The Effect of Age-related Differences in Body Size and Composition on Cardiovascular Determinants of VO₂max

Abstract

Background.

A reduction in maximal stroke volume (SV_max) and total blood volume (TBV) has been hypothesized to contribute to the decline in maximal oxygen uptake (VO₂max) with healthy aging. However, these variables have rarely been collected simultaneously in a board age range to support or refute this hypothesis. It is also unclear to what extent scaling size-related cardiovascular determinants of VO₂max affects the interpretation of age-related differences.

Methods.

A retrospective analysis of VO₂max, maximal cardiac output (Q_Cmax), TBV, and body composition including fat-free mass (FFM) in 95 (51% M) healthy adults ranging from 19–86 years.

Results.

Absolute and indexed VO₂max, Q_Cmax, and maximal heart rate decreased in both sexes with age (p ≤ .031). SV_max declined with age when scaled to total body mass or body surface area (p ≤ .047) but not when expressed in absolute levels (p = .120) or relative to FFM (p = .464). Absolute and indexed TBVs (mL/kg; mL/m²) were not significantly affected by age but increased with age in both sexes when scaled to FFM (p ≤ .013). A lower arteriovenous oxygen difference (a-vO₂diff) contributed to the reduction in VO₂max with age in treadmill exercisers (p = .004) but not in the entire cohort (p = .128).

Conclusion.

These results suggest (a) a reduction in absolute SV_max, and TBV do not contribute substantially to the age-related reduction in VO₂max, which instead results from a smaller Q_Cmax due to a lower maximal heart rate, and (b) body composition scaling methods should be used to accurately describe the effect of aging on physical function and cardiovascular variables.

MAXIMAL exercise capacity declines with healthy aging as evident by a reduction in maximal oxygen uptake (VO₂max) (1–6). Cross-sectional studies (3,7) have shown that a reduction in maximal cardiac output (Q_Cmax) significantly contributes to the age-related decline in VO₂max in healthy individuals. Although it is clearly established that maximal heart rate declines with age (3,7,8), previous findings regarding the effect of healthy aging on maximal stroke volume (SV_max) are conflicting (3,7,9–12).

Stroke volume during exercise is significantly influenced by total blood volume (TBV) via left ventricular end-diastolic volume and thus the Frank–Starling mechanism (13,14). Small sample studies have reported that TBV is reduced in both sexes with healthy aging as a result of a smaller plasma and erythrocyte volume (15–17). Consequently, it has been hypothesized that a reduction in TBV may contribute to a smaller SV_max with advancing age (15,17,18). However, VO₂max, TBV, and SV_max have rarely been collected simultaneously in healthy aging individuals to support or refute this hypothesis.

Confounding the interpretation of these data is the fact that cardiovascular anatomic and physiologic variables increase with body size (19,20). Accordingly, SV_max and TBV are frequently scaled to total body mass or body surface area (BSA) to compare populations of different body sizes (7,15–17,20,21). However, given that oxygen uptake during exercise is driven by metabolically active tissue (22), scaling cardiovascular determinants of oxygen transport to metabolically active tissue (fat-free mass, FFM) may provide a more precise assessment of cardiovascular capacity in populations of different body sizes and composition.

Therefore, the purpose of this study was twofold: (a) to determine whether the age-related decline in VO₂max is associated with a reduction in SV_max and TBV and (b) to examine to what extent scaling size-related cardiovascular determinants of VO₂max affects the interpretation of age-related differences. Given the age-associated increase in total body mass and adiposity, we hypothesized that SV_max and TBV would be reduced with age when scaled to total body mass and BSA but preserved when scaled to FFM.

Methods

Subjects were recruited within the Dallas-Fort Worth metroplex. All subjects were nonsmokers, were not taking any cardiovascular medications, had a 24-hour ambulatory blood pressure less than 140/90 mmHg, body mass index ≤ 30kg/m², and a normal ECG and exercise stress echocardiogram. No subjects were performing exercise for more than 90min/wk, with the majority (n = 92) performing less than 20min/d, less than 3x/wk. All subjects signed an informed consent form approved by the Institutional Review Boards of the University of Texas Southwestern Medical Center at Dallas and Presbyterian Hospital.

Maximal Oxygen Uptake and Q_Cmax

VO₂max was determined with an incremental cycling (n = 24) or treadmill (n = 71) exercise test to maximal exertion using the Douglas bag method as previously described elsewhere (23). Gas fractions were analyzed using mass spectrometry and ventilatory volumes by a Tissot spirometer. VO₂ and Q_C was assessed during seated or standing rest, one level of steady-state submaximal exercise corresponding to approximately ≈60% of maximal heart rate, and during maximal exercise. VO₂max was defined as the highest oxygen uptake measured from at least a 30s Douglas bag. Q_C was measured by the C₂H₂ rebreathing method, which has been previously validated in this laboratory (24). Heart rate was measured via a 12-lead ECG, and SV_max was calculated as Q_Cmax/maximal heart rate. Arteriovenous oxygen difference (a-vO₂diff) was calculated by rearranging the Fick equation (a-vO₂diff = VO₂/Q_C). Typical error of measurement expressed as a coefficient of variation (%) for test–retest reproducibility (25) for VO₂max, Q_Cmax, maximal heart rate, and SV_max for 56 subjects in this cohort were 4.8%, 7.6%, 2.7%, and 7.8%, respectively.

Body Composition

Body density and composition were determined by underwater weighing with correction for residual lung volume. Subjects were first weighed for total body mass in air. In the water, subjects were instructed to exhale normally to expel air from their lungs (to functional residual capacity) and then seat themselves in tucked position in an aluminum cage that was suspended in a tank of water (temperature ≈30±1°C). Residual lung volume was measured by the nitrogen washout method with a mass spectrometer. Each subject performed at least three adequate measurements defined as a definite plateau in underwater weight, and the mean value was calculated. FFM and fat content were calculated accordingly.

Total Blood Volume

On a separate day, TBV was measured using a carbon monoxide rebreathing method modified from that described by Burge and Skinner (26) and has been described in detail elsewhere (27). Typical error of measurement expressed as a coefficient of variation (%) for test–retest reproducibility for hemoglobin mass, which is used to determine the distribution of carbon monoxide, is less than 4% for repeated measures in our laboratory (27).

Physical Activity Levels

Daily energy expenditure (kcal) was assessed by a physical activity monitor (Acitical, Phillips, USA; RT3; Stayhealthy, USA) worn from 1–7 days.

Data Analysis

All statistical analysis was performed using SigmaStat version 11.0 (Systat Software, USA).

T-tests were used to determine sex differences in subject characteristics. Univariate correlations (Pearson’s) were used to determine the relationship between cardiovascular, metabolic and body size, and composition variables. Body size and composition measures, VO₂max, TBV and its compartments (plasma volume [PV], erythrocyte volume [ECV]), Q_Cmax, and SV_max were regressed on age by least squares linear regression analysis in all subjects and within each sex. Twenty-four subjects performed a maximal cycle test rather than treadmill test. As VO₂max and SV_max are ≈5%–10% lower during maximal cycle compared with treadmill exercise (28), linear regression analysis was repeated in 71 subjects (63% F, 55±13 years) who performed treadmill exercise testing only. The value of p < .05 was considered statistically significant. Data are presented as mean ± standard deviation (SD) in Tables 1–3.

Table 1.

Participants Characteristics

	Men	Women	All Subjects
Subjects, n	48	47	95
Age, y	46±17	53±15*	50±16
	(19–86)	(21–75)
Height, cm	176±8	164±6*	170±10
	(161–194)	(146–175)
Total body mass, kg	79±11	65±9*	72±12
	(58–99)	(48–90)
BSA, m2	1.96±0.17	1.72±0.14*	1.85±0.20
	(1.64–2.26)	(1.41–2.06)
BMI, kg/m2	25±3	24±3*	25±3
	(18–30)	(19–31)
Fat free mass, kg	58±7	42±5*	50±10
	(43–72)	(28–52)
Fat percent, %	27±7	36±6*	31±8
	(7–40)	(23–47)
Daily energy expenditure, kcal/d	2479±393	1964±352	2234±454
Daily energy expenditure, kcal/kg/d	31±3	30±4*	31±4
Daily energy expenditure, kcal/kgFFM/d	43±5	47±7*	45±6

Notes. BSA = body surface area; BMI = body mass index; FFM = free-fat mass.

Values are expressed as mean ± standard deviation. Numbers in parentheses represents range of values.

*p < .05 versus men.

Table 3.

Maximal Hemodynamic and Metabolic Parameters and TBV

	Men	Women	All Subjects
VO2max, mL/min	2479±458	1539±280*	2014±606
	(1294–3602)	(948–2145)
VO2max, mL/kg/min	32±6	24±4*	28±6
	(17–45)	(14–32)
VO2max, mL/kgFFM/min	43±6	37±5*	40±6
	(27–60)	(23–46)
SVmax, mL	96±18	65±12*	81±22
	(59–153)	(45–87)
SVmax, mL/kg	1.23±0.24	1.00±0.17*	1.12±0.23
	(0.84–1.86)	(0.75–1.37)
SVmax, mL/m2	49±8	38±6*	43±9
	(34–68)	(28–50)
SVmax, mL/kgFFM	1.67±0.25	1.57±0.26#	1.62±0.26
	(1.18–2.28)	(1.11–2.26)
TBV, mL	5503±23	4044±490*	4816±1002
	(4006–7613)	(2730–5426)
TBV, mL/kg	70±8	63±7*	66±9
	(58–101)	(49–83)
TBV, mL/m2	2796±291	2357±216*	2579±337
	(2261–3872)	(1860–3053)
TBV, mL/kgFFM	98±9	96±11	97±10
	(83–118)	(77–128)

Notes. VO₂ = oxygen uptake; SV = stroke volume; TBV = total blood volume; FFM = free-fat mass.

Values are expressed as mean ± standard deviation. Numbers in parentheses represents range of values.

*p < .05 versus men, ^# p = .051.

Results

Subject Characteristics

The men were younger, heavier, taller, leaner, and had more FFM compared with the women (p ≤ .047). Absolute and indexed levels (kcal/kg/d) of daily energy expenditure was greater in men versus women (p ≤ .060), but this result was reversed when scaled to FFM (p = .002; Table 1). At rest, absolute and indexed VO₂ (mL/kg/min) were greater in men compared with women (p ≤ .008), but this difference was eliminated when scaled to FFM (mL/kgFFM/min; p = .672). Likewise, resting stroke volume was smaller in the women (p ≤ .047), unless FFM was used as a scaling index (p = .775). Heart rate and a-vO₂diff were not different between the sexes (p = .544 and p = .352, respectively; Table 2).

Table 2.

Resting Hemodynamic and Metabolic Parameters

	Men	Women	All Subjects
VO2, mL/min	280±59	207±37*	244±61
VO2, mL/kg/min	4±1	3±1*	3.4±0.7
VO2, mL/kgFFM/min	5±1	5±1	5±1
SV, mL	52±12	38±8*	45±12
SV, mL/kg	0.68±0.20	0.59±0.13*	0.63±0.17
SV, mL/m2	27±7	22±5*	25±6
SV, mL/kgFFM	0.92±0.21	0.93±0.20	0.92±0.20
Heart rate, bpm	84±11	89±16	87±14
a-vO2diff, mL/100 mL	7±1	6±1	6±1

Notes. SV = stroke volume; MAP = mean arterial pressure; a-vO₂diff = arteriovenous oxygen difference.

Values are expressed as mean ± standard deviation.

*p < .05 versus men.

VO₂max ranged from 948 to 2,145mL/min in the women and 1,294 to 3,602mL/min in the men, with the mean value being greater in the men compared with the women (2479±458 versus 1539±280mL/min, p < .001; Table 3). Likewise, VO₂max relative to total body mass or FFM was greater in the men compared with the women (p < .001; Table 3). Irrespective of scaling approach, SV_max was also greater in the men compared with the women (p ≤ .051). Absolute TBV was larger in the men compared with the women (p < .001; Table 3). Although TBV remained larger in the men when scaled to total body mass and BSA (p < .001), sex differences were eliminated when scaled to FFM (p = .356). There was a positive and significant relationship between metabolic and cardiovascular variables and total body mass, BSA and FFM, but these relationships were negative for body fat percentage (Table 4).

Table 4.

Univariate Correlations Between Metabolic, Hemodynamic and Circulating Factors and Body Size and Composition Variables

	Pooled Data for Men and Women (n = 95)
	VO2max, L/min	SVmax, mL	TBV, L	Maximum a-vO2diff, mL/100mL
Total body mass, kg	0.69*	0.63*	0.79*	0.25*
BSA, m2	0.74*	0.69*	0.84*	0.24*
FFM, kg	0.88*	0.80*	0.86*	0.09
Fat mass, %	−0.54*	−0.49*	−0.37*	0.25*
	Pooled Data for Treadmill Exercisers (n = 71)
	VO2max, L/min	SVmax, mL	TBV, L	Maximum a-vO2diff, mL/100mL
Total body mass, kg	0.76*	0.73*	0.80*	0.38*
BSA, m2	0.79*	0.77*	0.84*	0.39*
FFM, kg	0.90*	0.80*	0.87*	0.47*
Fat mass, %	−0.49*	−0.35*	−0.38*	−0.28*

Notes. VO₂, oxygen uptake; SV; stroke volume; TBV = total blood volume; a-vO2diff = arteriovenous oxygen difference; BSA = body surface area; FFM = free-fat mass.

*p < .05.

Daily Energy Expenditure and Total Body Mass and Composition With Aging

Although daily energy expenditure (kcal/d) was not significantly affected by age (p = .610), it declined when expressed relative to total body mass (p = .049) but increased relative to FFM (p = .025). Total body mass and BMI were not significantly affected by aging (p = .659 and p = .063, respectively); however, fat mass whether expressed in total mass or percentage increased in both sexes with aging (p ≤ .028). For the entire cohort, FFM declined with aging (p = .027); however, this effect was not significant when men and women were analyzed separately (p = .545 and p = .328, respectively).

Resting and Submaximal Steady-State Exercise VO₂ and Hemodynamics With Aging

Irrespective of unit, resting VO₂ declined with advancing age (p ≤ .039). Resting stroke volume scaled to total body mass and size (mL/kg; mL/m²) appeared to decline with aging (p ≤ .006) but not when scaled relative to FFM (p = .334). Heart rate declined (p = .001), whereas a-vO₂diff was not affected by age (p = .259).

Heart rate (% of maximum) during steady-state exercise was not affected by age (p = .454), suggesting a similar relative level of exercise intensity across the ages. Absolute Q_C was reduced with age, (p = .008); however, stroke volume was unaffected irrespective of scaling approach (p ≥ .090; Figure 1).

Figure 1.

Absolute Q_c declined with age during steady-state exercise, but stroke volume (SV) was unaffected irrespective of scaling unit. Men (filled triangles) and women (filled circles). Overall regression line for men (solid line), women (dotted line), and all subjects (solid bold line). *p < .05 regression line against age. The linear regression equation coefficient (r) and p value are for n = 95.

VO₂max and Maximal Exercise Hemodynamics With Aging

Absolute and indexed VO₂max declined with age (p ≤ .020; Figure 2). Likewise, absolute Q_Cmax declined with age (p < .001), which was independent of body size or composition, as the age-related reduction in Q_Cmax was still evident relative to total body mass, BSA, and FFM (p ≤ .002; Figure 2). Although the reduction in Q_Cmax was robust in men regardless of scaling method (p ≤ .019), it only reached statistical significance in women when relative to BSA (p = .015) and total body mass (p = .004), underscoring the influence of fat mass in this calculation (Figure 2). Maximal heart rate declined in both sexes (7 beats/decade in women; 5 beats/decade in men; p < .001). SV_max declined with age when scaled to total body mass or BSA (p ≤ .047) but not when expressed in absolute levels (p = .120) or relative to FFM (p = .464; Figure 2). Maximal a-vO₂diff was not significantly affected by age (p = .128).

Figure 2.

VO₂max and Q_cmax declined with age. For the entire cohort, SV_max indexed to total body mass declined with age but not when scaled to body surface area or fat-free mass. Men (filled triangles) and women (filled circles). Overall regression line for men (solid line), women (dotted line), and all subjects (solid bold line). *p < .05 regression line against age. The linear regression equation coefficient (r) and p value are for n = 95.

Central Circulatory Factors With Aging

Absolute TBV increased with age in the men (p = .025) but not in the women (p = .452; Figure 3). When scaled to total body mass, the slope of TBV was essentially flat but positive when expressed relative to FFM in both sexes (p ≤ .013; Figure 3), further confirming that TBV is not reduced with age. There was a positive slope with age for PV and ECV in both sexes when indexed to FFM (p < .001; Figure 3).

Figure 3.

Total blood volume was preserved with age in both sexes but increased relative to FFM in both sexes. Men (filled triangles) and women (filled circles). Overall regression line for men (solid line), women (dotted line), and all subjects (solid bold line). *p < .05 regression line for age; ^† p = .054–.065 regression line against age. The linear regression equation coefficient (r) and p value are for n = 95.

Treadmill Exercisers Only

Resting and steady-state exercise, VO₂, Q_C, and stroke volume were not affected by age. Q_Cmax declined with age when indexed to total body mass or BSA (p ≤ .003) but not when expressed in absolute levels (p = .272) or indexed to FFM (p = .340), further demonstrating the significant influence of scaling choice on age-related differences. Although SV_max indexed to total body mass no longer declined with age (p = .896), SV_max indexed to FFM increased with age (p = .036). Maximal a-vO₂diff declined with age (p = .004). In treadmill exercisers, there was a positive and significant association between FFM and a-vO₂diff (r = 0.47, p < .001), but this relationship was negative for fat mass (r = −0.28, p = .028)

Discussion

This current study demonstrated that despite a decline in absolute and relative VO₂max, absolute levels of SV_max and TBV are well preserved with healthy aging. Consistent with our hypothesis, the decline in cardiovascular determinants of VO₂max was either attenuated (or surprisingly reversed) when FFM was used as the scaling variable compared with standard body size scaling variables. Therefore, these current results (a) do not support the hypothesis that a reduction in SV_max and TBV contributes to the age-related decline in VO₂max and (b) underscores the importance of determining the most appropriate scaling variables when investigating changes in cardiovascular determinants of VO₂max in populations of varying body sizes and composition.

The influence of age-related differences in body composition on cardiovascular determinants of VO₂max with aging has been reported previously (18). However, the findings of this previous study are limited by (a) all variables (VO₂max, stroke volume, TBV, FFM) were only collected in 27 subjects (≈26% of the entire cohort), (b) supine stroke volume and not maximal SV was collected, (c) TBV was only presented relative to total body mass, and (d) the influence of scaling index on age-related differences was not investigated. Therefore, our current study extends previous findings, which are important to delineate the limitations to exercise performance in healthy aging individuals, which in turn may allow for more effective prescription of lifestyle strategies including exercise training (29–31) to increase functional capacity in a globally aging population.

Aging and Cardiovascular Capacity

VO₂max declines 5%–20% per decade in healthy individuals aged 20–65, with greater rates of decline reported in those 70 years and over (1,2,5,6). VO₂max declines at a greater rate with age when scaled to total body mass or BSA rather than FFM (1,2,6), emphasizing the importance of scaling in interpreting these changes. Although an excess of fatness and body mass do not necessarily imply a reduced physiological ability to maximally distribute and utilize oxygen at the tissue level, it may instead limit submaximal and maximal exercise capacity due to a reduced ability to move excess mass through space during exercise (walking, running).

Nevertheless, VO₂max scaled to whole body or appendicular FFM declines with age (1–3), which could reflect either a reduced central component to oxygen delivery and/or a reduced utilization by metabolically active tissue. From a central component, it is clear that maximal heart rate declines (3,7,8); however, SV_max has been reported to decrease (3,9,12), increase (10,11,32), or be unaffected (7,33,34) with age in healthy individuals. Methodological differences including study design (longitudinal versus cross-sectional), sample size, measurement techniques, and exercise mode are likely to contribute to the discrepancies in previous findings. Although rarely investigated, two pieces of evidence suggest that differences in findings regarding SV_max may be heavily influenced by the approach taken to scale to body size measures. First, SV_max scaled to total body mass was smaller in healthy individuals aged 60 years and older compared with sex-matched young adults (<35 years); however, this effect was no longer statistically significant when SV_max was scaled to FFM (3). Second, scaling maximal cardiac power (Q_Cmax × mean arterial pressure) to metabolically active tissue was more precise at eliminating sex-related differences compared with BSA or total body mass in a group of subject ranging in age from 20 to 70 years (35). In this current cohort, the relationship between SV_max and age was unaffected or negative when scaled to total body mass but positive when scaled to FFM. Such findings highlight a potential confounding effect that may occur when standard scaling variables are used to examine cardiovascular capacity in aging individuals of varying body sizes and composition.

Our finding that TBV and its compartments are preserved (or increased) with age conflicts with previous findings (15–18). However, these previous findings are likely confounded by small sample size (15–17), restrictive age range (15–17), or failure to strictly control for physical activity levels (17). The findings of this current study suggest that irrespective of sex, TBV is reasonably well preserved with aging, which is consistent with previous findings (36,37). Moreover, we found that TBV compartments increased with age in both sexes when scaled to FFM. The only other study that has investigated age-related differences in TBV scaled to FFM reported that ECV and PV were reduced in seven older (mean age 66.1±1.8) men versus total body mass, BSA, and physical activity matched younger (24.7±1.2 years) men (15). The reasons for such divergent findings are unclear but may reflect a larger sample size that includes a broader age range in this current study. Nevertheless, one of the key findings of this current study is that irrespective of scaling approach, TBV is not significantly reduced with healthy aging.

Age-related Decline in a-vO₂diff

Consistent with previous findings (3), this current study demonstrates an age-related reduction in a-vO₂diff during maximal treadmill exercise. A lower a-vO₂diff with senescence may be multifactoral including a reduced number of mitochondria and associated enzymes, altered Q_C distribution, and increased vascular resistance to peripheral blood flow (38–42). However, a large portion of this reduction in a-vO₂diff may be related to physical deconditioning with aging, rather to the aging process itself, as aerobic exercise training increases a-vO₂diff in adults over the age of 60 due to an increased number of mitochondria and enzymes activity (43), increased vasodilator response, and enhanced blood flow redistribution to exercising muscle mass (39).

Scaling Cardiovascular Variables in Populations of Different Body Sizes and Composition

To accurately assess cardiovascular function and capacity in populations of different body sizes and composition, the most appropriate scaling model and scaling variables need to be identified based on physiological relevance. Previous reports (35,44,45) have commented that (a) based on physiologic and mathematical basis, an allometric scaling model is superior compared with the more conventionally used ratiometric model when examining cardiovascular physiologic and anatomic measures in populations of different body size and composition, and (b) based on blood flow distribution and metabolic demand during exercise, FFM would appear to represent a superior variable compared with the conventional scaling body size and mass variables. In early and late middle-aged adults (30–55 years), allometric modeling was more appropriate at distinguishing differences in arterial structure and function in obese versus nonobese individuals after differences in body size are considered (46), whereas Chantler et al. (35) showed that allometric scaling of maximal cardiac power to metabolically active tissue dramatically reduced gender-related differences compared with conventionally body size variables. The current findings underscore that FFM seems to be the most appropriate scaling index for cardiovascular variables, particularly in women, given the changes in total body mass and size, and more importantly, body composition with aging.

Methodological Considerations

There are several limitations to this current study. It is acknowledged that age-related differences in VO₂max may be in fact be nonlinear, with greater rates of decline observed in those older than 70 years compared with younger ages (2). As a result, the statistical analysis performed in this present study may not truly reflect the accelerated rate of decline in VO₂max in those 70 years and above. Moreover, it is also acknowledged that there were several results that were close to conventional levels of statistical significance, which likely reflects our sample size and retrospective methodological design. Despite its preferred usage, FFM does have some limitations. FFM is composed of highly metabolically active tissues during exercise such as skeletal and cardiac muscles but also less metabolically active tissues and organs such as the liver and the spleen. Moreover, the usage of FFM neglects adipose tissue blood flow, which increases in active limbs during exercise (47) but still remains much smaller compared with that of skeletal muscle (48). Finally, our subjects were nonobese, normotensive, and were screened for occult cardiovascular disease; therefore, it is unclear whether these current results are applicable to the larger population or to a physically active population.

Conclusions

In summary, this current study demonstrated that while absolute and relative VO₂max declines with age, SV_max and central circulatory factors are well preserved with aging in both sexes. A lack of an age-related decline in size-related cardiovascular determinants of VO₂max is evident when metabolically active tissue is used as the scaling variable. Collectively, the current findings do not support a hypothesis that a smaller SV_max and TBV contribute to the age-related decline in VO₂max. Instead, a reduction in maximal heart rate seems to be more influential in reducing VO₂max with advancing age.

Funding

This study was supported by the National Institutes of Health grant AG17479-02.

Conflict of Interest

There are no author conflicts to disclose.