significant correlation between postpartum maternal and infant systemic carotenoid status (P < 0.0001).

Conclusion

Prenatal carotenoid supplementation significantly increased maternal and infant systemic (skin and serum) carotenoid status, which may benefit pregnant women and their infants’ health.

This trial was registered at clinicaltrials.gov as NCT03750968.

Introduction

Xanthophyll carotenoids have recently gained considerable attention in research because of their emerging role in visual and cognitive performance throughout the human lifespan [[1], [2], [3]]. This recognition stems from their crucial function as antioxidants and blue light filters, which can enhance optical performance and cognitive function [4, 5]. Humans, like other animals, cannot synthesize carotenoids de novo. Therefore, carotenoids must be obtained exclusively through dietary sources such as green leafy vegetables, orange or yellow fruits and vegetables, eggs, and dairy products, or by supplementation [[6], [7], [8]]. Many carotenoids exist in nature; however, ∼ 50 are present in the human diet, and ∼10 are detectable in systemic circulation [[9], [10], [11]]. Of these, lutein (L), zeaxanthin (Z), and meso-zeaxanthin (MZ), a metabolite of L, uniquely accumulate in the macula, the central part of the retina, mediating distinct spatial vision. They are collectively referred to as the macular pigment (MP) [[12], [13], [14], [15], [16]].

MP plays an essential role in visual performance [[17], [18], [19], [20]], especially in mitigating the onset or progression of advanced age-related macular degeneration (AMD), the leading cause of irreversible blindness among adults in industrialized countries [[21], [22], [23]]. In addition, carotenoids’ functions in cognition have been demonstrated in Alzheimer’s disease and memory performance in children [[24], [25], [26], [27], [28]]. Furthermore, several observational studies have reported a direct relationship between carotenoid intake and positive health outcomes [29, 30]. These findings give credence to carotenoids’ unique role in human life.

During pregnancy, the fetus relies on the mother for all its nutrients. Therefore, prenatal maternal nutrition (through diet or supplementation) is crucial to pregnancy outcomes and maternal and infant health [31, 32]. Inadequate nutrient supply during pregnancy may predispose the infant to subsequent developmental delays and future metabolic diseases, particularly chronic noncommunicable diseases such as hypertension, diabetes, and coronary artery disease [[33], [34], [35]]. Prenatal micronutrient supplementation has been the standard-of-care for expectant mothers worldwide [36]. This supplementation ensures that mothers have enough essential nutrients to support the developing fetus. Although many of these prenatal micronutrients, such as DHA, folic acid, and vitamins A and D, have had extensive research, the benefits of carotenoids remain to be determined through adequately powered prospective trials [[37], [38], [39]].

The United States FDA and the European Food Safety Authority (EFSA) have designated carotenoids as GRAS for human consumption at doses of ≤1 mg/kg body weight/d L and 0.75 mg/kg body weight/d Z [40], yet many Americans consume only 1–2 mg L and 0.2–0.4 mg Z daily [41]. Therefore, an alternative approach that optimizes systemic carotenoid status, especially during pregnancy, is desirable. The addition of carotenoids to prenatal multivitamin formulations was nonexistent until Abbott Nutrition introduced a prenatal supplement with 6 mg L (Similac Prenatal Vitamins, distributed by Abbott Nutrition) into the American market in 2014 with the stated intentions of enhancing infant visual and neural development and maintaining maternal health.

Existing evidence supports the selective concentration of these dietary carotenoids in the human eye and neural tissues in utero, especially during the third trimester of pregnancy [[42], [43], [44]]. In addition, MP is detectable at birth, correlates with infant and maternal serum Z concentrations [45], and steadily increases through the first 7 y of life [46], suggesting that carotenoids may play a functional and protective role in infant visual and cognitive development. In the third trimester of pregnancy, mothers transfer their stores of nutrients via the placenta to support the infant’s CNS development [47], potentially putting mothers at risk of systemic and ocular carotenoid depletion. Therefore, the assertion that the addition of carotenoids to prenatal supplements will enhance the mother’s and her infant’s health seems biologically plausible but has yet to be prospectively confirmed. At least in part as a result of the lack of compelling prospective clinical trial data to support its claims, Abbott’s product achieved low market penetration and was subsequently withdrawn from the market.

At this point, no prospective study has examined the effect of prenatal carotenoid supplementation on biomarkers of maternal and infant systemic and ocular carotenoid status or whether the third trimester of pregnancy is truly a period of maternal carotenoid depletion. We hypothesized that prenatal carotenoid supplementation would significantly affect ocular and systemic biomarkers of maternal and infant carotenoid status relative to a matched, standard-of-care prenatal supplement without added L and Z. Furthermore, we speculated that infants with the highest systemic and ocular carotenoid status may have a more mature foveal structure at birth. We investigated the above hypotheses through a randomized, controlled clinical trial known as the Lutein and Zeaxanthin in Pregnancy (L-ZIP) trial.

Methods

A detailed study design and methodology have been reported previously [48]. Outlined below is a summary of the L-ZIP study design and procedures. The L-ZIP trial is a phase 2, single-site, prospective, active-controlled, double-masked, randomized clinical trial conducted at the John A. Moran Eye Center, University of Utah, Salt Lake City, UT, USA. The study participants were randomly assigned (1:1 allocation) to the 2 intervention groups: the Carotenoid and Control groups. The L-ZIP trial is registered at clinicaltrials.gov as NCT 03750968. The first participant’s first visit was in September 2019, and the last participant’s final visit was in January 2022.

Before study enrollment and assessments, all subjects provided written informed consent in accordance with International Conference on Harmonization (ICH) Guidance E6. The institutional review board of the University of Utah granted ethical approval for the L-ZIP study. The L-ZIP trial complied with the tenets of the Declaration of Helsinki, the ICH Harmonized Tripartite Guidance for Good Clinical Practice (IHC-GCP E6 [R1]), and fully adhered to the code of ethics regarding participant enrollment, study assessment, and data protection. A data safety and monitoring board (DSMB), consisting of a maternal-fetal medicine specialist and a pediatric ophthalmologist not otherwise involved in the study, met to review progress and adverse events every 6 mo. No changes to study methods were warranted upon study initiation.

L-ZIP inclusion criteria were: Pregnant women of all races and ethnicities with uncomplicated obstetric histories, aged ≥18 y, who planned to deliver at the University of Utah Hospital by either vaginal or caesarian section. Subjects were excluded from the study when they regularly (i.e., daily for the past 6 mo) consumed carotenoid supplements with >0.5 mg L and/or Z, had significant eye disease associated with MP abnormalities such as Stargardt disease, albinism, or macular telangiectasia type II (MacTel), or had any health conditions associated with high-risk pregnancies such as adolescent pregnancy, multifetal pregnancy, current or past history of diabetes, preeclampsia, previous premature delivery, drug abuse, or other significant medical illness. The screening assessments for eligibility included demographic information, visual function measures, and a comprehensive ophthalmic examination performed by a qualified ophthalmologist with a retinal subspecialty.

Upon establishing study eligibility, a research staff member randomly assigned subjects in a 1:1 allocation ratio using a computer-generated random sequence to the Carotenoid or the Control group. All study participants, clinicians, and research staff involved in the trial were masked by study assignment, as study formulations only had identification labels. The Carotenoid group subjects received active intervention opaque amber-colored softgels containing 10 mg L and 2 mg Z in safflower oil for daily consumption that were supplied by Kemin Health L.C. The Control group received an identical appearing softgel containing only the safflower oil vehicle. The study’s active intervention is identical to the carotenoid dosage used in the Age-Related Eye Disease Study 2 (AREDS2) [22], and active and control softgels were analyzed using HPLC every 6-mo to assure purity and stability. All subjects received a commercially available standard-of-care prenatal multivitamin for daily consumption (prenatal multivitamin + DHA, distributed by Walgreen Co.). Each prenatal multivitamin + DHA softgel contained 1200 μg retinol activity equivalent vitamin A, 60 mg vitamin C, 400 international units (IU) vitamin D3, 10 mg vitamin E, 1.5 mg vitamin B1, 1.7 mg vitamin B2, 18 mg niacin equivalent (NE) vitamin B3, 2.6 mg vitamin B6, 800 μg folic acid, 4 μg vitamin B12, 38 mg calcium, 27 mg iron, 25 mg zinc, and 250 mg DHA.

Although the subjects had no study-related dietary restrictions, they were asked to avoid carotenoid-containing supplements for the 6–8 mo of the study duration. Frequent phone calls and text messages were sent to remind study participants to ensure compliance with the study intervention intake. In addition, we counted the remaining pills at each study visit. The baseline first trimester study visit (T1) was before 14 wk of gestational age (GA). The second trimester study visit (T2) was between 22 and 26 wk, and the third trimester study visit (T3) was between 37 and 39 wk. The postpartum study visit (P) was within 2 wk after delivery.

Study outcomes

Maternal systemic biomarkers of carotenoid status (skin and serum carotenoid concentrations) and ocular carotenoids (to be presented in a subsequent report) were the predeclared primary outcome measures of the L-ZIP study [48]. The secondary outcomes assessed infant carotenoid status. Of note, there were no changes to study outcomes after trial commencement.

Demographic and lifestyle questionnaire

The demographic and lifestyle information of participants included contact details (i.e., name, date of birth, phone number(s), place of residence, e-mail address), self-reported race and ethnicity, occupation, medical history, ocular history, smoking habits (history and frequency), and alcohol consumption (mean intake per week and frequency). We assessed maternal dietary intake of carotenoids at each study visit using the LZQ™ quantitative food frequency questionnaire adapted from [7] and analyzed at Tufts University by Dr. Elizabeth Johnson. Although unpublished, the LZQ™ is validated against the Willett food frequency questionnaire and contains ∼ 90% of the L and Z foods consumed in the United States on the basis of the National Health and Nutrition Examination Survey data. At every study visit, each participant fills out the LZQ™. The screening LZQ™ was from the start of pregnancy to the screening visit (T1, before 14 wk GA), T2 was from the screening visit to the T2 visit (∼12 wk), T3 covered the time between T2 and T3 (∼13 wk), and P was between T3 and P (4 wk). Participants responded freely to how often they eat carotenoid-containing foods as the number of times per day, times per week, and times per month. The LZQ™ obtained from [7] has 10 food categories: bread, cereals, condiments, sauces, fruits, nuts, pasta, eggs, snacks, and vegetables. Each category has a list of food items with a prompt for clarification. There are 71 food items in total on the LZQ™ questionnaire. The bread category has 5 food items (i.e., yellow or white cornmeal, cornbread muffin, corn tortilla, and bread, roll, bun, or bagel). The cereals have 10 food items: Apple Jacks, Cap’n Crunch, Corn Chex, corn flakes, corn pops, Crispix, Froot Loops, frosted flakes, Life, and Reese’s Puffs. There are 2 condiments (fat-free mayonnaise and regular mayonnaise), 2 ready-to-serve sauces (salsa or pepper/hot sauce), 1 nut (pistachios), 3 types of pasta (macaroni and cheese, egg noodles, and spinach egg noodles), 1 large egg including the yolk, and 5 snacks (Chex Mix, Cheetos, Fritos, popcorn, and tortilla chips). There are also 14 and 29 food items in the fruit and vegetable categories, respectively.

Skin carotenoid measurement

Skin carotenoid measurements in infants and their mothers were measured using resonance Raman spectroscopy. This validated noninvasive device serves as a biomarker of fruit and vegetable intake and has been proven to significantly correlate with serum total carotenoid concentrations [[49], [50], [51]]. In brief, we measured skin carotenoid levels by placing a probe on the subject’s palm, the area extending between the little finger and the wrist (adults) or the heel of the foot (infants), with a 488-nm blue laser light that does not generate heat or skin damage and is harmless to one’s vision as long as it is not viewed directly (similar in risk to a laser pointer). The device collects back-scattered light and has a holographic notch filter that rejects Rayleigh-scattered light. The Peltier-cooled spectrograph analyzes the resulting fluorescence and Raman-shifted light. The peak intensity represents the C=C vibration of carotenoids at ∼1525 cm⁻¹ (known as Raman units [RU]). The device was calibrated daily, and each measurement required 30 s of skin contact and another 30 s to obtain the reading. We took 3 measurements and used the mean for statistical analysis.

Serum carotenoid concentrations

Blood samples from mothers and from infant umbilical cords were analyzed for serum carotenoid concentrations using well-established laboratory protocols that provide baseline separations of all common dietary carotenoids [52]. At each study visit, blood samples were collected in 6 mL tubes (BD Vacutainer K2EDTA; Becton, Dickinson and Company). The collection tubes were inverted a few times to ensure thorough mixing of the anticoagulant (K2EDTA). The blood samples were allowed to sit at room temperature for 30 min and then centrifuged at ∼ 5°C for 10 min at 3000 × g using a Sorvall ST 16 Centrifuge (Thermo Fisher Scientific Inc.). After centrifugation, the serum was separated from the whole blood, kept in a well-labeled storage tube, and stored at −80°C until further analysis. We extracted carotenoids and used HPLC to measure serum carotenoid concentrations as described previously [48].

Sample size calculation

We determined the statistical power and sample size for one of our primary outcome measures, which is maternal skin carotenoids at birth, using data from a previous study by Henriksen et al. [45]. At birth, that study reported a mean maternal skin carotenoid ± SD of 34,000 ± 8300 RU. This value is 20% lower than the mean Moran Eye Center skin carotenoid of ∼42,500 RU reported for our ancillary AREDS2 study [53]. Therefore, with a power of 0.90 and an alpha level of 0.05, we calculated a sample size of 20 for each study group to detect an 8500 RU difference. To account for potential ineligibility because of premature birth or other problems (∼10%) and to anticipate expected noncompliance and loss to follow-up (∼20%), a target enrollment of 30 subjects in each study group was considered appropriate.

Statistical analysis

All data analyses for this study utilized Stata 14.0 software (StataCorp). Only participants who completed the first and last study visits (n = 41) were included in the study’s statistical analysis in accordance with the predeclared L-ZIP protocol [48]. Repeated measures ANOVA was used for between-group comparisons of changes in outcome variables over time. Bonferroni adjustments were made for multiple comparisons. Paired t-tests compared the carotenoid status of infants whose mothers were in the 2 study groups. Correlation analysis determined the relationships between maternal and infant systemic carotenoid status. Statistical significance was set at P < 0.05 for all analyses.

Results

Enrollment and baseline characteristics

The first and last L-ZIP subjects enrolled in September of 2019 and July of 2021, respectively. Figure 1 provides the CONSORT flow diagram for the L-ZIP trial. Because of the ongoing COVID-19 pandemic, many potentially eligible participants declined enrollment, but those who enrolled rarely missed any study visits, and we had a <15% dropout rate in both groups. Since our enrolled subjects were highly engaged and compliant, we terminated enrollment at 47 randomly assigned subjects because it was apparent that we would reach our per-protocol target of 20 subjects per group. Table 1 summarizes the maternal baseline characteristics of the 41 study participants (Carotenoid group, n = 21, and Control group, n = 20) who completed the postpartum study visits.

FIGURE 1.

The CONSORT flow diagram for the L-ZIP trial. Abbreviation: L-ZIP, lutein and zeaxanthin in pregnancy.

TABLE 1.

Baseline maternal demographics, lifestyle, and systemic carotenoid concentrations of the Carotenoid and the Control study groups¹

Variables	Carotenoid group (n = 21)	Control group (n = 20)
Age, y	30.8 ± 3.1	29.1 ± 3.7
BMI, kg/m2	23.9 ± 3.7	25.4 ± 4.6
Race
Caucasian	21 (100.0)	20 (100.0)
Non-Caucasian	0 (0.0)	0 (0.0)
Ethnicity
Non-Hispanic	20 (95.2)	19 (95.0)
Hispanic	1 (4.8)	1 (5.0)
Smoking habits
Never smoked	21 (100.0)	20 (100.0)
Smoked	0 (0.0)	0 (0.0)
Alcohol frequency
Never	20 (95.2)	20 (100.0)
Occasional	1 (4.8)	0 (0.0)
Estimated dietary intake of L and Z, mg/d
Lutein	4.0 ± 5.7	3.4 ± 3.1
Zeaxanthin	0.3 ± 0.3	0.2 ± 0.2
Lutein + Zeaxanthin	4.3 ± 5.9	3.6 ± 3.1
Serum carotenoids, ng/mL
Lutein	219.7 ± 103.3	182.6 ± 68.2
Zeaxanthin	37.9 ± 16.8	34.4 ± 15.3
Lutein + Zeaxanthin	257.6 ± 116.7	217.0 ± 82.5
Total carotenoids	780.1 ± 347.9	674.7 ± 284.3
Skin carotenoids, RU	39981 ± 9495	39436 ± 8733

Abbreviations: BMI, body mass index; n, number of participants; RU, Raman units; Total carotenoids: summation of lutein, zeaxanthin, β cryptoxanthin, α carotene, β carotene, and lycopene.

To convert ng/mL L + Z to nmol/L, multiply ng/mL by 1.758.

¹Data shown are mean ± standard deviation for continuous data and frequency (%) for categorical data.

Compliance and adverse events

Of the 41 subjects who completed their postpartum visits, only one subject missed her T3 visit because of unexpectedly early labor. Subjects’ adherence to their assigned study supplement assessed by pill counting was 90%, 85%, and 76% at the T2, T3, and P visits, respectively. Supplement compliance did not differ significantly between the Carotenoid and Control groups over the study period (P > 0.05). One Carotenoid group infant missed their skin carotenoid assessment, and 2 cord blood samples were missing for each of the 2 experimental groups.

Table 2 details subjects’ reports of adverse events after the use of the study intervention over the study duration for all 47 randomly assigned subjects. Despite some subjects having multiple adverse events, there was no appreciable difference between the 2 study groups. In addition, most of the adverse events reported were not considered unusual occurrences during pregnancy by the investigators and the DSMB. Two infants whose mothers were in the Carotenoid group had severe adverse events. One had a clinical diagnosis of Down syndrome at 6 wk of age, which was confirmed using genetic testing. Another infant was born prematurely with multiple congenital anomalies; both the mother and child missed their postpartum study visit and were excluded from data analysis. These 2 severe adverse events undoubtedly originated before study enrollment and were therefore unrelated to the study intervention.

TABLE 2.

Maternal and infant adverse events¹

Adverse events	Carotenoid group (n = 24)	Control group (n = 23)
Maternal
Any adverse events	6	8
Gestational diabetes	1	2
Gestational hypertension	1	0
Preeclampsia	0	2
Anemia	0	1
Exacerbated acne	1	0
Itching	1	0
Torn placenta at delivery	0	1
Cholestasis	0	1
Sinus infection	1	1
Worsening depression	1	1
Nose laceration	1	0
COVID-19	2	1
Preterm delivery2	1	0
Infant
Small for gestational age	1	0
Heart murmur3	0	1
Down syndrome2	1	0
Multiple congenital anomalies2	1	0

Abbreviation: n, number of participants.

¹Adverse events after the initiation of the study intervention over the study duration for all 47 randomized subjects. Any adverse events represent the number of participants in each group that reported adverse events (which could be single or multiple for a participant).

²Severe adverse events unrelated to study intervention.

³Required no treatment.

Changes in maternal outcome variables over time

Change in maternal serum L + Z concentrations

Table 3 presents changes in maternal systemic carotenoid status at each study visit and the statistical significance between the Carotenoid and Control groups over the study period. Figure 2A illustrates the mean serum L + Z concentrations between the 2 study groups from baseline to postpartum. We found a statistically significant increase in maternal serum L + Z concentrations at all study time points in the Carotenoid group relative to the Control group after supplementation (P < 0.001). Notably, a decline in serum L + Z at the postpartum visit relative to the T3 visit was significant in the Carotenoid group (P < 0.001). There were no statistically significant changes in serum L + Z concentrations between the Control and Carotenoid group at all study time points. In Figure 2B, we observed a similar pattern for serum total carotenoid concentrations (the summation of L, Z, β-cryptoxanthin, α-carotene, β-carotene, and lycopene).

TABLE 3.

Change in maternal systemic carotenoid status from baseline to postpartum between study groups using repeated measures ANOVA¹

Variables	Carotenoid group (n = 21)	Control group (n = 20)	Mean difference	P value
Serum L + Z, ng/mL
T1	257.6 ± 25.5	217.0 ± 18.4	40.6 ± 97.6	1.000
T2	1193.5 ± 96.5	246.5 ± 18.8	947.0 ± 97.6	<0.0012
T33	1388.3 ± 133.5	301.1 ± 29.6	1087.2 ± 98.8	<0.0012
P	1030.2 ± 86.8	252.0 ± 33.4	778.2 ± 97.6	<0.0012
Serum total carotenoids, ng/mL
T1	780.1 ± 75.9	674.7 ± 63.6	105.4 ± 135.5	1.000
T2	1777.8 ± 144.4	716.4 ± 64.5	1061.4 ± 135.5	<0.0012
T33	1947.0 ± 144.5	719.2 ± 60.3	1227.8 ± 137.1	<0.0012
P	1416.0 ± 100.8	569.7 ± 48.7	846.3 ± 135.5	<0.0012
Skin carotenoid, RU
T1	39,981 ± 2072	39,436 ± 1953	545 ± 2868	1.000
T2	50,188 ± 1730	39,111 ± 1839	11,077 ± 2868	<0.0012
T33	57,100 ± 2309	38,151 ± 1893	18,949 ± 2903	<0.0012
P	55,419 ± 2421	38,475 ± 1932	16,944 ± 2868	<0.0012

Abbreviations: L + Z, lutein + zeaxanthin; n, number of participants; P, postpartum; RU, Raman units; Serum total carotenoid: summation of lutein, zeaxanthin, β cryptoxanthin, α carotene, β carotene, and lycopene; T1, first trimester/baseline; T2, second trimester; T3, third trimester.

The mean difference is Bonferroni adjusted. To convert ng/mL L + Z to nmol/L, multiply ng/mL by 1.758.

¹The data shown are the mean ± standard error.

²Statistically significant difference between the 2 study groups at 0.05 level.

³One subject missed her T3 visit because of early labor.

FIGURE 2.

Changes in maternal serum L+Z concentrations (A), serum total carotenoid concentrations (B), and skin carotenoids (C) between the Carotenoid (n = 21) and the Control (n = 20) groups over the study duration. There was a statistically significant difference in maternal serum L + Z concentrations, serum total carotenoid concentrations, and skin carotenoids between the Carotenoid and the Control groups at all study time points (i.e., T2 to P) after the initiation of the study intervention (P < 0.001, for all). Abbreviations: T1, first trimester/baseline (before 14 wk of gestational age); T2, second trimester (between 22 and 26 wk of gestational age); T3, third trimester (between 37 and 39 wk of gestational age); P, postpartum (within 2 wk after delivery). Error bars represent 95% CI. To convert ng/mL L+Z to nmol/L, multiply ng/mL by 1.758.

Change in maternal skin carotenoids

From Table 3 and Figure 2C, there was a statistically significant increase in maternal skin carotenoid status in the Carotenoid group relative to the Control group at all study time points after initiation of supplementation (P < 0.001). In addition, we noted a steady rise in skin carotenoid status from baseline to the third trimester and a statistically insignificant decline at postpartum (P = 1.000) in the Carotenoid group. In contrast, there was no significant change in skin carotenoid levels over time in the Control group.

Change in maternal dietary L + Z intake

Repeated measures ANOVA of the LZQ food frequency questionnaire between the Carotenoid and Control groups showed no statistically significant differences over the study duration (P > 0.05 for all).

Changes in infant outcome variables

Infants’ cord blood carotenoid concentrations

As shown in Table 4, infants whose mothers were in the Carotenoid group had a 5-fold increase in serum L + Z concentrations compared with the Control group, which was statistically significant (P < 0.0001). Likewise, cord blood total carotenoids concentrations significantly increased by nearly 3.5-fold for infants in the Carotenoid group relative to the Control group (P < 0.0001).

TABLE 4.

Demographic and systemic carotenoids status of infant’s whose mothers were in the Carotenoid and Control Study Groups¹.

Variables	Carotenoid group (n = 21)	Control group (n = 20)	Mean difference	P value
GA, d	277.3 ± 1.9	273.8 ± 1.3	3.6 ± 2.4	0.136
Age, d	4.5 ± 0.5	5.1 ± 0.6	-0.6 ± 0.7	0.446
Weight, kg	3.2 ± 0.1	3.3 ± 0.1	-0.1 ± 0.1	0.284
Sex				0.161
Female	13 (61.9)	8 (40.0)
Male	8 (38.1)	12 (60.0)
Delivery mode				0.929
Vaginal	16 (76.2)	15 (75.0)
C-section	5 (23.8)	5 (25.0)
Cord blood carotenoids, ng/mL
L + Z	270.5 ± 29.122	54.5 ± 6.63	216.0 ± 30.6	<0.00014
Total carotenoids	324.4 ± 31.62	93.0 ± 9.33	231.3 ± 33.7	<0.00014
Skin carotenoid, RU	31,656 ± 12795	23,011 ± 869	8645 ± 1546	<0.00014

Abbreviations: C-section, cesarean section; GA, gestational age; L + Z, Lutein + Zeaxanthin; RU, Raman units; Total carotenoid, summation of lutein, zeaxanthin, β cryptoxanthin, α carotene, β carotene, and lycopene.

To convert ng/mL L + Z to nmol/L, multiply ng/mL by 1.758.

¹Data shown are the mean ± standard error for continuous data and frequency (%) for categorical data.

²Number of participants is 19.

³Number of participants is 18.

⁴Statistically significant difference between the 2 study groups at 0.05 level.

⁵Number of participant is 20.

Infants’ skin carotenoids

Table 4 displays changes in infants’ skin carotenoid status between the 2 study groups. Infants whose mothers were in the Carotenoid group had significantly higher skin carotenoid status (38% increase) relative to infants whose mothers were in the Control group (P < 0.0001).

Relationships between postpartum maternal and infant systemic carotenoid status

Table 5 presents the interrelationships between systemic carotenoid concentrations of mothers and their infants at their postpartum visits after combining the results from both study arms. We found that postpartum maternal serum L + Z concentrations were positively and significantly associated with infants’ cord blood L + Z concentrations (r = 0.73, P < 0.0001), cord blood total carotenoid concentrations (r = 0.71, P < 0.0001) and skin carotenoids (r = 0.64, P < 0.0001). Postpartum maternal serum total carotenoid concentrations had a positive and statistically significant correlation with infants’ cord blood L + Z concentrations (r = 0.71, P < 0.0001), cord blood total carotenoid concentrations (r = 0.71, P< 0.0001) and skin carotenoids (r = 0.60, P < 0.0001). Likewise, postpartum maternal skin carotenoids had a positive and statistically significant correlation with infants’ umbilical cord blood L + Z concentrations (r = 0.70, P < 0.0001), umbilical cord blood total carotenoid concentrations (r = 0.70, P < 0.0001) and skin carotenoids (r = 0.51, P = 0.0008), respectively. Infants’ skin carotenoids significantly correlated with umbilical cord blood L + Z concentrations (r = 0.61, P = 0.0001) and total carotenoids (r = 0.61, P = 0.0001). In addition, there was a positive, statistically significant correlation between postpartum maternal skin carotenoids and serum L + Z concentrations (r = 0.70, P < 0.0001) and serum total carotenoid concentrations (r = 0.73, P < 0.0001).

TABLE 5.

Interrelationships between postpartum maternal and infant systemic carotenoid status¹

	Infants’ carotenoid status
	Cord blood L + Z, ng/mL2	Cord blood total carotenoids, ng/mL2	Skin carotenoid, RU3
Postpartum maternal
Serum L + Z, ng/mL	0.73 (<0.0001)	0.71 (<0.0001)	0.64 (<0.0001)
Serum total carotenoids, ng/mL	0.71 (< 0.0001)	0.71 (<0.0001)	0.60 (<0.0001)
Skin carotenoids, RU	0.70 (<0.0001)	0.70 (<0.0001)	0.51 (0.0008)

Abbreviations: L + Z, lutein + zeaxanthin; RU, Raman units; Total carotenoids, summation of lutein, zeaxanthin, β cryptoxanthin, α carotene, β carotene, and lycopene.

To convert ng/mL L+Z to nmol/L, multiply ng/mL by 1.758.

¹Data shown combine results from both study arms and present as Pearson co-efficient of correlation, r (statistical significance at 0.05 level).

²Number of participants is 37.

³Number of participants is 40.

Discussion

L-ZIP is a single-site, prospective, double-masked, randomized, active-controlled clinical trial that investigated the impact of prenatal carotenoid supplementation on biomarkers of systemic and ocular carotenoid status during pregnancy. This report concentrates on the systemic carotenoid status of the mother and her newborn infant, and a forthcoming manuscript will focus on ocular carotenoid status in the mother and child and potential influences on visual system function and development. We initiated this study to provide high-level evidence that prenatal supplementation at an AREDS2-level dose of 10 mg/d L and 2 mg/d Z [22] would be safe and well tolerated during pregnancy and that this formulation would be bioavailable to both mother and child. Pill-count compliance was high, and we had a lower-than-expected dropout rate. We had no serious adverse events attributable to the study intervention, and none of the recorded adverse events were unexpected during a routine pregnancy. We were also interested in determining whether carotenoid depletion was detectable in the mother, as she transfers a portion of her carotenoid stores to her child during the last trimester of pregnancy.

We documented a significant rise in systemic carotenoid concentrations in the Carotenoid group relative to the Control group at all study time points after the initiation of supplementation. Our results are relevant because maternal carotenoid intake and status are thought to positively influence fetal retinal and brain development [3, 54]. Given that the developing fetus undergoes various growth stages that require different amounts of nutrients, the third trimester is crucial, as there is an obligatory transfer of maternal carotenoids via the placenta to support these developmental processes, which may even deplete carotenoid stores in some pregnant individuals. The skin and serum carotenoid status within the Control group remained unchanged throughout the study. This finding contrasted with our prior cross-sectional study that included a large number of ethnically diverse lower socioeconomic mothers [45], possibly because many of our study’s subjects were hospital and university employees who were willing to enroll in a hospital-based clinical study that required multiple in-person visits during the COVID-19 pandemic. These individuals are likely to be more aware of the importance of good nutrition and lifestyle and their impact on health. This is evidenced by our population’s mean BMI at enrollment of ∼ 25 kg/m², the minimal use of tobacco or alcohol, and the rather high daily intake of dietary L and Z of ∼ 4 mg/d, considerably above the typical 1–2 mg/d for United States adults [6]. In addition, we specifically excluded diabetics, adolescents, and other high-risk pregnancies. We suspect that when this study is eplicated in a nutritionally compromised population, the postulated third trimester decline would be more evident.

The observed strong correlations of maternal skin and serum carotenoid confirm that the skin carotenoid measures could be a reliable biomarker to assess maternal carotenoid status during pregnancy [55]. Although serum carotenoid assessment using HPLC is considered the gold standard [1], skin carotenoid measurement with a validated noninvasive device could be a convenient means to assess human carotenoid status in remote areas [50, 56]. Given carotenoids’ roles in human cognition, visual performance, and general health across the lifespan [3, 54, 57], it could be worthwhile to use newer, portable reflection-based skin carotenoid measures of carotenoid status in nutrient-deficient regions and countries to identify at-risk individuals to target for dietary counseling and supplementation. Indeed, we have recently conducted one such study using the Veggie Meter (Longevity Link) in Nepal to show that skin carotenoids can be used as a biomarker for vitamin A deficiency in children and pregnant women [58].

Generally, carotenoid concentrations were lower in cord blood relative to the mothers. This is consistent with observational studies [45, [59], [60], [61]]. Although the exact mechanism underlying the infants’ lower carotenoid concentrations is not well understood, one likely reason may be that high-density lipoprotein, the primary transporter of L + Z, is lower in cord blood compared with maternal blood [1, 47]. Hence, the decreased concentrations of L + Z in cord blood may be attributable to a reduced transport capacity of carotenoids.

The strengths of our clinical study lie in the fact that it was randomized, active-controlled, and double-masked. Maternal and infant systemic carotenoid assessments were performed with established and validated devices and techniques that were initiated during the first trimester and continued throughout pregnancy. Moreover, an independent data safety and monitoring committee with expertise in the study area provided guidance throughout the trial. The desired sample size was achieved, and subject retention and compliance were excellent.

A limitation of this study was its lack of participant racial, ethnic, and socioeconomic diversity. This was driven by the population profile of Utah and the challenges of recruiting during a pandemic. In addition, by design, this first-in-humans, randomized clinical trial was limited to low-risk pregnancies. Our data and experience gained from the L-ZIP study can be used to design and power future prospective clinical trials in more diverse subjects and in medically and nutritionally higher-risk populations. This is of particular importance because we have recently shown that prenatal L and Z are protective against oxygen-induced retinopathy in a mouse model of retinopathy of prematurity [62]. Because of funding and logistical limitations, we could not follow-up with the mothers and their infants beyond their postpartum visit. This meant that we were not able to assess the role of maternal breastmilk as a continued source of L and Z during the first months of life with or without postnatal supplementation, nor could we evaluate any potential cognitive or visual function benefits of prenatal carotenoid supplementation. We hope to address this latter limitation in a future follow-up study on these children when they are old enough to test reliably.

In conclusion, prenatal carotenoid supplementation with 10 mg L and 2 mg Z per day substantially increased maternal and infant systemic (skin and serum) carotenoid concentrations. The ocular effects of this intervention in the mothers and their children will be addressed in a forthcoming report. Although not yet proven, we are optimistic that supplemental L and Z could provide health benefits to pregnant women and their babies comparable to more established prenatal nutrients such as folate and DHA. The results of the L-ZIP study will facilitate the design of future large-scale studies of prenatal L and Z supplementation in normal and high-risk pregnancies and guide policy decisions on prenatal carotenoid recommendations.

Funding

This study was supported by grants from the National Institutes of Health (EY029857 and EY014800; Bethesda, MD, USA) and an unrestricted grant from Research to Prevent Blindness (New York, USA) to the Department of Ophthalmology and Visual Sciences, University of Utah, Salt Lake City, UT, USA. Of note, the funding sources (NIH Grants EY029857 and EY014800 and the Research to Prevent Blindness) had no role in the study’s design, collection, analysis, and interpretation of the data, writing of the report, or the decision to submit the report for publication.

Author disclosures

The authors report no conflicts of interest.

Author responsibilities

The authors’ responsibilities were as follows –PSB and MWV designed the research; SJA and EKA collected study data; PSB and DYH provided study oversight and coordination; EKA, RA and JEG performed HPLC; EKA performed statistical analysis; and EKA and PSB wrote the paper. PSB had primary responsibility for final content. All authors have read and approved the final manuscript.

Data Availability

Data described in this manuscript will be made available from the corresponding author upon reasonable request.