Prospective Evaluation of Mango Fruit Intake on Facial Wrinkles and Erythema in Postmenopausal Women: A Randomized Clinical Pilot Study

Mangos are rich in β-carotene and other carotenoids, along with several phenolic acids that may provide oxidant defense and photoprotection to the skin. The objectives of this study are to investigate the effects of Ataulfo mango intake on the development of facial wrinkles and erythema. A randomized two-group parallel-arm study was conducted to assess 16 weeks of either 85 g or 250 g of mango intake in healthy postmenopausal women with Fitzpatrick skin type II or III. Facial photographs were captured at weeks 0, 8, and 16, and wrinkles at the lateral canthi and erythema at the cheeks were quantified. Skin carotenoid values were measured with reflection spectroscopy. Deep wrinkle severity decreased significantly in the 85 g group after 8 (p = 0.007) and 16 (p = 0.03) weeks compared to baseline measures. In contrast, those in the 250 g group showed an increase after 16 weeks in average wrinkle severity (p = 0.049), average wrinkle length (p = 0.007), fine wrinkle severity (p = 0.02), and emerging wrinkle severity (p = 0.02). Erythema in the cheeks increased with 85 g of mango intake (p = 0.04). The intake of 85 g of mangos reduced wrinkles in fair-skinned postmenopausal women, while an intake of 250 g showed the opposite effect. Further studies feeding 85 g of mangos are warranted.


Introduction
Skin aging is generally classified as intrinsic and extrinsic. Intrinsic aging includes genetic factors that influence pigmentation of the skin, skin composition and thickness, and hormonal composition [1]. Extrinsic aging leads to premature aging of the skin and includes lifestyle and environmental factors such as smoking, temperature and humidity, and ultraviolet (UV) radiation [1]. Photoaging involves changes in the skin induced by repeated exposure to UV radiation and is most often seen as the primary form of extrinsic aging [2]. Wrinkles are a common result of these factors [3].
Numerous dietary factors can modulate skin health. For example, consumption of carotenoid-rich kale extracts in humans reduced radical formation, prevented collagen I degradation, and improved the extracellular matrix [4,5], while supplementation with both carotenoids and vitamin C has been

Skin Carotenoids
Skin carotenoids (SCs), units in mm wavelength, were measured in the right index finger after cleaning with alcohol by reflection spectroscopy (Veggie Meter ® , Longevity Link Corporation, Salt Lake City, UT, USA). The device has been validated to correspond with plasma carotenoid levels and was calibrated before each use [21].

Blood Pressure and Lipids
Three blood pressure readings were obtained, five minutes apart after a fifteen-minute seated rest using an automated oscillometric unit and averaged to obtain systolic (SBP) and diastolic (DBP) values (Vital Spot, VSM 300, Welch Allyn, Skaneateles Falls, NY, USA). Plasma lipids were analyzed for cholesterol, low-density lipoprotein (LDL), non-high-density lipoprotein (HDL), and triglycerides at the UC Davis Department of Pathology and Laboratory Medicine.

Dietary Intake
A 24-h recall was conducted at each study visit using the validated Automated Self-Administered 24-h (ASA24) dietary assessment tool (https://epi.grants.cancer.gov/asa24). A compliance log was maintained, showing the date, time, and format of mango consumption.

Statistical Analysis
An a priori power analysis showed that there was greater than 80% power to detect a 10% difference in wrinkle severity between the 250 g and 85 g mango groups at Week 16, with recruitment of at least 15 subjects in each group with alpha set to p = 0.05. Statistical analyses were performed

Skin Carotenoids
Skin carotenoids (SCs), units in mm wavelength, were measured in the right index finger after cleaning with alcohol by reflection spectroscopy (Veggie Meter ® , Longevity Link Corporation, Salt Lake City, UT, USA). The device has been validated to correspond with plasma carotenoid levels and was calibrated before each use [21].

Blood Pressure and Lipids
Three blood pressure readings were obtained, five minutes apart after a fifteen-minute seated rest using an automated oscillometric unit and averaged to obtain systolic (SBP) and diastolic (DBP) values (Vital Spot, VSM 300, Welch Allyn, Skaneateles Falls, NY, USA). Plasma lipids were analyzed for cholesterol, low-density lipoprotein (LDL), non-high-density lipoprotein (HDL), and triglycerides at the UC Davis Department of Pathology and Laboratory Medicine.

Dietary Intake
A 24-h recall was conducted at each study visit using the validated Automated Self-Administered 24-h (ASA24) dietary assessment tool (https://epi.grants.cancer.gov/asa24). A compliance log was maintained, showing the date, time, and format of mango consumption.

Statistical Analysis
An a priori power analysis showed that there was greater than 80% power to detect a 10% difference in wrinkle severity between the 250 g and 85 g mango groups at Week 16, with recruitment of at least 15 subjects in each group with alpha set to p = 0.05. Statistical analyses were performed with JMP version 15 (SAS Institute Inc., Cary, NC, USA). p values of 0.05 or less were considered statistically significant. Each parameter was assessed for normality, or transformed (Log, Log10, or Johnson) to achieve normality before analyses. Data are presented as mean ± SD. Baseline participant characteristics were analyzed with the t-test or Wilcoxon signed-rank test as appropriate. All other outcome measures were analyzed using the Fit Model to perform Two Way ANCOVA, with treatment and time as factors, and BMI as a covariate. Post-hoc analyses were performed with Contrast tests. The Pearson and Spearman correlation coefficients were used to analyze the correlation between outcome measures that were normally and non-normally distributed, respectively.

Baseline Characteristics
Thirty-six healthy postmenopausal women were enrolled. Thirty-two individuals completed the study (Figure 2). Analysis of skin carotenoids, blood pressure, and dietary records included all participants. For the wrinkle analysis, four sets of data were removed due to technical errors in image capture. Those that showed no deep wrinkles at baseline were removed from the analysis of both deep and the average wrinkle score (which used deep wrinkle values), resulting in data from eight and nine participants for the left lateral canthus, and ten and eight for the right lateral canthus for the 85 g and 250 g groups, respectively. Data for fine and emerging wrinkles were available from 13 to 15 individuals in the 85 g and 250 g groups, respectively. Analysis of plasma lipids and glucose included 21 participants (13 and eight in the 85 g and 250 g group, respectively).

Baseline Characteristics
Thirty-six healthy postmenopausal women were enrolled. Thirty-two individuals completed the study ( Figure 2). Analysis of skin carotenoids, blood pressure, and dietary records included all participants. For the wrinkle analysis, four sets of data were removed due to technical errors in image capture. Those that showed no deep wrinkles at baseline were removed from the analysis of both deep and the average wrinkle score (which used deep wrinkle values), resulting in data from eight and nine participants for the left lateral canthus, and ten and eight for the right lateral canthus for the 85 g and 250 g groups, respectively. Data for fine and emerging wrinkles were available from 13 to 15 individuals in the 85 g and 250 g groups, respectively. Analysis of plasma lipids and glucose included 21 participants (13 and eight in the 85 g and 250 g group, respectively).  At the start of the study, the two groups were similar in age, blood pressure, lipids, and fasting blood glucose (Table 1). Participants had either Fitzpatrick Skin Type II or III (n = 9 and 19, respectively). Baseline BMI values were significantly different, with the mean for those in the 85 g group were classified as overweight, while the average for the 250 g group was in the normal range. Left side measurements were significantly different between the two groups at baseline for fine (FL) and emerging wrinkle length (EL). At the start of the study, the two groups were similar in age, blood pressure, lipids, and fasting blood glucose (Table 1). Participants had either Fitzpatrick Skin Type II or III (n = 9 and 19, respectively). Baseline BMI values were significantly different, with the mean for those in the 85 g group were classified as overweight, while the average for the 250 g group was in the normal range. Left side measurements were significantly different between the two groups at baseline for fine (FL) and emerging wrinkle length (EL). At the start of the study, the two groups were similar in age, blood pressure, lipids, and fasting blood glucose (Table 1). Participants had either Fitzpatrick Skin Type II or III (n = 9 and 19, respectively). Baseline BMI values were significantly different, with the mean for those in the 85 g group were classified as overweight, while the average for the 250 g group was in the normal range. Left side measurements were significantly different between the two groups at baseline for fine (FL) and emerging wrinkle length (EL).  At the start of the study, the two groups were similar in age, blood pressure, lipids, and fasting blood glucose (Table 1). Participants had either Fitzpatrick Skin Type II or III (n = 9 and 19, respectively). Baseline BMI values were significantly different, with the mean for those in the 85 g group were classified as overweight, while the average for the 250 g group was in the normal range. Left side measurements were significantly different between the two groups at baseline for fine (FL) and emerging wrinkle length (EL).  At the start of the study, the two groups were similar in age, blood pressure, lipids, and fasting blood glucose (Table 1). Participants had either Fitzpatrick Skin Type II or III (n = 9 and 19, respectively). Baseline BMI values were significantly different, with the mean for those in the 85 g group were classified as overweight, while the average for the 250 g group was in the normal range. Left side measurements were significantly different between the two groups at baseline for fine (FL) and emerging wrinkle length (EL).  At the start of the study, the two groups were similar in age, blood pressure, lipids, and fasting blood glucose (Table 1). Participants had either Fitzpatrick Skin Type II or III (n = 9 and 19, respectively). Baseline BMI values were significantly different, with the mean for those in the 85 g group were classified as overweight, while the average for the 250 g group was in the normal range. Left side measurements were significantly different between the two groups at baseline for fine (FL) and emerging wrinkle length (EL).  At the start of the study, the two groups were similar in age, blood pressure, lipids, and fasting blood glucose (Table 1). Participants had either Fitzpatrick Skin Type II or III (n = 9 and 19, respectively). Baseline BMI values were significantly different, with the mean for those in the 85 g group were classified as overweight, while the average for the 250 g group was in the normal range. Left side measurements were significantly different between the two groups at baseline for fine (FL) and emerging wrinkle length (EL).  At the start of the study, the two groups were similar in age, blood pressure, lipids, and fasting blood glucose (Table 1). Participants had either Fitzpatrick Skin Type II or III (n = 9 and 19, respectively). Baseline BMI values were significantly different, with the mean for those in the 85 g group were classified as overweight, while the average for the 250 g group was in the normal range. Left side measurements were significantly different between the two groups at baseline for fine (FL) and emerging wrinkle length (EL).  At the start of the study, the two groups were similar in age, blood pressure, lipids, and fasting blood glucose (Table 1). Participants had either Fitzpatrick Skin Type II or III (n = 9 and 19, respectively). Baseline BMI values were significantly different, with the mean for those in the 85 g group were classified as overweight, while the average for the 250 g group was in the normal range. Left side measurements were significantly different between the two groups at baseline for fine (FL) and emerging wrinkle length (EL).  At the start of the study, the two groups were similar in age, blood pressure, lipids, and fasting blood glucose (Table 1). Participants had either Fitzpatrick Skin Type II or III (n = 9 and 19, respectively). Baseline BMI values were significantly different, with the mean for those in the 85 g group were classified as overweight, while the average for the 250 g group was in the normal range. Left side measurements were significantly different between the two groups at baseline for fine (FL) and emerging wrinkle length (EL).  At the start of the study, the two groups were similar in age, blood pressure, lipids, and fasting blood glucose (Table 1). Participants had either Fitzpatrick Skin Type II or III (n = 9 and 19, respectively). Baseline BMI values were significantly different, with the mean for those in the 85 g group were classified as overweight, while the average for the 250 g group was in the normal range. Left side measurements were significantly different between the two groups at baseline for fine (FL) and emerging wrinkle length (EL). ge, blood pressure, lipids, and fasting Skin Type II or III (n = 9 and 19, t, with the mean for those in the 85 g 250 g group was in the normal range. he two groups at baseline for fine (FL) rticipants. Average and deep wrinkles measurements had an n of 8 and 9 in the left, and an n of 10 and 8 in the right, in the 85 g and 250 g group, respectively. ‡ Glucose measurement had an n of 12 and 9 participants for 85 g and 250 g group, respectively. BMI = Body Mass Index; WC = Waist Circumference; SBP = Systolic Blood Pressure; DBP = Diastolic Blood Pressure; HR = Heart Rate; RS = Reflection Spectroscopy; LDL = Low-density Lipoprotein; HDL = High-density Lipoprotein.

Dietary Intake
Both groups consumed approximately 1700 kcals at the baseline, with few significant differences in macronutrients (Table S1); the 85 g group had higher intakes of fiber, folate, and lutein plus zeaxanthin. After 16 weeks, the 85 g group reported increases in dietary potassium and cholesterol, while the 250 g group had a significant increase in total sugars, as well as potassium and folate.

Facial Wrinkles and Erythema
At baseline, no differences were noted in right lateral canthi measures, while left FL and EL were significantly lower in the 250 g compared to the 85 g group (Table 1). In the 85 g group, right deep wrinkle severity decreased by 23% after eight weeks and 20% after 16 weeks (both of which were significant (p = 0.007 and p = 0.03, respectively; Figure 3a). A trend for reduced left deep wrinkle severity (DS) was noted in the 85 g group (severity scores at baseline: 7575 ± 1063 vs. 16 weeks: 5339 ± 3153, p = 0.10). In contrast, an increasing trend in right DS was observed in the 250 g group (severity scores at baseline: 7537 ± 1338 vs. 16 weeks: 8056 ± 982, p = 0.07). Comparison of the two groups showed a trend for lower left DS after 16 weeks in the 85 g group relative to the 250 g group (severity scores for 85 g: 5339 ± 3153 vs. 250 g: 8105 ± 1840, p = 0.08). For the right DS, the 85 g mango group was significantly lower compared to 250 g at both week eight (p = 0.01) and week 16 (p = 0.02; Figure 3a). * Significantly different between 85 g and 250 g groups (p < 0.05). Statistical analysis by t test or † Wilcoxon's signed-rank test. Ŧ Average and deep wrinkles measurements had an n of 8 and 9 in the left, and an n of 10 and 8 in the right, in the 85 g and 250 g group, respectively. ‡ Glucose measurement had an n of 12 and 9 participants for 85 g and 250 g group, respectively. BMI = Body Mass Index; WC = Waist Circumference; SBP = Systolic Blood Pressure; DBP = Diastolic Blood Pressure; HR = Heart Rate; RS = Reflection Spectroscopy; LDL = Low-density Lipoprotein; HDL = High-density Lipoprotein.

Dietary Intake
Both groups consumed approximately 1700 kcals at the baseline, with few significant differences in macronutrients (Table S1); the 85 g group had higher intakes of fiber, folate, and lutein plus zeaxanthin. After 16 weeks, the 85 g group reported increases in dietary potassium and cholesterol, while the 250 g group had a significant increase in total sugars, as well as potassium and folate.

Facial Wrinkles and Erythema
At baseline, no differences were noted in right lateral canthi measures, while left FL and EL were significantly lower in the 250 g compared to the 85 g group (Table 1). In the 85 g group, right deep wrinkle severity decreased by 23% after eight weeks and 20% after 16 weeks (both of which were significant (p = 0.007 and p = 0.03, respectively; Figure 3a). A trend for reduced left deep wrinkle severity (DS) was noted in the 85 g group (severity scores at baseline: 7575 ± 1063 vs. 16 weeks: 5339 ± 3153, p = 0.10). In contrast, an increasing trend in right DS was observed in the 250 g group (severity scores at baseline: 7537 ± 1338 vs. 16 weeks: 8056 ± 982, p = 0.07). Comparison of the two groups showed a trend for lower left DS after 16 weeks in the 85 g group relative to the 250 g group (severity scores for 85 g: 5339 ± 3153 vs. 250 g: 8105 ± 1840, p = 0.08). For the right DS, the 85 g mango group was significantly lower compared to 250 g at both week eight (p = 0.01) and week 16 (p = 0.02; Figure  3a). Deep wrinkle length (DL) in the 85 g group was reduced by 32.7% after 16 weeks compared to baseline (baseline: 17.61 ± 5.05 mm vs. 16 weeks: 11.85 ± 7.56 mm, p = 0.07) that was significantly lower compared to the 250 g group (85 g: 11.85 ± 7.56 mm vs. 250 g: 18.21 ± 5.91 mm, p = 0.02). A betweengroup trend for increased right deep wrinkle width (DW) (85 g: 1.44 ± 0.63 mm vs. 250 g: 1.74 ± 0.20 mm, p = 0.08) and right emerging wrinkle severity (ES) (severity scores at 85 g: 4203 ± 94 vs. 250 g: Deep wrinkle length (DL) in the 85 g group was reduced by 32.7% after 16 weeks compared to baseline (baseline: 17.61 ± 5.05 mm vs. 16 weeks: 11.85 ± 7.56 mm, p = 0.07) that was significantly lower compared to the 250 g group (85 g: 11.85 ± 7.56 mm vs. 250 g: 18.21 ± 5.91 mm, p = 0.02). A between-group trend for increased right deep wrinkle width (DW) (85 g: 1.44 ± 0.63 mm vs. 250 g: 1.74 ± 0.20 mm, p = 0.08) and right emerging wrinkle severity (ES) (severity scores at 85 g: 4203 ± 94 vs. 250 g: 4299 ± 108, p = 0.05) was observed after 16 weeks in the 250 g. These between-group differences in right ES may be attributed to a 2.9% increase from baseline in the 250 g group (p = 0.02; Figure 3b).
A number of left side wrinkle measures significantly increased in the 250 g group, with no significant changes from baseline for the 85 g group. Compared to baseline, left average wrinkle severity (AS) significantly increased by week 8 (p = 0.03), which persisted to week 16 (p = 0.049; Figure 4a). A 25% increase from baseline in left average wrinkle length (AL) was observed (p = 0.007), which was also significantly higher than the value for the 85 g group at week 16 (p = 0.01; Figure 4b).
Left fine wrinkle severity (FS) increased from baseline after 8 weeks (p = 0.048) and after 16 weeks (p = 0.02; Figure 4c). 4299 ± 108, p = 0.05) was observed after 16 weeks in the 250 g. These between-group differences in right ES may be attributed to a 2.9% increase from baseline in the 250 g group (p = 0.02; Figure 3b).
A number of left side wrinkle measures significantly increased in the 250 g group, with no significant changes from baseline for the 85 g group. Compared to baseline, left average wrinkle severity (AS) significantly increased by week 8 (p = 0.03), which persisted to week 16 (p = 0.049; Figure  4a). A 25% increase from baseline in left average wrinkle length (AL) was observed (p = 0.007), which was also significantly higher than the value for the 85 g group at week 16 (p = 0.01; Figure 4b). Left fine wrinkle severity (FS) increased from baseline after 8 weeks (p = 0.048) and after 16 weeks (p = 0.02; Figure 4c). For erythema measures, left cheek erythema was significantly increased after 16 weeks in the 85 g group (degree of intensity % at baseline: 21.2 ± 18.0 vs. 16 weeks: 27.2 ± 16.0, p = 0.04), while no changes were observed in the 250 g group.

Skin Carotenoids
At baseline, SCs were generally lower in the 85 g compared to the 250 g group (85 g: 363 ± 78 mm wavelength vs. 250 g: 432 ± 105 mm wavelength, p = 0.06). Mango intake did not result in a significant change within or between groups over time. While no significant changes in SCs were observed for those who were normal weight in either group, those who were overweight or obese showed a significant increase from baseline after eight (348 ± 59 mm wavelength, p = 0.01) and 16 (352 ± 76 mm wavelength, p = 0.03) weeks, regardless of group assignment. These findings may be due, in part, to the fact that participants who were overweight or obese had significantly lower SCs at baseline compared to those who were of normal weight (329 ± 77 mm wavelength vs. 452 ± 104 mm wavelength, respectively, p = 0.008).

Skin Carotenoids
At baseline, SCs were generally lower in the 85 g compared to the 250 g group (85 g: 363 ± 78 mm wavelength vs. 250 g: 432 ± 105 mm wavelength, p = 0.06). Mango intake did not result in a significant change within or between groups over time. While no significant changes in SCs were observed for those who were normal weight in either group, those who were overweight or obese showed a significant increase from baseline after eight (348 ± 59 mm wavelength, p = 0.01) and 16 (352 ± 76 mm wavelength, p = 0.03) weeks, regardless of group assignment. These findings may be due, in part, to the fact that participants who were overweight or obese had significantly lower SCs at baseline compared to those who were of normal weight (329 ± 77 mm wavelength vs. 452 ± 104 mm wavelength, respectively, p = 0.008). No significant changes were observed in blood glucose with either 85 g or 250 g of mango intake after 16 weeks.

Discussion
In this exploratory trial, we observed that 85 g (0.5 cup) of mango intake for two to four months reduced facial wrinkles while 250 g (1.5 cups) increased them. We report here a significant decrease in right DS and a trend in the reduction for left DS and DL in the 85 g group. In contrast, the 250 g group showed a significant increase in left AS, AL, FS, and right ES and trended towards an increase in right DS and DW. While further research is needed to explore the mechanisms behind these findings, the reduction in wrinkles with 85 g of mango intake may be due to the beneficial effects of carotenoids, flavonoids, and mangiferin, all of which, as part of a whole food complex, could lead to improvements in collagen bundles and a reduction in epidermal thickening as seen with the mouse study noted above [18]. The increase in wrinkles in the 250 g group was notable and unexpected. Since a significant increase in total sugar intake was noted in this group after eight and sixteen weeks, this increased sugar intake may have led to glycation of collagen fibers, thereby disrupting the collagen structure [22]. The effects of whole food intake on skin health in humans is a relatively new area of research. A recent study reported a significant reduction in facial wrinkles in postmenopausal women after regular intake of almonds [23]. Although dietary and skin carotenoids were positively correlated with erythema, a significant increase in erythema was only observed in the 85 g mango group. Therefore, it is unlikely that erythema was caused by the amount of β-carotene consumed from the fruit. Furthermore, β-carotene has been shown to decrease erythema in other studies [9,24].
Favorable changes in markers of cardiovascular risk were noted in the 250 g group. Serum cholesterol significantly decreased by 7.4%, while a trend for reduced LDL and non-HDL cholesterol was observed. This may possibly be attributed to plant sterols that are abundant in mangos (24.4 mg/100 g fruit), as these compounds have been associated with a reduction in LDL cholesterol [25][26][27]. The 250 g of mango intake may also have improved cholesterol levels by providing a significant source of both soluble (28.2 g/100 g dry matter) and insoluble (41.5 g/100 g dry matter) fiber [28,29]. Our data is consistent with the observation of reduced total and LDL cholesterol levels after Ataulfo mango pulp intake in a rat model [30].
We also observed a 4% reduction in blood pressure with higher levels of mango intake. The inverse correlations observed between dietary and skin carotenoids with blood pressure are consistent with findings from other studies that report negative correlations between β-carotene and SBP, as well as with cardiovascular mortality [31,32]. Taken together, these results suggest a potential role of mango intake on cardiometabolic health but require confirmation through future trials powered explicitly for these outcomes.

Limitations
Two levels of mango intake were used in this exploratory study to determine whether either amount would produce a change in facial wrinkles. Thus, a control group consuming no mangos was not employed. However, a strength of the study is that the design allowed for the assessment of an amount-based response and allowed for better control of any placebo effect that could be present as women in both groups received mangos. Finally, the study was limited to healthy postmenopausal women with Fitzpatrick skin type II and III; therefore, the findings reported here may not be generalizable to other groups.

Conclusions
Results from this pilot study support the concept that regular intake of modest amounts of mangos may improve facial wrinkles. The apparent beneficial effects of mangos on skin health may be lost if the intake of mangos is particularly high. The effects of whole food intake on skin health are limited but promising. Further prospective studies are warranted.
Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6643/12/11/3381/s1, Table S1: 24-h reported dietary intake. Funding: This study was supported in part by a grant from the National Mango Board (NMB), grant number A18-2693, which also supplied the fresh mangos for the study. The NMB had no role in the study design, data collection, data analysis, manuscript preparation, or publication decision.