================================================================================ OXYGEN UPTAKE IN WHEELCHAIR FENCING Direct assessment and criterion validity of a sport-specific, heart rate-based estimation method ================================================================================ Dataset accompanying the article published in PLOS ONE. Authors: Xavier Iglesias, Rafael Tarrago, Diego Chaverri, Jordi Montraveta, Daniel Muniz-Pumares, Ignacio Fernandez-Jarillo, Lindsay Bottoms Contact: Xavier Iglesias - xiglesias@gmail.com INEFC, Universitat de Barcelona (Spain) Corresponding author of the article: Lindsay Bottoms - l.bottoms@herts.ac.uk University of Hertfordshire (United Kingdom) Article: [DOI OF THE ARTICLE - TO BE COMPLETED ON PUBLICATION] Dataset DOI:10.18745/ds.00027179 Licence: CC BY 4.0 Funding: Ministerio de Universidades (Spain), PRX21/00210; Ministerio de Cultura y Deporte, Consejo Superior de Deportes (Spain) and European Union, EXP_74847 (2023). HOW TO CITE THIS DATASET Iglesias, X., Tarrago, R., Chaverri, D., Montraveta, J., Muniz-Pumares, D., Fernandez-Jarillo, I. and Bottoms, L. (2026). Oxygen uptake in wheelchair fencing [Data set]. University of Hertfordshire. (Complete with the DOI issued by the repository on deposit.) SUGGESTED DATA AVAILABILITY STATEMENT FOR THE ARTICLE "Data relevant to this study are available from: [Iglesias, X., Tarrago, R., Chaverri, D., Montraveta, J., Muniz-Pumares, D., Fernandez-Jarillo, I. and Bottoms, L. (2026). Oxygen uptake in wheelchair fencing [Data set]. University of Hertfordshire.]" -------------------------------------------------------------------------------- 1. STUDY IN BRIEF -------------------------------------------------------------------------------- Eleven elite wheelchair fencers (7 class A, 4 class B; 3 women, 8 men) completed a wheelchair fencing-specific incremental test (WFST) with continuous heart rate (HR) and breath-by-breath oxygen uptake (VO2; COSMED K5). An individual linear equation VO2 = a + b*HR was derived for each fencer. They then performed two 15-hit epee bouts, one of which was monitored, to compare directly measured VO2 with VO2 estimated from HR. A subsample of seven fencers took part in a simulated epee competition in which HR was recorded continuously and VO2 was estimated from the individual equation. Data were collected in the summer of 2022 in Troyes (France) and Bath (UK). Ethics: Clinical Research Ethics Committee of Sport in Catalonia, ref. 012020CEICEGC. -------------------------------------------------------------------------------- 2. DE-IDENTIFICATION -------------------------------------------------------------------------------- Each fencer is identified only by an anonymised code (F1-F11). The codes are the same as in Table 2 of the article, so table and data can be linked. Rows are ordered by calibration coefficient; the codes do NOT correspond to the order of assessment. Demographic and clinical characteristics of individual participants (age, sex, functional class, Tier, impairment, body mass, stature) are NOT included. Elite wheelchair fencing is a very small population and the combination of those variables would make individual participants identifiable. Those characteristics are reported in aggregate form in Table 1 of the article. -------------------------------------------------------------------------------- 3. FILES -------------------------------------------------------------------------------- 01_test_and_bouts.csv Calibration and validity data (3 sheets) 02_competition.csv Simulated competition, second-by-second (7 sheets) 03_competition_F7_tidy.csv Input for the competition figure script 04_bland_altman_equivalence.R Script for the Bland-Altman and equivalence figure 05_competition_figure.R Script for the competition figure README.txt This file -------------------------------------------------------------------------------- 4. 01_test_and_bouts.xlsx -------------------------------------------------------------------------------- Sheet "Summary_per_fencer" - one row per fencer (n = 11) Fencer Anonymised code (F1-F11) n_WFST Number of 10-s intervals used for the individual regression a Intercept of the individual equation VO2 = a + b*HR (mL/min) b Slope of the individual equation (mL/min per beat/min) r Pearson correlation of the individual HR-VO2 relationship R2 Coefficient of determination of that regression SEE Standard error of the estimate (mL/min) SEE_kg Standard error of the estimate (mL/kg/min) HR_WFST_mean / _peak Mean and peak HR during the WFST (beats/min) VO2_WFST_mean / _peak Mean and peak VO2 during the WFST (mL/kg/min) Duration_bout_min Duration of the monitored bout (min) HR_bout_mean / _peak Mean and peak HR during the monitored bout (beats/min) pct_HRpeak_bout Mean bout HR expressed as a fraction of WFST peak HR (provided for completeness; not reported in the article) VO2_bout_measured_mean/_peak Mean and peak VO2 directly measured during the bout (mL/kg/min) VO2_bout_estimated_mean Mean VO2 estimated from HR with the individual equation (mL/kg/min) Diff_measured_minus_estimated Measured minus estimated (mL/kg/min); positive = underestimation Error_abs Absolute value of that difference (mL/kg/min) Error_pct_MAPE Absolute percentage error of the individual estimate (%) Lactate_baseline Blood lactate at baseline (mmol/L). A single sample taken at the start of the day, before the WFST; it is the reference baseline for both the WFST and the bouts, which were performed the same day. Lactate_test_post3min/_post5min Blood lactate 3 and 5 min after the WFST (mmol/L) Lactate_bout_post3min/_post5min Blood lactate 3 and 5 min after the bout (mmol/L) Lactate_comp_baseline Blood lactate at baseline on the competition day (mmol/L) Lactate_comp_post3min/_post5min Blood lactate 3 and 5 min after the competition (mmol/L) RPE_test_general / _arm Rating of perceived exertion after the WFST, whole body and fencing (armed) arm (Borg CR-10) RPE_bout_general / _arm Same, after the bout RPE_comp_general / _arm Same, after the competition Missing values are left blank. Not all variables are available for all fencers: RPE_bout_arm n = 10; competition lactate and RPE n = 6-7 (only seven fencers competed). Sheet "WFST_raw" - 372 rows Fencer Anonymised code HR Heart rate (beats/min), 10-s average VO2 Oxygen uptake (mL/min), 10-s average VO2_kg Oxygen uptake (mL/kg/min), 10-s average These are the paired HR-VO2 data from which each individual equation was fitted. Artefacts were removed as described in the article, so the number of intervals per fencer is smaller than the nominal test duration. Sheet "Bout_raw" - 502 rows Fencer Anonymised code Time Elapsed time of the recording (hh:mm:ss) HR Heart rate (beats/min), 10-s average VO2 Oxygen uptake (mL/min), 10-s average VO2_kg Oxygen uptake (mL/kg/min), 10-s average Directly measured values during the monitored bout, used as the criterion. -------------------------------------------------------------------------------- 5. 02_competition.xlsx -------------------------------------------------------------------------------- One sheet per fencer (F3, F4, F5, F7, F8, F9, F10), second-by-second. Only these seven fencers took part in the simulated competition. Fencer Anonymised code Crono_hms Elapsed competition time (hh:mm:ss), starting at 00:00:00 Clock_time Time of day (hh:mm:ss) HR_bpm Heart rate (beats/min). Blank where the signal was lost. VO2est_mL_kg_min VO2 estimated as (a + b*HR)/body mass, using that fencer's individual equation from the WFST (mL/kg/min). Blank where HR is missing, since no estimate is possible without HR. Bout "Yes" during a bout, "No" between bouts Phase Competition round: POULE (pool) or ED (direct elimination) Note on negative values: a small number of estimates are slightly negative. This is not a data error. The individual equation is fitted over the HR range of the WFST, and between bouts HR can fall below that range, where the linear extrapolation crosses zero. Values are reported as calculated, without truncation. -------------------------------------------------------------------------------- 6. 03_competition_F7_tidy.csv -------------------------------------------------------------------------------- Long-format extract of sheet F7 of 02_competition.xlsx, used by 05_competition_figure.R to produce the competition figure of the article (representative participant). ID Anonymised code (F7) sec Elapsed competition time (s) vo2 Estimated VO2 (mL/kg/min) bout "Yes" during a bout, "No" between bouts -------------------------------------------------------------------------------- 7. HOW TO REPRODUCE THE MAIN RESULTS -------------------------------------------------------------------------------- All results reported in the article can be recomputed from these files. Individual equations, r and SEE Fit VO2 = a + b*HR per fencer on sheet "WFST_raw" (columns HR and VO2). Reproduces columns a, b, r, R2, SEE of "Summary_per_fencer". Group mean r = 0.78; mean SEE = 3.07 mL/kg/min. Measured VO2 during the bout Mean of VO2_kg per fencer on sheet "Bout_raw". Reproduces VO2_bout_measured_mean. Group: 23.3 +/- 6.1 mL/kg/min. Estimated VO2 and agreement Apply (a + b*HR)/body mass to the bout HR, or use the columns already provided. Group: estimated 21.2 +/- 7.4; bias (measured - estimated) 2.07 mL/kg/min, 90% CI [0.16, 3.99]; MAE 3.24; MAPE 14.45%. Script 04_bland_altman_equivalence.R reproduces the Bland-Altman and equivalence figure (requires the R packages ggplot2 and patchwork). Competition From 02_competition.xlsx: mean VO2est over all rows = session VO2 (13.4 +/- 5.9 mL/kg/min); mean over rows with Bout = "Yes" = VO2 during bouts (21.9 +/- 6.5); maximum of those rows = peak (30.7 +/- 8.3). Counting rows with HR gives the session duration (187.7 +/- 8.5 min) and counting rows with Bout = "Yes" gives combat time (33.3 +/- 9.5 min). Software used in the article: jamovi 2.7.32.0; R 4.5.2 with TOSTER, SimplyAgree and ggplot2. -------------------------------------------------------------------------------- 8. UNITS -------------------------------------------------------------------------------- VO2 relative mL/kg/min (mL*kg-1*min-1) VO2 absolute mL/min Heart rate beats/min Blood lactate mmol/L (mmol*L-1) RPE Borg CR-10 scale (0-10) Time hh:mm:ss or seconds, as indicated per column