*2.4. Ultrasound—Multi-Compartmental, Indirect (or Doubly Indirect If Prediction Equations Used)*

A relatively new method in the measurement of body composition is ultrasound (US). This method measures uncompressed subcutaneous adipose tissue thickness through US imaging [39] by transmitting a high-frequency beam through the skin at a chosen surface anthropometry site(s) via a handheld transducer head. Once the beam meets an interface (such as subcutaneous fat or muscle tissues) an image is partially echoed back to the transducer, whereby specific types of tissues transmit differing acoustic impedance (i.e., FM content increases the time required for sound reflections off BMC to return to the probe [40]). An image can then be produced from the echo reflected back to the transducer, allowing the integrated software to produce an estimation of the thickness of the tissue (for a more in-depth review see [39]). The use of US can be an accurate measure of subcutaneous adipose tissue; however, it can be altered by (1) the chosen US frequency; (2) pressure and orientation of the probe causing measurement error; (3) the technician's ability to choose the representative sites for measurements. Furthermore, US can be expensive and impractical within the applied setting, even with improved portability. One reported advantage of US is that intra-abdominal fat can be assessed more reliably than anthropometric measures [41], although, whilst this is useful to assess metabolic risk factors for cardiovascular disease, it is not routinely required in applied sport.

With regards to the application of US in applied sport, it is considered a reliable [42] and accurate measure [39] of subcutaneous adipose tissue thickness, but is limited by the plasticity of fat tissue and uneven tissue borders. More recently, Gomes et al. [43] reported US provides similar results for FM in comparison to skinfold and DXA across a range of athletes. However, the aforementioned study used a high-resolution B-mode US unit, which can be time-consuming (>20 min) and requires costly medical devices (>£11,000) along with expensive analysis software (>£1000) [44], that limits the application of such a method in an applied setting. Within an applied context, one of the more widely used methods is A-mode US devices which are commercially available at a considerably lower cost (such as the BodyMetrix® (Intelmetrix, Brentwood, CA, USA) device (<£2000)). However, research on the validity of such devices is at best equivocal, with poor agreeability reported in comparison with DXA in both non-athletic [45–47] and athletic young adults [48], and when compared to BOD POD in NCAA Division I athletes [44]. Moreover, Peréz-Chirinos Buxadé and colleagues [49] reported that the A-mode US devices produced significantly lower skinfold thickness scores in comparison with skinfold caliper measures performed by The International Society for the Advancement of Kinanthropometry (ISAK) qualified technicians. In terms of the reliability of A-mode devices, mixed results have been reported, from excellent [44,50], to acceptable [51] and also poor [49]. However, it is important to note there are a range of devices available and results may be specific to the device assessed. It should also be noted that the US method involves converting uncompressed subcutaneous adipose tissue thickness into a percentage body fat using regression equations, which adds another layer of inaccuracy that is discussed in more detail in Section 3.2. One advantage of the US device is that it has better inter-rater reliability than skinfolds in novice (non ISAK trained) practitioners [52]. Whilst the use of portable US may prove an exciting avenue for assessment of body composition in the future, this requires further development and research to assess the accuracy and reliability of available devices; however, it could be an effective tool to produce repeatable data when there is no access to suitably trained skinfold practitioners (see Section 2.7).

#### *2.5. 3D Photonic Scanning—One Compatmental, Doubly Indirect*

The use of 3D scanners, a form of digital anthropometry, originates from the assessment of human body shape for garmen<sup>t</sup> manufacture [53]. Three-dimensional scanners are now used for a variety of purposes, including assessment of body composition. Briefly, data with a 3D scanner involves the use of visible and infrared light to create an avatar of the human body, with the subject required to stand still in a particular posture whilst wearing minimal clothing. The reflection of the light off the body allows for a series of points to be captured with triangulation [54]. These points are connected to create a 3D mesh, with the use of landmarks to calculate circumferences, volumes, lengths and surface areas (for a more in-depth explanation of data acquisition and processing, see [55]). In comparison to DXA and computerised tomography (CT) scans, 3D scanners do not require ionising radiation or principal component analysis, the outcome can be used to create a pseudo-DXA scan [55]. The use of 3D scanners is time efficient (a scan takes approx. 10 s), which provides advantages over other time-consuming methods. Theoretically, this method could be employed on a regular basis within athletic populations for frequent assessments of body composition, providing visualisations for retrospective comparisons. Conversely, the cost of using 3D scanners and the operative expertise required may make it a prohibitive method in most applied settings.

The first published data on the use of 3D scanners to assess the body composition of athletes were by Schranz et al. [56], who compared elite Australian rowers to agematched non-athletic controls. They observed elite rowers had greater segmental volumes and cross-sectional areas, variables which cannot be measured with a one compartment method. Therefore, the authors suggested for talent identification purposes, 3D scanning may be implemented in the testing of potential athletes, with the same research group subsequently observing that 3D methods are better than 1D methods for predicting junior

rowing performance [57]. Evidence exists for the accuracy and reliability of 3D scanning when estimating body composition [56,58–60], although there are differences between commercially available scanners, which is partly due to the differences in the algorithms used [60]. Indeed, the development and refinement of the most valid algorithm and post-processing technique are still required [61]. Nonetheless, not all data support the validity of 3D scanning [62,63]. Cabre et al. [63] observed high typical errors and significant under and over predicting of BF%, FM and FFM using 3D scans compared to a four component model (combined DXA, ADP and BIS) in a non-athletic population. However, the authors observed no significant differences between 3D and DXA measures. Conversely, Tinsley et al. [60] observed high limits of agreemen<sup>t</sup> (LoA) for BF% (7–9.5%) and FM/FFM (5.3–7.2 kg) when compared to a four component model. As these LoA are in excess of expected changes from typical dietary and exercise interventions, longitudinal data that compare 3D scanning to other methods of body composition assessment during interventions aimed to alter body composition are required. In summary, given the expense and lack of athlete-specific validation data, this technique is uncommon in applied sport and is mainly suitable for laboratory-based research.

#### *2.6. Dual-Energy X-ray Absorptiometry—Three Compartmental, Indirect*

Whilst DXA was first developed for the measurement of bone mineral density (BMD), it has been extensively utilised in athletic populations for the assessment of body composition [11,64–66]. Indeed, DXA is now considered by many in the field as the 'criterion standard' of body composition assessment, despite being used in clinical settings for the diagnosis of bone related disorders such as osteoporosis. DXA operates by passing both high and low energy x-ray photons in either a pencil or fan-based beam, through differing body regions. The energy of these beams is attenuated by the density and volume of differing tissues, with soft tissues such as FM and LM allowing greater passage of photons when compared with denser tissues such as bone. The system software then produces an image in a rectilinear fashion and measurements of cross-sectional areas are completed with quantification of FM, LM and BMC in a two-dimensional image, calculated from the coefficient of two differing peaks in order to generate an R-value. For an further in depth description of the technical aspects of DXA measurement in differing systems, readers are directed to a recent review by Bazzocchi and colleagues [67]. Although DXA can be a reliable measure of body composition [68,69], utilisation of this method is not without its limitations, inclusive of legal and ethical constraints and technical considerations that will be discussed in more detail in subsequent sections. The major strength of DXA is the ability to measure BMC, which is growing in importance in due to the increasing awareness of low energy availability and the consequences of this on bone mineral content [70]. Furthermore, DXA provides limb-specific estimations of FM and FFM which can be useful when tracking injured athletes and the magnitude of fat loss in weight-making athletes [9,10].

#### *2.7. Skinfold Thickness—Two Compartmental, Indirect (or Doubly Indirect If Prediction Equations Used)*

Skinfold thickness assessment involves the use of a caliper to measure a double fold of gripped skin, over a range of differing sites to establish an overall measurement of subcutaneous adiposity [71]. This method is an inexpensive technique, requiring minimal equipment (calibrated calipers and anthropometric tape measure), allowing assessment to be conducted in a number of different field-based settings making it a popular method for estimating FM [71,72]. Detailed methodology of specific protocols are outlined in a number of texts [73,74]. The number of anatomical sites measured and equations used to predict both body density and FM using this technique varies significantly, which can create discrepancies in the data collected. As a consequence, ISAK was founded in 1986 to provide training courses and accreditation worldwide, setting professional standards for using skinfold thickness to assess body composition, with the eight site method now considered by many as best practice in applied sport settings. Although traditionally the most popular and suitable method for field testing, this doubly indirect method was

previously deemed unsuitable for the assessment of FFM and estimation of BF% [30]. The limitations and practical application of this technique will be covered in more detail in Section 3.

#### **3. Practical Considerations When Using DXA and Skinfolds as Measures of Body Composition in Applied Sport Practice**

Although traditionally skinfold thickness measurement has been the most popular method of body composition assessment in applied settings, DXA assessment has become increasingly more common in recent years [68], most likely due to a greater availability of machines and a belief that this is now the criterion standard. Despite these methods being used extensively in applied practice, both techniques produce outcomes based on a number of assumptions and require a high degree of standardisation for both accurate and reliable assessment, which is often ignored or not considered in applied sport settings. The following sections will serve to draw attention to these considerations, in the context of an applied sport setting, whilst providing a framework for best practice should these methods be considered and utilised.
