Article Text

Download PDFPDF

Increased running speed and previous cramps rather than dehydration or serum sodium changes predict exercise-associated muscle cramping: a prospective cohort study in 210 Ironman triathletes
  1. Martin P Schwellnus1,2,
  2. Nichola Drew1,
  3. Malcolm Collins1,2,3
  1. 1UCT/MRC Research Unit for Exercise Science and Sports Medicine, Department of Human Biology, University of Cape Town, Cape Town, South Africa
  2. 2International Olympic Committee (IOC) Research Centre, Cape Town, South Africa
  3. 3South African Medical Research Council, Cape Town, South Africa
  1. Correspondence to Martin P Schwellnus, UCT/MRC Research Unit for Exercise Science and Sports Medicine, Department of Human Biology, University of Cape Town, 3rd Floor, Sports Science Institute of South Africa, Boundary Road, Newlands, Cape Town 7700, South Africa; martin.schwellnus{at}


Background Despite the high prevalence of exercise-associated muscle cramping (EAMC) in endurance athletes, the aetiology and risk factors for this condition are not fully understood.

Aim The aim of this prospective cohort study was to identify risk factors associated with EAMC in endurance triathletes.

Methods 210 triathletes competing in an Ironman triathlon were recruited. Prior to the race, subjects completed a detailed validated questionnaire and blood samples were taken for serum electrolytes. Immediately before the race, pre-race body weight was obtained. Body weight and blood samples for serum electrolyte concentrations were obtained immediately after the race. Clinical data on EAMC experienced during or immediately after the race were also collected.

Results 43 triathletes reported EAMC (cramping group) and were compared with the 166 who did not report EAMC (non-cramping group). There were no significant differences between groups in any pre-race–post-race serum electrolyte concentrations and body weight changes. The development of EAMC was associated with faster predicted race times and faster actual race times, despite similarly matched preparation and performance histories in subjects from both groups. A regression analysis identified faster overall race time (and cycling time) and a history of cramping (in the last 10 races) as the only two independent risk factors for EAMC.

Conclusion The results from this study add to the evidence that dehydration and altered serum electrolyte balance are not causes for EAMC. Rather, endurance runners competing at a fast pace, which suggests that they exercise at a high intensity, are at risk for EAMC.

View Full Text

Statistics from


Several medical problems are encountered by Ironman triathletes.1 Of particular interest are consistent reports that exercise-associated muscle cramping (EAMC) is one of the most common medical problems encountered by the competitors in an Ironman event.1,,4

EAMC can be defined as a syndrome of involuntary painful skeletal muscle spasms that occur during or immediately after physical exercise.5 6 EAMC usually presents as painful localised muscle cramping that occurs spasmodically in different exercising muscle groups—usually the calf, hamstring or quadriceps muscles.7 8 EAMC should be distinguished from generalised involuntary muscle contractions in non-exercising muscles which are associated with a number of acute or chronic and congenital or acquired medical conditions.6 9

Although EAMC is common in triathletes, the aetiology and risk factors of the condition are still not well researched, and remain controversial.6 10 11 A number of hypotheses concerning the aetiology of EAMC have been proposed and the evidence in support of these has been recently reviewed.6 The more traditional hypotheses state that heat,12 dehydration12 and electrolyte depletion10,,12 are the cause of EAMC.5 The evidence supporting these traditional hypotheses is limited and has recently been reviewed.6 In contrast, we and others have published data from prospective cohort studies that showed no relationship between serum electrolyte changes and the development of EAMC,13,,15 or measures of dehydration and the development of EAMC.14 15 Despite these findings, triathletes, coaches and scientists10 11 16,,20 still remain to be convinced that the development of EAMC is not necessarily related to dehydration and electrolyte depletion.

In 1997, we5 introduced a novel ‘muscle fatigue’ hypothesis for the development of EAMC which suggested that the development of ‘neuromuscular fatigue’ leads to abnormal neuromuscular control. More recently, we reviewed the scientific basis for the ‘abnormal neuromuscular control’ hypothesis for EAMC in detail.6 We concluded that a number of factors may result in the development of abnormal neuromuscular control during exercise, and these factors may all predispose to the development of EAMC. It has been documented that fatigue increases the muscle spindle afferent activity and decreases the Golgi tendon organ afferent activity,21 22 which may result in increased alpha motor neuron activity. Therefore, as progressive muscle fatigue develops, there may be an increased risk of developing EAMC.5 It has been shown that there is increased baseline electromyographic activity in triathletes suffering from EAMC immediately after the race, when compared with a non-exercising control muscle.15 Therefore, endurance athletes who compete at a pace that is faster than their usual training pace (higher relative exercise intensity) may develop muscle fatigue earlier during a race and this may be a risk factor for EAMC. However, this has not yet been studied.

The Ironman triathlon, where the risk of developing EAMC is high, is an ideal model to study the risk factors and possible aetiological mechanisms for the development of EAMC. The aim of this study was therefore to determine the risk factors for EAMC in a cohort of triathletes participating in an Ironman triathlon.



All 1136 triathletes (970 male (85.4%) and 166 female (14.6%)) who entered an Ironman triathlon (3.8 km swim, 180 km cycle and a 42.2 km run) were considered as potential subjects. In the 2 months prior to the event, information about the study and research team contact details were posted on an official race website. Prior to the study, the protocol was approved by the Faculty of Health Science's Research Ethics Committee of the University of Cape Town (reference 425/2005), as well as the general organising committee and the medical subcommittee of the Ironman triathlon.

Recruitment for the study took place at race registration during the 3 days prior to the event. A detailed flow diagram summarising the recruitment of the subjects for this study is shown in figure 1. Of the triathletes who entered the event, 992 (87.3%) started the race and 399 (40.2% of race starters) consented to participate in this study. After giving written informed consent, 304 (76.2%) participants completed a modified previously validated pre-race questionnaire.15 23 24 The following information was used for this study: (1) demographic details (including age, height, weight and gender); (2) previous participation in cycling, running and Ironman events (including personal best times); (3) training details in the last 15 weeks before the event (training distances and hours spent training); (4) frequency of flexibility training and stretching exercises; (5) family history of EAMC and nocturnal cramps; (6) anticipated fluid intake during the event (type, amount and frequency); (7) personal general medical history, including previous EAMC and recent flu-like symptoms and, finally, (8) medication and supplement usage.

Figure 1

Flow diagram depicting the recruitment of subjects. EAMC, exercise-associated muscle cramping.

Pre-race and post-race blood sample and biochemical analysis

During registration a pre-race 5 ml venous blood sample was drawn into a lithium heparin vacutainer tube from the ante-cubital vein while the subject was seated. An immediate post-race blood sample was obtained in a similar fashion after body weight had been recorded.

The samples were immediately centrifuged at 3000×g for 10 min at 4°C and the serum stored at −20°C until analysis. Serum sodium concentrations were determined using an EasyLyte PLUS Na/K/Cl analyser (Medica Corporation, Bedford, Massachusetts, USA). Absolute changes in serum sodium concentrations during the race were calculated as the difference between the post-race and pre-race values. The absolute change was also divided by the pre-race concentration and expressed as a percentage.

Pre-race body weight measurement

Pre-race body weight (in swimming costumes) was measured on the morning of the race, within 60 min before the start of the event using calibrated (with a standard weight) electronic scales (Beurer GmbH & Co., Ulm, Germany) placed on a hard flat surface. For triathletes wearing light clothing, the pre-race body weight was adjusted by subtracting an average weight of the clothing items.23 24

Immediate post-race body weight measurement and calculations

Using the same scales, post-race body weights were immediately recorded in a designated research area located 70 m from the finish line with triathletes wearing running gear but no shoes. In triathletes where post-race weights were recorded with running shoes, this was corrected for by subtracting an average weight of a pair of either male or female running shoes.23 24 The percentage change in body weight during the race, which was used as a proxy of hydration status, was calculated as the difference between the post-race and pre-race values, divided by the pre-race value. Body mass index (BMI) was calculated from the self-reported weight and height (obtained from the pre-race questionnaire), using the formula mass (kg) divided by height (m)2.

Documentation of EAMC

All episodes of EAMC, which were defined as painful, spasmodic, involuntary contractions of skeletal muscle that occurred during or in the 6 h post-race period,8 were included in this study. In order to include the post-race period, a post-race questionnaire was used as the tool to determine which athletes in the cohort developed EAMC. The questionnaire was emailed within 7 days following the event to all 399 participants, who were informed regarding the clinical symptoms and signs of EAMC at recruitment (figure 1). In the email, the triathletes were only asked whether or not they had experienced any EAMC during or within 6 h after the race. Two hundred and seventy-four (68.7% response rate) of the triathletes replied to this post-race email question. Of these, 62 triathletes indicated that they experienced cramps during or immediately after the race, while 212 triathletes did not experience any cramping.

A second email questionnaire was then sent to those 62 triathletes who reported that they did experience EAMC. In this email, further details about their cramping during or after the race were requested. Triathletes were specifically asked to answer questions related to the nature and severity of the cramps, the specific leg of the triathlon in which the cramps occurred and how these cramps were treated. Over 90% (56/62) of the triathletes with EAMC completed the second component of the post-race email questionnaire.

Of the 56 triathletes who responded with details of their cramps during the race, 43 also completed both the triathlon and the full pre-race questionnaire. Complete pre-race and post-race data on cramping were therefore available for 43 triathletes, and these represent the final cramping (CR) group. Similarly, of the 212 athletes that reported no cramping during the event, 166 had also completed both the triathlon and the pre-race questionnaire, and these represent the final non-cramping (NC) control group (figure 1).

Environmental conditions

Data on the weather conditions on race day were obtained from the South African Weather Service. The average temperature during the race was 20°C (maximum 21°C, minimum 19°C). The average relative humidity during the race was 70% and the average wind speed was 37 km/h.

Statistical analysis

Data were analysed with the STATISTICA V.7.0 (StatSoft Inc., Tulsa, Oklahoma, USA) and GraphPad InStat V.2.05a (GraphPad Software, San Diego, California, USA) statistical programs. Any significant differences between the groups were determined by a one-way analysis of variance or χ2 analysis. Once significant differences in the characteristics of the groups were identified, these categorical and continuous variables were used in a Logit linear and non-linear regression model to determine factors which best predicted EAMC during the Ironman triathlon. Statistical significance was accepted when p<0.05.


Subject and cramping characteristics

There were no significant differences between the CR and NC groups with respect to age (NC, 39±9 years, n=158; CR, 38±7 years, n=39; p=0.638), height (NC, 179±8 cm, n=149; CR, 179±10 cm, n=39; p=0.882), self-reported normal body weight (NC, 75.1±10.9 kg, n=159; CR, 76.7±11.2, n=42, p=0.385), BMI (NC, 23.5±2.3, kg m−2, n=147; CR, 24.0±2.7 kg m−2, n=39; p=0.230) and male sex (NC 82.5%; CR 93.0%; p=0.089).

Triathletes reported that they cramped either once or multiple times during the swim-leg (n=8, 18.6%), cycle-leg (n=13, 30.2%), run-leg (n=31, 72.1%) and/or 6 h immediately after the triathlon (n=17, 39.5%). The majority of the triathletes in the CR group (n=29, 70.7%) reported experiencing only mild (<5 min and were able to continue exercising) muscle cramps, while 9 (22.0%) and 3 (7.3%) reported moderate (5–15 min and were able to continue exercising) or severe (>15 min or had to stop exercising) muscle cramps, respectively. The average duration of these cramps was 12.0±31.8 min (n=34) and ranged from 0.1 to 180 min. Ninety-five per cent of the reported cramps were either in the calf (n=17), hamstring (n=19), quadriceps (n=26) or foot muscles (n=10). None of the athletes were admitted to the medical facility at the race, or were confused or comatosed as a result of experiencing muscle cramps. The majority of triathletes (n=29, 67.4%) reported that their muscles were ‘tired’. Six triathletes (14.0%) also reported that they had dark urine. The majority (n=29, 67.4%) of the triathletes stretched the muscle to relieve the cramping.

Performance and training history

The pre-race self-predicted cycle-leg, run-leg and overall times were on average significantly faster in the CR group (table 1). Similarly, the actual race-day cycle-leg and overall finishing times were also on average significantly faster in the CR group (table 1). There were no significant differences between the predicted and actual swim-leg times between the two groups. In addition, both the NC and the CR triathletes completed each leg and the overall triathlon on average slower than their predicted times.

Table 1

The predicted and actual performance times of the non-cramping (NC) and cramping (CR) triathletes during the Ironman Triathlon

When co-varied for sex and age, there were no significant differences in any of the previous career and recent performance parameters between the two groups, indicating that both groups were similarly matched for previous performances (from sprint triathlons to Ironman triathlon distances and from 5 km to ultra-marathon running) (table 2 and other data not shown). Both groups of triathletes also competed in the various triathlon and running disciplines for a similar number of years and a similar number of events in each discipline (data not shown).

Table 2

Triathlon (standard and Ironman) and road running (21.1 km and 42.2 km) career personal best times and best times achieved over the last 12 months (triathlons) or 15 weeks (road running) of the non-cramping (NC) and cramping (CR) triathlete groups

There were no significant differences in training frequency or volume (distances and durations) for the various disciplines between the CR and the NC groups during the 15 weeks and 1 week before the event (table 3).

Table 3

Self-reported swimming, cycling, running and/or total training frequency distances and durations for the 1-week and 15-week period before the Ironman triathlon of the non-cramping (NC) and cramping (CR) triathletes

There were also no significant differences in the average cycling speeds of the NC and CR groups, while training during the 15 weeks (p=0.127) and 1 week (p=0.668) prior to the triathlon, as well as during a race over 80 km during the 15-week period prior to the triathlon (p=0.539) (data not shown). The average distance of the cycling race was 107±25 km (n=123) and 105±22 km (n=32) for the NC and CR groups, respectively (p=0.660). However, the triathletes in the CR group cycled on average relatively faster during the Ironman triathlon than those in the NC group when compared with their personal best time over a race longer than 80 km during the 15-week period (p=0.042 and p=0.031 when co-varied for the race distance).

Serum electrolyte and body weight changes

There were no significant differences in the pre-race, post-race or pre-race–post-race changes in body weight or serum electrolyte (sodium, potassium and chloride) concentrations between the groups (table 4).

Table 4

Pre-race and post-race as well as absolute and relative changes in the serum sodium concentrations [Na+] and weight in the non-cramping (NC) and cramping (CR) triathletes during the Ironman triathlon

Estimated fluid intake and strategy

Triathletes in the CR and the NC groups estimated similar volume fluid intakes for the swim- (data not shown), cycle- (NC 4.2±1.6 l, N=155; CR 4.2±1.5 l, N=41; p=0.836) and run-legs (NC 2.7±1.4 l, N=148; CR 2.9±1.5 l, N=41; p=0.419). They also predicted similar fluid intake strategies for the event (data not shown). These observations did not alter when co-varied for performance times (data not shown).

Past cramping history

Triathletes in the CR group (82.9%), had a significantly higher reported history of EAMC (OR=5.8, 95% CI 2.4 to 13.9, p<0.001) compared with the NC group (45.5%). There was no significant difference in the prevalence of family history of EAMC (CR 32.5% vs NC 23.7%, p=0.351) or nocturnal cramping (CR 26.2% vs NC 22.9%, p=0.812) between the groups.

There was no significant difference in the number of years of cramping in the triathletes with a history of EAMC in both the NC and CR groups (p=0.194) (table 5). Similarly, there was no significant difference in the number of triathletes in each group who have experienced EAMC during the last year (p=0.232). Triathletes in the CR group, however, reported (1) a significantly greater mean number of cramps in their last 10 races (p=0.003) but not the last 10 training sessions (p=0.814), (2) a greater frequency of cramping during running (p=0.030) and (3) a greater frequency of whole body cramping (p=0.006), compared with the triathletes in the NC group (table 5). In both groups, cramping is experienced in the last quarter or immediately after the race in >80% of triathletes (data not shown). The majority of triathletes also reported that they experience mild muscle cramps which usually last less than 5 min and the triathlete was able to continue exercising (table 5). The major sites of cramping in both groups were the calves, hamstrings, quadriceps and/or feet (data not shown).

Table 5

Cramping history of triathletes in the cramping (CR) and non-cramping (NC) groups who had a history of cramping

Flexibility training

There were no significant differences between the two groups with respect to frequency or duration of stretching, total weekly stretching time and the muscle groups that were stretched (table 6 and data not shown).

Table 6

Flexibility training history of the non-cramping (NC) and cramping (CR) triathletes who participated in the Ironman triathlon

Risk factors for EAMC in triathletes

The significant independent risk factors for EAMC in Ironman triathletes were the reported number of cramp episodes in the past 10 races (p=0.011) and overall finishing time for the event (p=0.010) (table 7). In particular, post-race serum sodium concentrations and relative change in body weight (as a measure of hydration) were not risk factors for EAMC in these athletes. In a separate analysis where only two variables were considered, both relative cycling speed (percentage of their personal best cycling speed in the 15 weeks before the Ironman event) (estimate±SE=−0.05±0.02, p=0,037) and the number of reported cramps in the last 10 races (estimate±SE=−0.39±0.11, p<0.001) were also significant predictors of EAMC in these triathletes.

Table 7

Regression analysis for the determination of independent risk factors for EAMC in Ironman triathletes


This study is, to our knowledge, the largest prospective cohort study to determine risk factors for EAMC in athletes. The main novel finding of this study was that the two independent risk factors for the development of EAMC in Ironman triathletes were a history of muscle cramping (in particular the number of EAMC reported in the last 10 races) and an overall faster race time during the Ironman triathlon. Triathletes who developed EAMC during or in the 6 h after the Ironman triathlon predicted faster overall cycle and run times before the race, and then proceeded to cycle and run faster during the race when compared with non-cramping triathletes who reported similar training and past performance parameters. Other significant findings from this study were that there was no difference in the serum sodium concentrations, body weight changes (as an indicator of hydration status), flexibility training, BMI, age or predicted habitual fluid intake in the cramping and non-cramping triathletes.

To date, there are only observational data in marathon runners to indicate that EAMC is more likely during high intensity racing.13 25 In this study, the group of cramping athletes predicted significantly faster run, cycle and overall finishing times than those triathletes who did not cramp. In our study, it is important to note that (1) the time predictions by the athletes were made 1–3 days before the event; (2) both groups of athletes reported similar personal best times and performances, ranging from sprint triathlon events to Ironman distance triathlons and from 5 km to ultra-marathon running distances; (3) all-time best performances and recent best performances were similar in the two groups and (4) there was no significant difference between the groups with respect to their training volume or intensity before the Ironman race.

Therefore, despite the cramping and non-cramping athletes being matched for performance and training parameters pre-race, triathletes in the cramping group predicted significantly faster running, cycling and overall performance time. Studies conducted in other Ironman triathlons have shown that those triathletes who performed better had done a significantly greater training volume before the race.26 Both groups of athletes in our study had performed similar amounts of training, yet the cramp group predicted faster times. It can thus be concluded that, although both cramping and non-cramping triathletes were similar in ability and preparation, triathletes in the cramping group intended to perform better. This would necessitate racing at a faster pace and subsequently at higher exercise intensity.

Data collected on the actual race day confirmed that both cycling and overall performance times of the cramping group of triathletes were significantly faster compared with the non-cramping group. Furthermore, there was a trend for the run times to be faster in the cramping group. Therefore, it appears that triathletes in the cramping group predicted and then subsequently raced at a higher intensity compared with non-cramping triathletes. Although we did not measure relative exercise intensity during the race, it appears that the cramping group exercised at a higher relative intensity during the race compared with the non-cramping group. In support of this were the data showing that triathletes in the cramping group cycled at an increased relative cycling speed compared with triathletes who did not cramp during the race.

Data from observational studies show that EAMC is related to self-reported poor conditioning for an event,25 exercising at a higher intensity such as during racing compared with training,25 and that EAMC is more common in the latter stages of a race.13 25 Our findings provide further evidence that the development of muscle fatigue is likely to be an important predictor of the development of EAMC during triathlons. It is interesting to note that in our study, as in previous studies, most triathletes developed EAMC in the latter part of the race or immediately after the race. Only a small number (n=8) of triathletes reported muscle cramping during the swim stage of the event. The mechanism for cramping during swimming may well differ from that during weight bearing sports, and requires further investigation.

Although we cannot directly link the faster race pace to the development of muscle fatigue in our study, data from observational studies in animals21 22 and laboratory-based studies in humans27,,30 support the hypothesis that repetitive muscle contraction results in fatigue and that this can lead to cramping, probably by altering neuromuscular control.6

A history of cramping was also a significant risk factor for EAMC in our cohort of triathletes. In particular, a history of cramping in the past 10 races, but not training sessions, was strongly associated with EAMC during the Ironman triathlon. However, cramping triathletes did not report a greater frequency of a positive family history of EAMC, or a positive family history of nocturnal cramping.

Another important finding from our study was that pre-race and post-race serum electrolyte (sodium and chloride) concentration changes and body weight changes did not differ between cramping and non-cramping triathletes. Our findings confirm those of other published prospective studies where no relationship between either serum electrolyte concentrations,13,,15 or hydration status14 15 and the development of cramping during endurance events was documented. The hypothesis that either serum electrolyte changes, in particular serum sodium changes, or dehydration is related to EAMC, is therefore not supported by the findings in our study.

Other previously reported possible risk factors for EAMC include higher BMI, older age group, longer history of running, family history of cramping and poor flexibility training.31 In our study, we did not find a relationship between any of these variables and the risk of EAMC. Although there was a trend for male sex to be associated with the development of EAMC in our study, this finding was not confirmed when other variables were entered into the model predicting risk factors for cramping.

What is already known on this topic

One of the most common medical conditions in endurance athletes is exercise-associated muscle cramping (EAMC). Despite this, there is still controversy surrounding the aetiology and risk factors for this condition. The most common hypotheses for the cause of EAMC are electrolyte depletion and dehydration. However, recent evidence challenges these hypotheses and suggests that abnormal neuromuscular control resulting from muscle fatigue and other factors may be responsible for EAMC.

What this study adds

The results of this study show no relationship between serum electrolyte disturbances or dehydration and the development of EAMC in Ironman triathletes. Rather, independent risk factors for EAMC were a history of muscle cramping (in particular the number of EAMC reported in the last 10 races) and an overall faster race time during the Ironman triathlon. Prolonged exercise at a relatively higher intensity, compared with training, is therefore associated with the development EAMC in triathletes.

The main strength of this study is that it is the largest prospective study investigating possible risk factors for EAMC in endurance athletes. A large cohort was recruited before the race, and measurements were taken before and immediately after the race. One limitation of this study is that the diagnosis of EAMC was self-reported and could not be confirmed by clinical examination by a health professional. However, the diagnosis of EAMC is mainly based on symptoms which are clear and could be explained to an informed athlete. A further limitation is that some subjects in the original cohort were lost to follow-up as they did not respond to the email questionnaires. The overall response rate to the first email questionnaire was almost 70%, and the response rate to the second questionnaire requesting details of the cramping from the EAMC subgroup was over 90%.

In summary, the results of our study confirm that there is no correlation between EAMC and changes in electrolyte concentrations or changes in hydration status. Rather, this study shows that EAMC is associated with a history of muscle cramping (in particular the number of EAMC reported in the last 10 races) and an overall faster race time during the Ironman triathlon. Prolonged exercise at a relatively higher intensity, compared with training, is therefore associated with the development EAMC in triathletes.


The final preparation of this study for publication was supported in part by the International Olympic Committee (IOC) Research grant to the Clinical Sports Medicine Group of the UCT/MRC Research Unit for Exercise Science and Sports Medicine of the University of Cape Town.


View Abstract


  • Funding This study was supported in part by funds from Discovery Health, the South African Medical Research Council (MRC) and the University of Cape Town Staff Research Fund.

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the Research Ethics Committee, Faculty of Health Sciences, University of Cape Town.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.