TABLE 1
Triathlon Race Distances (km)

With exception, competitors focus on either ‘long’ or ‘short’ competition (Bentley et al., 2002a). They also race at either ‘age-group’ or ‘elite’ level. ‘Age-group’ athletes compete in 5-year age categories and in a separate race to ‘elite’ triathletes. World Championship races are available, over the ‘Olympic distance,’ to both ‘age-group’ and ‘elite’ triathletes. Elite triathletes may also compete over ‘World Cup’ or ‘International’events organised by either the world or their regional governing body. Athletes who finish within the top 50 in a World Cup or World Championship race, or within the top 10 in an International race, are awarded world-ranking points. The higher the finishing position of the athlete, the more ranking points he or she obtains. These world-ranking points are accumulated over each 4-year qualification cycle for the Olympic Games. They are used to determine the world ranking of the athlete in question. Only triathletes with a world ranking of less than 125 at the end of each qualification cycle, or winning a place for their National Olympic Committee at the World Championships immediately preceding the Olympiad, are eligible for selection for the Olympic Games. They are termed ‘world ranked elites’, to distinguish them from ‘elite’ triathletes in general, in this thesis. ‘Elite’ triathletes are defined as members of a National Squad who compete regularly at senior international level. ‘Sub-elite’ triathletes are defined as athletes (typically members of a National Development Squad) who compete immediately below senior international level. They regularly place within the top twenty finishers of the 20-29 year age group at national level. Non-elite triathletes are defined as those athletes who compete in age-group triathlons on a recreational level. They neither feature within a National Triathlon Squad nor within the top 20 finishers at either national or international age-group events (Lothian and Donovan, 1999).

A recently published review of the technical and tactical specificities of triathlon competition (Bentley et al., 2002) clearly indicates that the needs of the sport are unique from, and are not simply a combination of, those of swimming, cycling and running. Although the ‘Olympic distance’ triathlon has received the bulk of recent research attention, work in this field is still in its infancy. Just over 500 articles with ‘triathlon’ or ‘triathlete’ as a key word have been published in ‘PubMed’ indexed journals to date. Little is known either about how triathletes can best train for this event or how training patterns may differ between elite and sub-elite triathletes. Very little epidemiological data describing the relationship between training volume (km), training duration (hours), or training intensity and performance, and injury exist. ‘Not enough is (therefore) known about planning individual training programmes scientifically (O’ Toole et al., 1987a).

O’ Toole et al. (1988) stated that ‘a fine line exists between the level of training that is required for optimal performance, and that which induces problems’. Research in other endurance sports has shown that where there is uncertainty about the balance between training stress and recovery that is required for optimal adaptation, training practice may be excessive and inappropriate (Steinacker et al., 1998). Under these circumstances, negative overreaching may result. Negative overreaching is defined as an accumulation of excess training and non-training stress resulting in a short-term decrement in performance capacity with or without related physiological and psychological signs (Fry et al., 1992a). Alternatively, some specific system may break down. For example, overuse injury, defined as that caused by the repetitive stress of identical movements, (van Mechelen, 1992) may result.

It is not clear from the literature to what extent negative overreaching, as manifested by overuse injury, occurs in the elite and sub-elite triathletes who specialise in ‘Olympic distance’ competition. This is due to a number of factors including the use of subject groups that are heterogeneous with respect to ability, event distance or sex in the studies that have been carried out so far. Nor has adequate examination of the training related factors for overuse injury been undertaken. Only two prospective studies of training and injury in triathletes (Burns et al., 2003 ; Korkia et al., 1994) have been carried out. Almost no data relating how training structure and intensity change over the athlete’s year are available. It is not clear at which points in this year the athlete may most be at risk of maladaptation.

The aim of this review is to provide both a brief overview of training theory and of the data relating to it that may be gleaned from the triathlon literature.

For positive over-reaching and supercompensation to occur, the athlete must be allowed a period following training stress during which no further homeostatic imbalance occurs (i.e. recovery). The length of the recovery period required depends on the time needed to complete the recuperative processes. This in turn depends on the degree of fatigue induced in the athlete by the training stress. Performance improves when successive training stimuli are imposed when the athlete’s functional systems are in a state of supercompensation. If the resting period between two stimuli is either too short or too long, a training effect will not be achieved. If the resting phase is too long, the effects of overcompensation on performance will be reversed. If no secondary stimulus for adaptation is applied during the supercompensation period, the degree of adaptation reverses and the training effect will be lost. Alternatively, if a second stimulus is applied before the athlete has experienced full regeneration, the athlete will still be suffering from the fatigue of the previous stimulus (Busso et al., 2002). Under such conditions the training load will result in greater homeostatic stress than previously, and the processes of adaptation will be placed under additional stress. In this situation the athlete will be placed under greater metabolic load to achieve the same degree of work. If the athlete is repeatedly exposed to fatigue inducing stimuli without sufficient recovery, his capacity to adapt is placed under continuous strain. This stress may manifest itself in symptoms of negative overreaching and eventually to exhaustion of the athlete’s capacity to adapt or overtraining syndrome (Fry, Morton and Keast, 1992a).

The requirement that training loads be administered in a logical fashion, to promote positive adaptation and prevent overtraining, implies a systematic and well-planned approach to the development of a training programme. Training has to be organised into phases of training with high load so as to induce a training response, and phases of lower load to allow recovery (Bompa, 1989). The process whereby these objectives are incorporated into the training programme has been termed periodisation of training (Fry, Morton and Keast, 1992b). According to this process, the training year is normally divided into three major macro-cycles, each lasting between one and six months. The reason for this division is that not all training objectives can be achieved on at the same time (Harre, 1982 ; Sleamaker, 1989 ; Steinacker et al., 1998). In addition, the development of some athletic abilities is prerequisite to others. Certain aspects of performance have to be emphasised for certain periods while others are simultaneously stabilised or maintained (Harre, 1982 ; Fry, Morton and Keast, 1992b).

These major macro-cycles referred to above are termed preparation (or endurance base), competition and transition (Fry, Morton and Keast, 1992b). They are repeated, with modifications, on an annual basis.

The training year involves a gradual progression of emphasis from general training- through specific training- to full competition training. Each macro-cycle is itself (generally) divided into 3 to 6- week blocks of training called meso-cycles (Fig. 1). Two or three meso-cycles are generally followed by one or two recovery meso-cycles. Each meso-cycle is itself normally composed of micro-cycles of 7-, 10-, or 14-day blocks of training. The latter may be interspersed with recovery days. Both microcycles and individual training sessions (which are referred to as training units) should be structured in alternating periods of high, moderate and low load. In this manner, recovery between training bouts should be optimised..

Minimal information is available on how to periodise training load in the elite triathlete, such as to minimise the chance of him becoming overtrained, (Lucia et al., 1996 ; Dreano, 1998). The most detailed reports of the training structure followed by elite triathletes so far published are those provided by Dreano (1998, Table 2) and Millet et al. (2002a).

Table 2 shows case study data (from Dreano, 1998) for a world ranked elite male triathlete specialising in Olympic distance competition. The athlete in question trained more than 696 hours over a 298-day period. He covered over 12,000 km in the process. No breakdown of training at different intensities in the different training months and or macro-cycles covered by the study was given. It is not known to what extent the periodisation structure that was followed was appropriate to the athlete. No information relating to what extent the training stress was reflected by performance, by overuse injury, or by illness was available. Nor were the data even representative of the athlete in question. His training was severely interrupted at one point as the result of a traumatic injury sustained during a cycle crash (Dreano, personal communication, 1997). Clearly, more data on how (groups of, rather than individual) elite triathletes structure their training over the year is required.

TABLE 2
Number of races competed in and training workload completed by an Elite male triathlete

Most reports of the number of training hours done per year by ‘Olympic distance’ triathletes are gleaned from studies that did not have the gathering of such information as their primary aim. Unfortunately, despite the fact that training volume and intensity commonly vary throughout the training year, few authors report the training phase to which their results pertain. Data from two studies clearly show that to report volume for a ‘typical training week,’ without referring to the training phase to which the data pertains, can be misleading. Lucia et al. (1996) assessed nine elite members of the Spanish National Triathlon Squad in the pre-competitive (winter to January), competitive (spring to June) and rest periods (autumn to October) of their training year. They found significant differences between the pre-competitive and competitive periods in both weekly training volume and duration in each discipline. Rowbottom et al. (1997) assessed eight elite and sub-elite triathletes every six weeks over a nine-month season. Swimming, cycling and running distance increased 55 %, 65 % and 71 % over the duration of the study. These increases corresponded to mileage changes from 14.6 (± 3.1) to 22.6 (± 33.7) km per week in swimming, from 156 (± 26) to 257 (± 32) km per week in cycling, and from 26.3 (± 3.0) to 45.1 (± 4.8) km per week in running. These data highlight the importance of assessing the variation in training load and intensity in triathletes over both the calendar and training year. There is a need to know within which months and or training macro-cycles ‘Olympic distance’ triathletes are most at risk of developing negative over-reaching or overtraining.

A further potential source of inaccuracy in the literature lies in the manner by which most authors report training information. In addition to not reporting which cycle of the training year that they are referring to, the majority of authors report training data for an average week. They do this without apparently having checked whether the athletes under study were training according to seven-day micro-cycles. If, for example, the athletes are operating in micro-cycles lasting 10- or even 14- days, training data collected over one-week might be misleading. It is therefore suggested that future reports of training data in triathletes refer to the number and duration of the macro-, meso-, or micro-cycles to which they refer.

Nor is it clear whether weekly total exposure to training stress varies with sex, ability, and event specialisation. Chatard et al. (1998) reported elite ‘Olympic distance’ males to train over 20 hours per week on average but the values reported in a recent study by Millet et al. (2002a) add up to just over 16.5 hours per week. The mean of the literature for weekly training time of competitive ‘Olympic distance’ males (15.15 hours per week) appears lower than that of the equivalent ‘Ironman distance’ group- but is estimated from only 16 subjects. Most of the limited number of papers that do report weekly training time do not differentiate between elite, sub-elite and non-elite triathletes ; between males and females ; or between ‘Olympic distance’ and ‘Ironman distance’ specialists. The recent study by Egermann et al (2003), that reported no relationship between the weekly training time of 656 triathletes competing in ‘Ironman’ Europe 2000, age and sex, is an exception. Egermann et al. (2003) noted no significant difference in weekly training time between athletes with a competition time of less than 10 hours, 10-12 hours, or more than 12 hours (17.6 ± 6.0 hours, 15.7 ± 5.5 hours and 14.6 ± 6.2 hours, respectively), but did not report the number of athletes in each group. More detailed comparative data is required (Table 3).

Nor do many data assessing whether training mileage differs between ‘short’ and ‘long’ distance (Bentley et al., 2002), elite and non-elite, or male and female triathletes exist. In some cases (Kohrt et al., 1989) the reported data might again be unrepresentative of the athlete group referred to. Kohrt et al. (1989) reported weekly swimming, cycling and running mileage values of 3.9 km.wk^-1, 181.3 km.wk^-1, and 51.8 km.wk^-1 respectively for a group of triathletes competing over a 1.9/90.3/21 km triathlon. The swim mileages were well below the mean of the literature (Table 4).

Data from other studies may also be misleading with respect to the training mileage of ‘Olympic distance’ triathletes because the authors did not split their subjects by sex (Migliorini, 1991), by ability (Korkia et al., 1994) or by event distance specialisation. Collated information from elite, sub-elite or non-elite triathletes, from ‘Olympic distance’ and ‘Ironman distance’ triathletes, and or from both males and females may not give an accurate picture of any of these groups. Holly et al. (1986) compared the training volume for males finishing within the top 15 (n=4) and outside the top 15 positions (n=2) at the Hawaii Ironman competition. They showed daily volume to differ significantly (2.7 km vs. 1.6 km swimming, 84.0 km vs. 49.9 km cycling, 13.6 vs. 7.8 km running) between the groups. Further investigation of the extent to which training mileages may differ between elite, sub-elite and non-elite athletes ; between male and female competitors ; or between ‘Olympic distance’ and ‘Ironman distance’ specialists, is needed.

TABLE 3
Weekly training time data (h) (± SD) reported by the triathlon literature

It is not clear whether triathletes allow themselves adequate recovery between successive training loads. Few studies report the frequency of training sessions. Even fewer studies include data specific to a particular triathlete ability, distance or sex group. Lindeman (1990) reported a group of 46 male and 6 female ‘Olympic distance’ triathletes to complete 2.5 (± 2.3) sessions per day. Korkia et al. (1994) detailed training data for a group of British triathletes of both sexes (n=155). Their male sample (n=124) comprised 14 long distance specialists and 101 ‘Olympic distance’ athletes. Ten of the latter were classed as elite, 9 as sub-elite and 14 as recreational athletes. The 31 females surveyed comprised one long distance competitor, and 27 ‘Olympic distance’ athletes of whom — were sub-elite and — were recreational athletes. The authors showed that ‘Olympic distance’ athletes did 2.4 (± 1.2) swims, 2.8 (± 2.0) cycle sessions, and 2.6 (± 1.6) run sessions per week. This load was less than triathletes concentrating on longer events, who completed 3.1 (± 1.3) swim sessions, 4.3 (± 2.3) cycle sessions and 2.6 (± 1.4) run over an equivalent period. Unfortunately, Korkia et al. (1994) did not undertake a comparative analysis of training frequency for the different ability groups. Moreover, it is likely that some of the athletes in the study should have been classed as ‘age-group’, rather than as ‘elite’, triathletes. Elite ‘Olympic distance’ triathletes might train more frequently than Korkia et al. (1994) reported. The male elite triathlete whose training was reported by Dreano et al. (1998) averaged 3.85 swim sessions, 6.2 cycle sessions and 2.5 run sessions per week. These values were averaged over a period over which the traumatic injury that he suffered was likely to have hampered him from training as frequently as normal. Unpublished case data from another elite male, ranked in the world top 50 and coached by one of these authors (Vleck, personal communication), indicated him to commonly train three and sometimes even four times a day. Busso et al. (2002) have pointed out that training frequency will influence the magnitude and duration of fatigue induced by a single training session. A short time between sessions leads to more persistent fatigue. Although the relative extent to which recovery is required from a swimming stress as opposed to a cycling stress or running stress is not known, it is suggested that elite or sub-elite ‘Olympic distance’ triathletes may often commence training sessions before full adaptation to previous session(s) has taken place.

TABLE 4
Means of the triathlon literature for training distance (km).

It is not known to what extent elite and sub-elite ‘Olympic distance’ triathletes distribute ‘hard’, ‘moderate’ and ‘light’ intensity training sessions within each training disciplines and within a training micro-cycle. Nor is it clear whether the approach taken is appropriate. Those studies that do quote triathlon training intensity tend to do so in terms of average pace per week rather than give a break down of training format. O’Toole et al. (1989b) for example, observed average weekly training paces for swimming, cycling and running of 27 (± 5.50), 18.8 (± 2.3) and 8 (± 1.28) minutes per mile, respectively. Since response to a specific workload is athlete specific this approach is not enlightening for the purposes of this review. What is an appropriate training load for one athlete may cause breakdown in another (Kreider et al., 1998a). It would be far better, when assessing the way training is structured, to report number, time or percentage of sessions spent in each of various defined intensity levels.

Future studies in the area of quantification of triathlon training should expand on the examples set by Korkia et al. (1994), Lucia et al. (1996), Rowbottom et al. (1997), Dreano (1998) and Banister et al. (1999). They should cite training data in terms of time spent at intensities that are, as far as possible, objectively defined. Lucia et al. (1996) reported training volume of elite male triathletes in terms of the percentage of their training time that they spent below ‘ventilatory threshold’, between the ‘first ventilatory threshold’and ‘second ventilatory threshold, and above the ‘second ventilatory threshold’(Beaver et al., 1986 ; Wasserman et al., 1973). The ideal epidemiological study of training structure would entail the daily prospective recording of time spent within training intensity levels that are defined according to the results of regular physiological testing. The time required for laboratory testing, however, also means that such studies could only be conducted with relatively small subject numbers. It is suggested that a prospective training studies of (National Squad) elite and sub-elite ‘Olympic distance’ triathletes, not requiring laboratory testing, be also carried out. The reader is referred to Hopkins et al. (1991) for a discussion of the advantages and disadvantages of the various methods by which training data may be quantified.

Given the lack of suitable training guidelines, the probability of the triathlete following an inappropriate training programme may be high. A poorly conceived training programme is considered to be one of the chief causes of the overtraining syndrome (4^th ACSM / USOC consensus statement on ‘overtraining’, 1999). The risk of negative ‘overreaching’ in triathletes that is already added to by the lack of properly researched, triathlon specific, training guide-lines, is probably exacerbated by the current lack of triathlon specific coaches to provide feedback on and adjust training. The Annual report of the British Triathlon Association for 2002 [⁴⁹] attests to the existence of only 30 registered ‘Level III’ and three ‘High Performance ‘triathlon coaches in Great Britain. ‘Level III’ coaches are those considered by their National Federation to be competent to plan, coach and evaluate an annual coaching programme for committed performers up to sub-elite level.’ ‘High performance’ coaches are considered competent to assume full responsibility for the organisation, management and delivery of all elements of a world-class performers or teams preparation for competitive international success. This small number of qualified coaches is in stark contrast to the number of registered triathletes in Britain. This currently stands at 4967 and has grown by an average of 11.7 % per year since 1998 (British Triathlon Association annual report, 2002). ‘The coaching structure’ is not ‘broad enough to deal with the increase in both numbers and standard of athletes’ (Lothian and Donovan, 1999).

Even if triathletes have access to separate swimming, cycling and running coaches, such coaches may work independently of one other. Triathletes may therefore be stressed within their training for each of the three component sports of the triathlon to a level only appropriate for athletes who train exclusively in that sport. The stress imposed by training in the other disciplines of the triathlon may not be taken into account. It is possible that when a swimmer, cyclist or runner trains and then subsequently rests, the triathlete training alongside him goes on to do another full quality session, in the same day, in another discipline. Whether he or she can undertake further session(s) at a time in the day that allows for maximum inter-session recovery, is limited by constraints such as having to fit training around swimming pool availability, running club evenings and having to do long cycling sessions during the weekend and daylight hours (Jacques, 1996).

Lack of coaching quantity and quality is considered to be one of the main reasons why talented swimmers, cyclists and runners do not achieve their competitive potential (Moore and Row, 1998). The same may apply for triathletes. The environmental constraints to training ; the absence of training guidelines ; and the lack of easily available methods of monitoring training adaptation ; may all contribute to a difficulty in obtaining a good balance between training stress and recovery. This imbalance, and the fact that triathletes have less desire to avoid harm in their training than swimmers, cyclists and runners (Clingman and Hilliard, 1987) may mean that triathletes are highly susceptible to negative over-reaching.