Caves are abundant around the world and are explored for recreational and physical activity purposes by thousands of cavers. Moreover, caves are often explored for scientific purposes. Cave environments are particularly hostile and extreme for humans because of high humidity, mud, darkness, and slippery. They are found in a great variety of settings, from cold alpine to warm tropical rain forests. For this reason, temperature can vary greatly in absolute terms, even though it is usually constant in each cave (White and Culver, 2012). Depths of several hundreds meters and kilometric lengths are not unusual in caves, explorations can last many hours or even days and require extensive climbing and ropework (Gunn, 2004). Thus, given the uniqueness of the environment and the duration of effort, caving can be considered a peculiar sports activity. However, very little is known about the physical capacity of cavers or about their energy expenditure (EE) during caving.

Considering the medical and logistical challenges associated with cave rescue operations, not surprisingly most of the available physiological literature deals with caving injuries (Stella-Watts et al., 2012; Cowart et al., 2014; Schneider et al., 2016). Only some specific studies have addressed the physiological challenge of the Naica Caves, Mexico, where temperature exceeds 45°C and humidity is well over 90% (Giovine et al., 2009) while limited attention has been devoted to the modification of biochemical and hematological parameters during conventional speleological practice (De Felice, 1952; Bregani et al., 2005; Stenner et al., 2006). To the best of our knowledge, to date there is only one study dealing with EE in caving (Antoni et al., 2017), but this investigation is focused mostly on body composition and hydration, while EE in relation to cavers' physical capacity has been only marginally addressed.

Given the scarce information on the EE and level of effort during caving, the physical training of cavers lacks a functional model. Thus, coaches build training programs without any knowledge of the level of effort reached by cavers during explorations.

This investigation was devised to study the physical capacity of a group of skilled cavers. Moreover, the EE during a typical session of a subterranean cave exploration was measured and then compared to the level of physical capacity previously assessed in a laboratory setting to understand the level of effort experienced by cavers in relation to their physical capacity. This information would be useful for coaches to design specific training programs on the basis of the specific EE required by caving.

All participants completed the study protocol. The Kolmogorov-Smirnov test confirmed the normal distribution for all the parameters examined.

Table 1 shows the results of the IET. In detail, cavers had a maximum oxygen uptake (VO_2max) of 2,248.7 ± 657.8 ml·min⁻¹ (i.e., 32.4 ± 6.4 ml·kg⁻¹·min⁻¹). The level of W_max was 205.7 ± 14.5 W, while anaerobic threshold (AT) occurred on average at 74.5% of VO_2max. Figure 1A shows the EE time course (1 h average) in each subject during caving. Two horizontal dotted lines identify the EE mean group level reached at W_max and AT during the IET. Figure 1 also shows separate graphs for female and male subjects (Figures 1B,C respectively). Results of the ANOVA test applied to the mean values are displayed in Figure 2A, while panel Figures 2B,C shows separate data for female and male subjects respectively. From this figure it can be gleaned that EE at W_max was higher than EE achieved at AT (vs. 674.3 ± 197.7 vs. 502.9 ± 154.6 kcal·h⁻¹) and also higher than the average EE during caving. Moreover, EE at AT was more elevated than during caving (502.9 ± 154.6 vs. 268.8 ± 54.8 kcal·h⁻¹).

Results from caving sessions provided an average time spent in the cave of 9.4 ± 1.2 h while the average EE was 268.8 ± 54.8 kcal·h⁻¹, which corresponded to about 40% of the maximum aerobic capacity measured during the IET. A mean distance of 10.6 ± 2.2 km was covered by subjects. TS were 10,344.8 ± 1,756.4.

The present investigation is aimed at studying the EE in a group of skilled cavers. In particular, we were interested in information on the EE during caving in relation to parameters related to cavers' physical capacity, which was previously assessed with a classical IET. To this end, EE was measured during a typical session of about 10 h subterranean cave exploration. Average EE level was then compared to the level of physical capacity measured in a laboratory setting.

Results suggest that cavers do not need to have a high level of aerobic physical capacity. Actually, the VO_2max (i.e., 32.4 ± 6.4 ml·kg⁻¹·min⁻¹) was only slightly higher than that of a sample (n = 15) of sedentary subjects with similar age and body composition selected from the database of our laboratory (30.6 ± 5.8 ml·kg⁻¹·min⁻¹). This suggests that a high aerobic fitness is not needed to cope with the energy demands of caving. Moreover, throughout caving the EE was on average well below the level of AT. Furthermore, looking at the EE time course of each subject (Figure 1A) it can be gleaned that EE was quite stable in all subjects, without showing abrupt increments and always below the average AT level. Thus, results indicate that caving is not physically demanding. However, in absolute terms, the total EE was high (i.e., 2,672.3 ± 576 kcal in total during a single session) because of the long time spent in exploration. This may be challenging for the body homeostasis, and suggests that the assumption of caloric food portions may be recommendable during long lasting explorations.

Notwithstanding the fact that participants performed well below the level of AT and that their physical fitness was not high in terms of aerobic capacity, it should be taken into consideration that explorations need climbing, rappelling, and ropework (Gunn, 2004). Hence, a good level of muscular strength is probably essential during caving. Moreover, this kind of effort, which can be prolonged for several hours, may increase considerable fatigue, thereby increasing the risk of injuries. It is then conceivable to speculate that specific training programs aiming at increasing muscle strength are useful in this activity.

In the past, several investigations have employed HR in an attempt to estimate EE during various kinds of exercise and daily activities (Gilman and Wells, 1993; Li et al., 1993; Hiilloskorpi et al., 1999; Achten and Jeukendrup, 2003; Green, 2011). The development of accurate and inexpensive telemetry HR monitors has made this method attractive and is widely utilized in freely moving athletes during their field performance. This method is based upon the assumption that there is a direct relationship between HR and VO₂, which in turn reflects EE. Thus, EE may be calculated from HR. However, while the HR/VO₂ relationship assessed during incremental laboratory exercise testing is stable and reliable, in the field setting, i.e., when a subject moves freely without a fixed incremental protocol, this relationship is affected by several potentially perturbing factors. In these situations HR may not reflect VO₂ accurately. One of these conditions is the phenomenon of “cardiac drift,” which appears when exercise duration exceeds 20 min or longer and causes HR to rise slowly and continuously despite stable VO₂ values. Also heat stress and dehydration increases HR with respect to the real metabolic engagement. All these situations may perturb the HR response, which in turn results excessively high for a given VO₂, thus overestimating the athlete's actual EE (Crisafulli et al., 2006; Milia et al., 2014; Tocco et al., 2015). Taking into consideration these limits, i.e., the long duration of these sessions, the possibility of heat stress and dehydration during caving, we preferred not to use HR to monitor the EE.

One potential limit of the present study was that two methods of measuring of EE were employed: indirect calorimetry conducted in a laboratory setting and the Armband Fit Core during field testing. Although the Armband Fit Core has been proven sufficiently accurate in estimating EE in comparison with indirect calorimetry (Johannsen et al., 2010; Brazeau et al., 2016), it should be noticed that this device has several times been found to underestimate EE during high-intensity exercise (Drenowatz and Eisenmann, 2011; Koehler et al., 2011; Benito et al., 2012; van Hoye et al., 2014). Therefore, we cannot know whether in the present study a systematic underestimation of EE was introduced because of the presence of bouts of high intensity effort during caving. Moreover, considering that climbing and rappelling require the recruitment of anaerobic energy sources, it is reasonable to assume that a substantial amount of energy requirement was derived from anaerobic metabolism. However, the Armband Fit Core has been validated only against indirect calorimetry, which is able to measure EE derived from the aerobic metabolism but not that obtained from anaerobic energy sources. Thus, the Armband Fit Core probably underestimated the total EE for two at least two reasons: because of its low reliability during high-intensity exercise bouts and because it did not take into consideration the energy derived from the anaerobic energy sources.

This is the first study to assess EE during real caving in relation to cavers' physical capacity. Results show that cavers did not need a high aerobic capacity to meet the EE demands of caving in the setting of our experiment However, in absolute terms the total EE was high because of the long duration of sessions. The present information could be useful for coaches to build dedicated training as well as diet programs able to develop the specific adaptations and food intake required by this activity.

VP, SM, GS, GG, GA, FT, AR, and AC conceived the study, conducted experiments, designed and wrote the paper. SV, SO, EM, NC, SC, and PF conducted experiments, designed and wrote the paper. All authors read and approved the final version of the manuscript.