Carotenoids, especially -carotene, have attracted recent attention due to numerous reports of an inverse relationship between plasma carotenoid level, or dietary carotenoid intake, and risk of certain chronic diseases (Mares-Perlman et al. 1995, Seddon et al. 1994, Ziegler 1989). Because of the potential biological effects of carotenoids, a better understanding of the physiological factors influencing plasma carotenoid levels and tissue deposition in humans is required. Body composition may be one such factor.

Carotenoids are lipophilic substances and become deposited in adipose tissue and liver (Parker 1989). It is presently not known whether adipose tissue carotenoids exchange with the plasma carotenoid pool and, if so, at what rate of exchange. If uptake is rapid but release relatively slow, adipose tissue may represent a sink for carotenoids, at least over relatively short periods of time. However, carotenoids exist in many non-adipose tissues in humans (Schmitz et al. 1991), probably taken up via LDL internalization. Low density lipoprotein is the primary plasma transporter of hydrocarbon carotenoids (- and -carotene, lycopene), whereas HDL is an important carrier of the more polar carotenoids such as lutein and cryptoxanthin (Clevidence and Bieri 1993). Limited research on the relationship between body composition and steady-state plasma carotenoid concentrations has shown that plasma -carotene levels are inversely related to body mass index (BMI)⁶ (Ringer et al. 1991, Rock and Swendseid 1993). However, BMI is a relatively inaccurate measure of adiposity, because it does not distinguish between body mass contributed by fat and lean body mass. More specific measurements of body composition, such as fat-free mass (FFM) and fat mass (FM), are warranted to partition the effects of different body compartments on plasma carotenoid concentrations. Fat mass represents the total mass of fatty tissues in the body (predominantly adipose tissue), whereas FFM represents the combined mass of non-fatty tissues such as muscle and bone and organs such as the liver and kidney. Specific compositional compartments may modulate either steady-state plasma concentrations or the magnitude of changes in concentration in response to either carotenoid supplementation or dietary depletion. Plasma -carotene response to long-term supplementation has been reported to be inversely related to BMI (Constantino et al. 1988), whereas reports of the effect of adiposity (estimated by bioelectrical impedance or anthropometry) have been inconsistent (Henderson et al. 1989, Johnson and Russell 1992, Rock and Swendseid 1993, Sugerman et al. 1991). At the present time, there are no published reports concerning the comparative effects of FM and FFM on either steady-state plasma carotenoid levels or on changes in plasma carotenoid concentrations in response to supplementation or depletion.

We selected the five most prevalent plasma carotenoids (-carotene, -carotene, lutein, lycopene and -cryptoxanthin) to examine the relationship between plasma carotenoid concentrations and body composition. Together these carotenoids constitute about 75-80% of the total carotenoid concentration in the plasma as determined by HPLC (Cooney et al. 1995, Micozzi et al. 1992). Food values of these five carotenoids are now available via the USDA-NCI Carotenoid Database (Mangels et al. 1993), and these carotenoids represent a range of structural classes and polarities. Parker (1993) reported that adipose tissue contains higher concentrations of some polar carotenoids relative to their concentrations in plasma. The differential transport of polar and nonpolar carotenoids in plasma lipoproteins (Clevidence and Bieri 1993) indicates a possible structure-specific or tissue-specific nature of carotenoid deposition and the importance of examining a diverse group of carotenoids.

The aims of the present study were to investigate 1) the relationship between steady-state carotenoid concentration and body composition of healthy older men as measured by hydrodensitometry and 2) the relationship between body composition and changes in plasma carotenoid concentration in response to carotenoid depletion or -carotene supplementation.

MATERIALS AND METHODS

All subjects were supplied with a standard diet for 28 d after the baseline blood sampling was conducted. This standard diet was similar to a low carotenoid diet used in a previous study (Fuller et al. 1993) except that it included one daily serving of carrots. On average, this diet provided 15 µg/d of -carotene from the basal diet as analyzed by HPLC (Fuller et al. 1993) and 1.5 mg/d of -carotene from the carrot serving as estimated from the 1993 USDA-NCI Carotenoid Food Composition Database, Version I (Mangels et al. 1993). Thus, total -carotene intake from the standard diet was about 1.51 mg/d, or only about 5% of the supplement level. This diet supplied roughly 13% of total energy from protein, 54% from carbohydrate and 33% from fat. Because the fat content of the diet was kept constant during the study, the effect of meal fat intake on -carotene absorption was controlled. Food-energy intake was adjusted at least weekly for individual weight maintenance, based on the initial weight of each individual and weight changes reported in the diaries. Initial and final body weights were determined in the laboratory.

Subjects came to the laboratory every 2 or 3 d to pick up food and were instructed to fill out diaries to monitor daily compliance and weight change. They were also given a list of prohibited food items and a list of food items that were not provided but permitted if recorded in the diary. Diaries were routinely checked at each visit.

Table 1. Age, body size and body composition of the subjects1 [View Table]

The effects of body size and composition on the changes in plasma carotenoid levels caused by treatment (increase in plasma -carotene in the supplemented group or decrease in the placebo group) were tested first by simple linear regression with one of the following as the independent variable: weight, FFM, FM, BMI or % BF. The effects of FFM, FM and BMI on post-treatment plasma carotenoid levels were further tested by multiple regression analysis (Kleinbaum et al. 1988) in models that included baseline concentration and treatment group as covariates. A P value of less than 0.05 was considered statistically significant. All analyses were performed using MINITAB software (MINITAB version 10, Minitab, State College, PA).

RESULTS

At baseline (pre-treatment), there were no significant differences between the two groups in plasma concentrations of either individual carotenoids or total carotenoids (SUM5) (Table 2). In the placebo group, there were significant decreases from pre-treatment to post-treatment in plasma concentrations of -carotene, lutein, lycopene and -cryptoxanthin and in SUM5 (P < 0.05). Concentrations of -carotene did not change significantly. This reflects the relative lack of carotenoids in the basal diet with the exception of -carotene and -carotene in the daily carrot serving. In the -carotene-supplemented group, there was a large and significant increase in plasma -carotene concentration, and there were significant decreases in the concentrations of lutein, lycopene and -cryptoxanthin, carotenoids lacking in the basal diet. As in the placebo group, no significant change was observed in -carotene concentration. The increase in plasma -carotene concentration in the supplemented group ranged from 1.5- to 29-fold among subjects.

Table 2. Plasma carotenoid concentrations of placebo and -carotene-supplemented men before and after treatment1 [View Table]

Baseline total plasma carotenoid concentration was significantly and negatively correlated with FFM but not with FM (Table 3). Consistent with previous reports, total carotenoid concentration was also negatively correlated with both BMI and BW (P < 0.05). Baseline plasma -carotene concentration exhibited a significant and negative correlation with BMI and BW and with both FFM and FM (P < 0.05).

Table 3. Correlation coefficients between plasma carotenoid concentrations and body size and body composition at baseline in 31 men1 [View Table]

In the -carotene-supplemented group, the magnitude of increase in plasma -carotene and total carotenoid concentrations was significantly and negatively associated with body weight, BMI and FFM but not with FM (Table 4). Thus, subjects with higher body mass or higher FFM exhibited smaller increases in plasma -carotene. In the placebo group, the decrease in total carotenoid concentration was significantly positively associated with BMI, body weight and FFM (P < 0.05) but not with FM. Therefore, subjects with greater FFM exhibited smaller (less negative) decreases in plasma total carotenoid concentration accompanying dietary carotenoid deprivation. The decline in lycopene concentration was positively related to body weight, FM and FFM, and the decrease in -carotene concentration was significantly and positively associated with both body weight and BMI. Notably, in neither groups were decreases in the polar carotenoids, lutein and -cryptoxanthin, significantly correlated with any of the body composition variables measured.

Table 4. Correlation coefficients between changes in plasma carotenoid concentrations in response to -carotene or placebo treatment and body size and composition in men1 [View Table]

Figure 1 illustrates the relationship between FFM and the magnitude of change in plasma total carotenoids for the -carotene-supplemented group (panel A, r = 0.64, P < 0.05) and for the placebo group (panel B, r = 0.57, P < 0.05). With both treatments, higher FFM was associated with smaller changes in plasma total carotenoids.

When results for all subjects were pooled, multiple linear regression analysis revealed that FFM and BMI were significantly and negatively associated with post-treatment -carotene and total carotenoid concentration, after controlling for baseline level and treatment group (Tables 5 and 6). Neither FM nor % BF were associated with these carotenoid concentrations. There was a significant and negative interaction between the effects of treatment group and FFM on the final concentrations, indicating that the direction of effect of FFM (positive or negative) on final -carotene or total carotenoid concentration depended on treatment (i.e., supplementation or depletion). This result was consistent with that of the simple regression analyses (Fig. 1), i.e., the effect was positive in the placebo group and negative in the supplement group (P < 0.02). Both analyses reflect the finding that individuals with greater FFM tended to exhibit smaller increments or decrements in plasma carotenoids in response to -carotene supplementation or dietary carotenoid deprivation, respectively.

Because age has occasionally been reported to be associated with plasma carotenoid concentrations (Brady et al. 1996, Ringer et al. 1991) or response to single dose supplementation (Sugerman et al. 1991), the effect of age on baseline total carotenoids, and on change in total carotenoid or -carotene concentration, was examined. However, none of these carotenoid concentration variables was significantly associated with age in this cohort, a finding consistent with one multiple-dose study (Constantino et al. 1988) and two single-dose studies (Johnson and Russell 1992, Rock and Swendseid 1993).

DISCUSSION

However, BMI is a measurement of total body mass relative to height and does not distinguish between the contribution of FM and lean body mass to total body mass. Using hydrodensitometry to more accurately assess body composition, we were able to examine the individual contributions of these two body compartments. No such studies have been reported previously. Fat mass is reflective of the total mass of adipose tissue, and FFM reflects the total mass of non-fatty tissues in the body. In the current study, plasma -carotene concentration at baseline was found to be inversely related to both FM and FFM, whereas the baseline concentration of total carotenoids was inversely related to FFM only. This suggests an important role of non-adipose tissues in determining steady-state plasma carotenoid levels resulting from usual diets. Such a role is supported by the data of Rock and Swendseid (1993), who reported elevated plasma -carotene levels in patients with anorexia nervosa but not in obese subjects who had recently lost weight, relative to -carotene levels of normal controls. The differences in plasma -carotene could not be attributed to differences in dietary carotenoid intake. Thus modulation of both FM and FFM, which generally occur together in anorectic subjects, resulted in marked changes in carotenoid status, whereas changes in FM alone (obese subjects) apparently had no such effect. Although plasma -carotene concentrations among the entire cohort were negatively correlated with BMI (r = 0.59), there was no relationship with triceps skinfold thickness.

The extent of increase in plasma -carotene concentration in response to -carotene supplementation is highly variable among individuals (Dimitrov et al. 1988). Although the reasons for this heterogeneity are unclear and likely to be multiple, body composition may represent one such factor. In the current study, 28 d of -carotene supplementation resulted in substantial but variable increases in plasma -carotene concentration. The magnitude of increase in plasma -carotene concentrations was significantly and inversely related to BMI. This result is consistent with that of Constantino et al. (1988), who studied elderly men supplemented with -carotene for 10 mo. In the current study, examination of the differential effects of FFM and FM showed that the increase in -carotene was significantly and negatively correlated with FFM (r = 0.64, P < 0.05) but not with FM. That is, subjects with greater FFM had lower plasma -carotene concentrations following the supplementation period, whereas FM showed no such effect. This suggests that non-adipose tissues may play a more active role than adipose tissue in the disposition of newly absorbed -carotene during the non-steady-state conditions that prevail during the first few weeks of supplementation.

Table 5. Multiple regression analysis of the effects of fat-free mass (FFM) or fat mass (FM) on post-treatment plasma -carotene concentration in men1 [View Table]

Table 6. Multiple regression analysis of the effects of fat-free mass (FFM) or fat mass (FM) on post-treatment plasma total carotenoid concentration in men1 [View Table]

The current study is the first to investigate the relationship between adiposity and plasma carotenoid concentrations in response to regular -carotene supplementation. Previous reports utilized single-dose protocols. We did not observe a significant association between adiposity (either percentage of body fat or FM) and plasma -carotene response. This finding is in contrast to that of Sugerman et al. (1991), who reported that the percentage of body fat, as estimated by bioelectrical impedance, was inversely correlated with the single-dose plasma -carotene response in a subject pool that included both young and elderly men. Our finding of a lack of effect of adiposity, however, is consistent with three other single-dose studies (Henderson et al. 1989, Johnson and Russell 1992, Rock and Swendseid 1993). In the latter, the plasma -carotene response to a single 50-mg dose was similar between non-obese and obese subjects, despite greatly elevated BMI, body weight and tripceps skinfold thickness in the latter. In contrast, anoretic subjects exhibited a 30-h plasma -carotene response nearly double that of normal or obese subjects, but because of wide variation in response within all three groups, the mean increments were not significantly different.

Fat-free mass includes the mass of bone, muscle, and organs such as liver, kidney and adrenal. The adrenal generally exhibits the highest carotenoid concentrations in humans (Raica et al. 1972), which probably results from the high rate of LDL uptake by this organ, a characteristic also shared by several other non-adipose tissues such as liver and kidney. In rats, these three organs exhibit receptor-mediated LDL clearance rates approximately 200, 100 and 28 times greater, respectively, than adipose tissue (Spady et al. 1985). Thus, FFM, in contrast to FM, may serve as a dynamic reservoir that actively takes up lipoprotein-associated -carotene from plasma. In individuals with greater FFM, a larger proportion of newly absorbed -carotene would enter this reservoir, resulting in a dimished increment in plasma -carotene concentration. This hypothesis could be tested by standardizing the -carotene dose for FFM, in which case the plasma -carotene response should be more similar among subjects, regardless of FM.

Body composition might also be expected to influence the rate of decline of plasma carotenoids resulting from dietary carotenoid deprivation. There have been no published reports concerning this issue. Both placebo and -carotene supplement groups exhibited similar decreases in plasma lutein, lycopene and cryptoxanthin concentrations, indicating that -carotene supplementation had no effect on the elimination of these carotenoids. Placebo group subjects with higher body weight, BMI or FFM exhibited smaller decrements in plasma total carotenoid concentration. This relationship was not significant for FM. Thus BMI and FFM exhibited the same muting effect on changes in plasma carotenoids during depletion as were observed during -carotene supplementation. This finding is consistent with the hypothesis that FFM serves as a dynamic reservoir for carotenoids, because individuals with greater FFM would possess a greater exchangeable pool and thus exhibit smaller decrements in plasma carotenoids, at least over some initial period of time. Assuming adipose tissue carotenoids exchange with the plasma pool at a much slower rate than FFM carotenoids, as is the case for tocopherols (Bieri 1972, Machlin et al. 1979), adipose tissue stores may affect plasma carotenoid concentration only after prolonged periods. Our finding of an effect of FM on baseline (steady-state) -carotene concentration, but not on shorter-term changes in concentration, is consistent with this notion.

The relationship between body composition and carotenoid disappearance seemed to involve primarily the hydrocarbon (LDL-associated) carotenoids, because among specific carotenoids this relationship was significant only between -carotene and BMI (and body weight) and between lycopene and FFM (and body weight). The FFM-lycopene relationship was not observed in the -carotene group, even though both groups exhibited similar decrements in plasma lycopene, a finding for which we have no explanation at present.

Two previous reports concerning the rate of decline in plasma -carotene concentration, one from baseline levels (Rock et al. 1992) and one following cessation of supplementation (Micozzi et al. 1992), indicate a biphasic elimination curve, suggestive of two body pools with different turnover rates. It was suggested that one such pool is rapidly responsive to recent intake, whereas the other influences long-term elimination (Rock et al. 1992). Whether FFM and FM may represent these pools is speculative at the present time. Novotny et al. (1995) proposed two hepatic -carotene pools but only a single total body irreversible loss component, based on a single-dose compartment model in a single subject. Additional studies of this type are needed.

In summary, fat-free body mass was found to be a significant determinant of 1) steady-state plasma carotenoid concentration, 2) the extent of increase in plasma -carotene concentrations in response to 4 wk of -carotene supplementation and 3) the extent of decline in plasma total carotenoid concentration during 4 wk of dietary carotenoid restriction. Because this study was conducted in older men, few of which were particularly obese, an important role for adipose tissue in other subject populations cannot be excluded. However, the results strongly support a role for body composition, and non-adipose tissues in particular, in influencing plasma carotenoid concentrations during chronic dietary or supplement interventions.