Participants were recruited using social media, online climbing forums and posters in local climbing centres (Sheffield, UK). Forty climbers (20 females, 20 males) volunteered to participate. Participants were required to meet the following inclusion criteria: age ≥ 18 years, ≥2 years climbing experience, currently taking part in climbing or climbing specific training ≥2 × per week, in good health with no acute or chronic illness that may influence dietary intake.

Participants answered a series of questions to identify years of climbing experience, predominant climbing discipline (bouldering/sport), weekly training/climbing volume, dietary preference (vegan/vegetarian or omnivorous) and highest climbing difficulty grade attained in the last 6 months. Self-reported climbing ability has been shown to be a valid representation of actual climbing ability (34). Climbing grades were converted from Font (bouldering) and French/sport (sport climbing) grading systems to the International Rock Climbing Research Association Reporting Scale to support a common approach to the statistical analyses within rock climbing research (35).

Body mass was measured to the nearest 0.1 kg using electronic scales (Tanita, Japan), and height to the nearest 0.1 cm using a wall-mounted stadiometer (Holtain Ltd., UK). Body composition was assessed using the International Society for the Advancement of Kinanthropometry (ISAK) 8-site skinfold profile (36), carried out by an ISAK certified (Level 1) practitioner with an average technical error of measurement of 1%. Skinfold thickness was measured to the nearest 0.2 mm at eight sites (biceps, triceps, subscapular, iliac crest, supraspinal, abdominal, anterior thigh, and medial calf) using research standard calipers (Harpenden, UK). Duplicate measures were taken at each site and, where the technical error of measurement (TEM) was <5%, the mean value was reported. Where the TEM was >5%, a third measure was taken, and the median value reported. Girth measurements (relaxed arm, waist, gluteal and calf) were taken using a Lufkin metallic tape (W606PM, ATG, US). Body fat and fat-free mass (FFM) percentages were calculated using the Durnin and Womersley (37) equation, which has previously been validated against dual-energy X-ray absorptiometry in elite sport climbers (38).

Resting metabolic rate (RMR) was calculated using the Cunningham (1980) equation [RMR (kcal·day⁻¹) = 500 + 22 (Fat Free Mass)]. This equation was chosen for its established application in highly active individuals (39). The RMR for each climber was multiplied by a physical activity factor of 1.443 to represent a “moderate” exercise level. Due to the highly variable energy requirements of climbing, and lack of reliable predictive models, this calculation was used as a conservative estimate to allow comparisons to be made between the actual intake of the climbers, and the energy needs to support a “moderately active” training programme. The self-reported data collected suggested that all participants had a training volume that met or exceeded the criteria for a “moderate” exercise level (3 × hard/5 × light sessions per week).

Participants were instructed to weigh (in grams) or measure (in millilitres) all foods, fluids and supplements consumed in a 3-days non-consecutive diary, an accepted method used in previous research to collect this type of data (40). Measurements were taken using self-owned, commercially available digital kitchen scales or by reporting manufacturer weights. An electronic template was provided with guidance notes detailing the requirements for accurate reporting. Participants were instructed to choose three non-consecutive days within a 7-days period to record, and to consider capturing days with a range of training demands (e.g., rest day; indoor climbing training; outdoor climbing session). When consuming pre-packaged products, participants were required to report the brand name, the weight of the product and manufacturer nutritional values. The diet diaries were analysed by a registered sports nutritionist using Nutritics software (Version 5.096). Where exact foods were not listed within the software database, similar foods with matched macronutrient composition were selected. Dietary supplement intake was included within the dietary analysis, therefore, contributing to the energy, macro-, and micro-nutrient content of the log.

A random, non-fasted venepuncture blood sample was taken by a trained phlebotomist in line with the World Health Organisation guidelines on drawing blood (41). Blood sampling was not restricted to a specific time of the day in line with diurnal variation data in markers of iron status (42), with fasting conditions recommended only when assessing iron overload (43). Blood was collected in a serum vacutainer (BD, USA) for analysis of serum ferritin and transferrin saturation, and an EDTA vacutainer for follow up haemoglobin analysis. Serum samples were allowed to clot for 30–60 min at room temperature, then processed in a centrifuge for 10 min at ~2,000 × g (Clinispin Horizon 642E, Drucker Diagnostics, USA) before being securely packaged and posted, as per testing laboratory guidelines. Serum ferritin and transferrin saturation were analysed by an external nationally recognised medical laboratory (TDL, London, UK), using a Roche “Cobas 8000” blood analyser (Roche Diagnostics GmbH, Germany). Serum ferritin was analysed using the sandwich principle; iron using a colorimetric assay; and unsaturated iron binding capacity (UIBC) via direct determination with FerroZine. Transferrin saturation was calculated as; transferrin saturation = [(iron) × 100/(iron+UIBC)]. Haemoglobin was analysed immediately after blood collection using the azide-methemoglobin method (Hb 201 System; HemoCue AB, Sweden) in participants with sub-optimal iron status; serum ferritin (<35 μg·L⁻¹) and/or transferrin saturation (<20%) (44, 45). The sub-optimal iron status cut-off points specified were the same for both genders.

Statistical analysis was performed using SPSS software (version 24, IBM, USA). Data was checked for homogeneity of variance using Levene's test, and normality using Shapiro-Wilk's. Non-conforming data sets were transformed using Log-10. The differences in variables between groups (e.g., males vs. females) was analysed using an independent samples t-test, with significance set at p ≤ 0.05. A Pearson correlation coefficient determined the relationship between data sets (e.g., ability and energy intake). Correlation values (R²) were set as <0.2: weak correlation, 0.5: medium correlation, and >0.8: strong correlation (46). Data are presented as means ± standard deviation (SD), unless otherwise stated.

Participant demographics are shown in Table 1. Forty experienced climbers (n = 20 male, n = 20 female; 8.8 ± 6.6 years' experience) aged 18–46 (mean age 30.3 ± 6.7 years) participated in the study. The average climbing ability of the cohort using the International Rock Climbing Research Association (IRCRA) scale was 22.2 ± 3.7; 47.5% (n = 19) of the climbers were classed as intermediate to advanced level [IRCRA score: 10–23 for males, 10–20 for females; (35)], with 52.5% (n = 21) meeting classification criteria for elite or higher elite (IRCRA score: 24–32 for males, 21–32 for females (35). Twenty-seven climbers identified “bouldering” as their primary climbing discipline, with the remaining thirteen climbers reporting “sport climbing.” One climber reported a previous case of iron deficiency anaemia. No other known health issues, or the use of medications that could impact dietary intake were reported. Average BMI was 21.6 ± 1.7; a BMI of <18.5, defined as potentially “underweight” (47) was reported in one female participant. Over a quarter (27.5%) of the climbers reported being vegan (n = 5) or vegetarian (n = 6).

Body composition results are shown in Table 2. With the exception of gluteal girth, statistical analysis revealed significant gender differences across all the measured parameters (p ≤ 0.05). No significant differences were observed between ability groups (intermediate/advanced vs. elite/higher elite) for males or females. However, weak to medium correlations were seen in males between the IRCRA ability score and body mass (R² = 0.506, p = 0.02), and height (R² = 0.478, p = 0.03).

Iron status results are shown in Table 4. Data from two male participants was omitted from the iron status analysis. The aforementioned participant was excluded due to taking a high dose iron supplement, while hemochromatosis was incidentally identified in one participant during the study (serum ferritin 515 μg·L⁻¹).

Sub-optimal iron status was found in 16.6% of males (n = 3) and 45% of females (n = 9). One quarter of females (n = 5) met the criteria for stage 1 iron deficiency (ferritin <35 μg·L⁻¹, Hb > 115 g·L⁻¹, transferrin saturation > 16% (44), and one female was identified with Stage 2 iron-deficient non-anaemia (ferritin <20 μg·L⁻¹, Hb > 115 g·L⁻¹, transferrin saturation <16%). Follow-up testing revealed only one male participant with anaemia (Hb <130 g·L⁻¹).

There was no significant correlation between serum ferritin and energy intake (kcal·kgFFM⁻¹·day⁻¹) for males (R² = −0.075, p = 0.77), or females (R² = 0.05, p = 0.83). Furthermore, there were no significant differences between intermediate/advanced and elite/higher-elite level climbers for serum ferritin in males (111.7 ± 59.2 vs. 70.3 ± 25 μg·L⁻¹; p = 0.08), or females (46 ± 28.2 vs. 53.7 ± 23.1 μg·L⁻¹; p = 0.42). When analysed separately, there was no significant difference between vegan and vegetarian climbers for serum ferritin in females (30.7 ± 11.1 vs. 37.0 ± 24 μg·L⁻¹; p = 0.82). However, serum ferritin was significantly lower in vegan/vegetarians combined when compared to omnivores in female climbers (33.2 ± 14.8 vs. 57.5 ± 24 μg·L⁻¹; p = 0.05) (Figure 2).

Anthropometric and body composition data in the present study were similar to those previously reported in the literature (3, 5, 12, 38), with mean values for height, mass, BMI and body fat % of 177 cm, 69.4 kg, 22.1 (BMI), and 12.0% for males, and 166.8 cm, 58.5 kg, 22 (BMI), and 22.9% for females, respectively. Early research presented climbing athletes as short in stature, with a low body mass (19). Conversely, the present study shows significant weak to medium correlations in males between the IRCRA ability score and body mass, as well as height. However, it is difficult to draw conclusions as the existing anthropometric data in climbing research varies considerably, particularly body fat %, with a range of mean values reported between ~5–13% for males, and ~10–25% for females (3, 5, 12, 19, 38). Explanations for this variation might include the method used to assess body composition [Skinfold vs. DEXA; (38)], ability level, whether the climbers climb/compete indoors or exclusively climb outdoors, the timing of the measurement in relation to peak conditioning (particularly relevant to competition athletes), and the changing demands of the sport as it has developed over the years. Nevertheless, this study adds further data in climbers, where recent body composition figures are lacking.

There were no significant differences found in any of the anthropometric characteristics between ability groups for males or females, which again supports previous findings (5, 50). Furthermore, Laffaye et al. (2) determined that the only significant anthropometric difference found between novice and elite boulderers was an increased ape index, concluding that anthropometric variables explained only 4% of performance variation. Similarly, Mermier et al. (12) concluded that when anthropometric characteristics are similar, climbing performance is determined to a greater degree by training variables, rather than physique (58.9 vs. 0.3% of total variance). Considering the large variation in the physiological requirements of each climbing route or boulder problem, we have previously proposed that an ideal physique may not exist in climbing (51). Due to the reduced load and friction requirements, a lower mass may be favoured on a route with small holds and where there is more time spent static. Conversely, on routes with higher friction and more explosive, strength reliant moves, an athlete might benefit from greater muscle hypertrophy and consequently, increased force development (51).

Iron deficiency has been shown to negatively impact aerobic power, with larger deficiencies correlating with greater reductions in oxygen transport to the working muscles (32). Reduced aerobic power is likely to place greater demands on anaerobic metabolism during climbing (66), especially on steeper routes where it is considered the predominant energy source (9, 67), and may exacerbate decrements in performance. Despite resulting in impaired physiological function, iron deficiency is a commonly reported issue in athlete populations, affecting ~3–11% of male athletes, with a higher prevalence of ~15–35% seen in females (32). In the present study, 31.6% of participants had a sub-optimal iron status as defined in the literature (44, 45). The prevalence of sub-optimal iron status was greater in females (45%) compared to males (~17%), with one quarter of females meeting the criteria for stage 1 iron deficiency (ferritin <35 μg·L⁻¹, Hb > 115 g·L⁻¹, transferrin saturation > 16%) (44).

It has been suggested that low energy availability (LEA) may be partially induced by and result in the development of iron deficiency (30). The proposed mechanism is related to the iron deficiency induced perturbation of thyroid function, leading to decreased appetite and impaired metabolic efficiency; which can result in a reduced energy intake, increased energy expenditure and potentially, further exacerbation of LEA in athletes (30).

Considering the high prevalence of both sub-optimal energy intake and iron status in the climbing population assessed, it would be reasonable to suggest an interaction of this nature, however, statistical analysis revealed that there was no significant correlation between serum ferritin and energy intake for males, or females. Therefore, further research should consider iron status in relation to energy availability, rather than overall intake. Another factor to consider in the incidence of ID in this population is exercise-induced haemolysis, as previous research has suggested that this is exacerbated by muscle damaging exercise (68), and increased blood lactate levels (69); both of which are considered physiological features of climbing and training for climbing (70). Finally, menstrual blood losses of iron in females, may contribute toward the higher prevalence of ID seen within this population group (61). Larsson et al. (71) reported lower menstrual blood loss (~50%) in combined oral contraceptive pill (OCP) users compared to non-users, with OCP administration increasing serum ferritin levels by ~21–29% in subjects with pre-existing low iron stores (ferritin <10 μg·L⁻¹), therefore, OCP use offers a therapeutic intervention for women who struggle to maintain iron stores due to heavy menstruation, with the potential additional benefit of protection from soft tissue injury (72). It is crucial to ensure that the dietary intake of the female athlete supports the energy needs of eumenorrhea and includes iron rich foods to attenuate a continued loss via menstruation. Future research may also consider controlling for variables which may affect iron loss in females.

Although there was no significant difference in iron intake between the dietary preference groups, serum ferritin was significantly lower in vegan/vegetarians when compared to omnivores in female climbers (33.2 vs. 57.5 μg·L⁻¹), with no difference seen between vegan vs. vegetarian climbers when analysed separately. This could be explained by the different mechanism by which heme iron from animal products is absorbed compared to non-heme iron derived from plants, resulting in more efficient absorption which is less affected by accompanying dietary factors, and therefore, significantly greater bioavailability (73, 74). Whilst the presence of vitamin C can enhance the absorption of non-heme iron, chemicals (polyphenols and phytates) and minerals (calcium) that are found in tea, coffee, whole grains, legumes and dairy products, can inhibit the absorption of non-heme iron within a meal (75) and therefore, the overall intake and timings of these foods should be carefully considered by vegan/vegetarian athletes at risk of ID.

Limitations in the assessment of iron status in this study may lead to confounding results. For example, as an acute state reactant, serum ferritin (SF) may be artificially raised in response to intense exercise (20). Although participants were instructed to avoid exercise in the hours preceding the lab visit, exclusion of this variable relies on strict adherence to instruction. In addition, research suggests that SF may be decreased during menstruation in female participants (76), which is affected by variables such as the menstrual cycle phase (61), the use of birth control (71), or amenorrhea; induced by menopause, or LEA (28). Due to a lack of control of these variables, it is not possible to exclude this interaction and therefore, the data should be considered as preliminary at this stage, although the defined cut off values for ID are consistent irrespective of these variables. Furthermore, C-reactive protein, a marker of inflammation, was not measured and therefore it cannot be ensured that SF was not confounded by infection, inflammation, or injury (77), however, all participants reported to be in good health at the time of testing.