Bioimpedance phase angle in elite male athletes: a segmental approach

Assessing body composition (BC) in the field of sport has proven to be crucial in detecting individual characteristics critical to performance, evaluating the effects of training and managing weight strategies in weight-category sports (Turnagöl 2016, Fields et al 2018, Sonksen 2018). Therefore, its evaluation may interest athletes involving in different sport disciplines, because of their different training programs and energy requirements. Bioelectrical impedance analysis (BIA) is a portable, non-invasive, easy-to-use and inexpensive method widely used in the field of sport for assessing BC by means of predictive equations (Moon 2013). Specifically, BIA measures impedance (Z), the opposition of the body to an alternating current resulting from resistance (R) and reactance (Xc), while phase angle (PA) is identified as the arctangent of the Xc to R ratio (Norman et al 2015).

An increasing amount of evidence has focused on the health significance of raw bioimpedance data. PA has gained attention because it has proven to be a proxy of water distribution (extracellular/intracellular water ratio) and body cell mass (BCM) (Campa et al 2020, Marini et al 2020). Thus, high PA suggests greater cellularity (e.g. more BCM relative to fat-free mass (FFM)), cellular integrity and adequate cell functions (Norman et al 2012, Stobäus et al 2012, Beberashvili et al 2014, Zhang et al 2014). While several studies have evaluated PA in the clinical setting, fewer (Carrasco-Marginet et al 2017, Silva et al 2019, Di Vincenzo et al 2019) have explored its role in sport, as a helpful proxy parameter in evaluating changes in water distribution and in monitoring FFM hydration status.

PA values varies depending on gender, age or BMI ranging in healthy subjects from 5 to 7.5 degrees (Kyle et al 2001, Barbosa-Silva et al 2005, Bosy-Westphal et al 2006), whereas in trained athletes it may reach 8.5 degrees (Marra et al 2016). In this context, a recent systematic review evaluated the variability of PA among athletes and its relationship with sports performance (Di Vincenzo et al 2019). Only a few studies showed that PA was higher in athletes vs controls and it appears still uncertain to what extent PA differs among various sports. Additionally, there were no clear relationships of PA with sport performance or the level of training. Moreover, although these investigations reported data on whole-body PA values, to date, limited data are available concerning the segmental assessment (whole-body and limbs) that in the field of sport could be useful for assessing differences among athletes and slight variations due to training in specific body segments.

Cycling, water polo and ballet dance are very different sport specialties and the type of physical exercise and training varies among these disciplines. To the best of our knowledge, no studies have examined differences on total and segmental PA among these athletes.

Based on this background, the objectives of the present study were as follows: (a) to assess whether whole-body and segmental PA varied among elite male athletes practicing cycling, water polo and ballet dance; (b) to assess whether PA differed between elite athletes and healthy control subjects; and (c) to explore the relationship between PA and anthropometric and BC variables.

2.1. Subjects

2.2. Anthropometric and BIA measurements

2.3. Statistical analysis

Participants included three groups of male athletes (18 cyclists, 20 water polo players, 18 ballet dancers) and three groups of healthy control subjects, each of which similar for general characteristics (one to one) to the previous ones (15, 15 and 22 subjects, respectively). Individual characteristics of the participants are presented in table 1; BC variables and PA are summarized in table 2.

3.1. Cyclists vs control group

3.2. Water polo players vs control group

3.3. Ballet dancers vs control group

3.4. Cyclists vs water polo players vs ballet dancers

3.5. Correlations between anthropometric and body composition variables and phase angle

Pearson's correlation coefficients assessed the association of both whole-body and limb PA with both anthropometric and BC variables in athletes and controls (table 3).

Table 3. Pearson's correlation for the association of phase angle with both anthropometric and body composition variables.

	Whole-body			Upper limb			Lower limb
				Athletes (n = 56)
	r	p		r	p		r	p
Age	−0.122	0.537		−0.175	0.372		−0.006	0.974
Weight	0.508	0.006		0.372	0.051		0.238	0.222
Stature	0.069	0.726		−0.069	0.726		−0.098	0.621
BMI	0.639	0.000		0.527	0.004		0.371	0.052
FFM	0.651	0.000		0.629	0.000		0.347	0.070
FM	0.417	0.027		0.245	0.209		0.213	0.275
FM%	0.325	0.091		0.150	0.445		0.154	0.434
				Controls (n = 52)
Age	0.099	0.503		0.402	0.002		−0.181	0.218
Weight	0.284	0.050		0.417	0.003		−0.041	0.782
Stature	0.113	0.443		0.288	0.047		−0.082	0.580
BMI	0.280	0.042		0.365	0.011		0.016	0.915
FFM	0.212	0.149		0.479	0.001		−0.143	0.332
FM	0.270	0.062		0.153	0.298		0.145	0.326
FM%	0.242	0.097		0.016	0.915		0.210	0.153

BMI = body mass index; HGS = handgrip strength; FFM = fat-free mass; FM = fat mass.

We found that in athletes whole-body PA had a strong association with body weight (r = 0.508; p = 0.006; figure 2), BMI (r = 0.639; p < 0.001) and FFM (r = 0.651, p < 0.001; figure 1). Findings in controls showed a weak correlation with body weight (r = 0.284, p = 0.050; figure 2) and BMI (r = 0.280; p = 0.042).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Upper limb PA in athletes was positively associated with BMI (r = 0.527, p = 0.004; figure 3) and FFM (r = 0.629, p < 0.001; figure 4) whereas in controls it was correlated with age (r = 0.402; p = 0.002), body weight (r = 0.417; p = 0.003), stature (r = 0.288; p = 0.047), BMI (r = 0.365; p = 0.011) and FFM (r = 0.479; p = 0.001). No correlations were found between lower limb and the studied variables in both groups.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

3.6. Variables influencing phase angle

The main purpose of the present study was to assess whether segmental PA differed among elite male athletes practicing cycling, water polo and ballet dance. To our knowledge, this is the first study to focus on the differences in PA values among these groups of athletes. Previously, limited cross-sectional studies have compared PA among athletes practicing different sports but none of them have focused on the segmental PA evaluation (Kim et al 2010, Koury et al 2014, Galanti et al 2015). The development of muscle mass varies depending on sport type; the morphological characteristics of athletes are different mostly due to the specific requirements of each sport. The major muscles involved in cycling are leg muscles (especially quadriceps and thigh muscles) (Mujika et al 2016), whereas in water polo and dance training involves upper and lower muscle groups, with a more balanced distribution of increased muscle mass (Koutedakis and Jamurtas 2004, Lozovina 2009, Ferrari et al 2013, Cox et al 2014).

Our findings did not show any significant differences between the three groups of athletes, even after adjusting for FFM, although whole-body and limb PAs were slightly higher, albeit not significantly, in water polo players. That may be due to the small groups of participants.

As another aim, we evaluated whether PA differed between athletes and control subjects, since current data are limited (Di Vincenzo et al 2019). Only few papers have addressed this issue: a marked increase in whole-body PA was observed in bodybuilders (Piccoli et al 2007), dancers (Marra et al 2009), ballet dancers (Marra et al 2018b), cyclists (Marra et al 2018b), marathon runners (Marra et al 2018a) and volleyball players (Di Vincenzo et al 2020). The results of this study are in line with literature data. Additionally, in our study interesting evidence emerged with the segmental evaluation. Specifically, we found that lower limb PAs were higher in all groups of athletes than in controls, probably due to the severe involvement of the lower limbs in sport activity. Similar results emerged for upper limb PA values in water polo players and ballet dancers but not in cyclists who, surprisingly, showed similar upper limb PA values compared to their controls. We attributed these results to the relatively less development of the arms during training and competitions.

Furthermore, an interesting issue was to explore whether and to what extent anthropometric and BC parameters were related to PA in elite athletes. As far as we know, no consistent data are available in the literature on the topic (Di Vincenzo et al 2019). Based on our results, anthropometric and BC variables such as body weight, BMI and FFM were stronger associated with whole-body and upper limb PA in athletes than in controls.

Additionally, the multivariate regression model showed that FFM was the strongest determinant of both whole-body and upper limb PA, followed by stature. Surprisingly, no associations were found for lower limb PA in both athletes and controls.

Despite the useful information that the segmental PA evaluation could provide in sport science, very few studies focused on changes in body segments due to training. Specifically, a study by Marra et al (2016) on professional cyclists, suggested the use of segmental PA measurements to monitor athletes' FFM modifications during medium- and long-term competitions. They have demonstrated a clear reduction in lower limb PA accompanied by a substantially unmodified upper limb PA values during the same assessment and period of time. Conversely, resistance-trained females showed increased both whole-body and limb PA values after a 8 weeks training programme (Tinsley et al 2019). These contrasting results may be explained by the different type of sport activity: a strenuous exercise about 200 km cycling per day for 20 consecutive days (Marra et al 2016) vs a training programme for three times per week for 8 weeks (Tinsley et al 2019).

Finally, we evaluated and compared BC variables (such as FFM and FM) in the three groups of athletes and between athletes and controls. We found that water polo players had a substantially higher amount of FM than cyclists and ballet dancers and similar to their control group, probably because water polo players may retain slightly more fat to assist with floating during training and match (Andreoli et al 2003, Lozovina 2009, Cox et al 2014). On the other hand, as expected, ballet dancers had both absolute and percentage FM clearly lower than other athletes (possibly due to reduced dietary energy/calorie intake, high performance and because of body image issues) and, especially, than their control group. These results demonstrate that specific training for several years has influenced BC of ballet dance athletes, in line with previous evidence reported from the latest literature (Ferrari et al 2013, Stokic et al 2016, Kruusamäe et al 2016, Amorim et al 2017, Marra et al 2019). Indeed, it should be reminded that the Sun equation, used to predict FFM in all the participants, was derived using a four compartment model and is usually considered very reliable for the general population (Caucasians) but so far it has not been specifically validated in athletes.

The strength of this study are the assessment of segmental BIA and the use of a single BIA device, which limits the device-related PA variability. This study suffers from a few limitations such as being retrospective and the low number of participants because of the difficulty to enroll elite athletes.

This study further supports the idea that PA is directly related to the level of physical activity. The novelty of this study is represented by the direct evaluation of segmental PA among athletes practicing different sports, which can be employed by researchers and specialists in assessing variations in physiological variables, for example in response to interventions.

In fact, while whole-body PA could represent a proxy parameter of soft tissue mass quality of the overall body, the limbs evaluation can be useful for monitoring the condition of the athlete preparing for the competition, for assessing differences among athletes and slight variations due to training in specific body segments.

While our results are promising, large cross-sectional studies and, possibly, longitudinal studies (both also considering the size of major muscle groups) are needed to confirm that the current use of BIA is valuable in assessing muscle quality, and in evaluating differences due to gender, training volume, playing position, levels, etc.

The authors have no affiliations with or involvement in any organization or entity with any financial interest or non-financial interest in the subject matter or materials discussed in this manuscript.